I just gotta say your comments at the end of the video are so true, spot on! It really bothers me that so many people will comment about EVERYTHING they feel you did wrong, whether they're correct or not. However, probably 99.99998% of the ones that leave those comments will never go but all those motors, film thier "Correct" testing, and then put it up for the community! I really commend everyone in the community that does do things like this to make it easier for us COMMONERS and/or NOOBS. Because of a few of my favorite "Machinist TH-camrs" I went and bought a series 60 Monarch Lathe, in fact it's the same one another pretty famous TH-cam machinist has, don't want to name drop here, his channels name is the one that has a play of word from "it's always sunny in Is Philadelphia"... Anyways, thank you for all your help and videos, we very much appreciate it!
I’ve been on the edge of my seat for any and all of the ELS videos 🤗 Some companies should put you on their payroll, since you’re fixing a lot of errors and saving other consumers!!! Thank you sir for everything you’ve done on the ELS! Not only has it been awesome to watch you develop it in each video, but it’s taught me about steppers and servos!
Hi James. Another masterclass. This time in how to deal with a manufacturer that clearly was in need of some skilled and knowledgeable help. Maybe they had looked your channel up first and were hoping for this level of support. In any event, very nicely done. It's a pleasure watching you wade through the morass! Cheers. BobUK.
Agree with Stepper Online - I've seen some even running windows XP or NT in my industry. Companies paid 10's of thousands for a tool, they won't change it if it works, or worse still, runs a real risk of causing it to no longer work
I really appreciate your testing methodology and the fact that you continue to improve on your equipment recommendations. This is why I feel confident recommending your ELS on all of the home machining forums that I belong to. As a long-time owner of a 9x20 lathe, I am a big proponent of the ELS, mainly to overcome the need for change gears and the back gear mod that is almost a requirement for this class of lathe. I especially appreciate that you continue to improve your recommendations with the goal of providing the best operation at an affordable price. Not all of us hobby machinists can afford the cost of a Clearpath solution so it's nice when someone provides an affordable alternative more in line with my budgetary constraints.
I was impressed with your persistence! I deal with software/hardware issues on a daily basis and something new like that which presented so many up front challenges would have been tossed well before getting to where you got to. Well done to you sir on a what seems like a good find.
Holy crap! Who's got the time for all this malarky? You are a man of endless patience, on your course of investigation and explication. For the rest of us - go with Teknic Clearpath. It works. It keeps on working. Nothing else to say.
I commend your patience and persistence. I would have bailed many emails ago. "Are you running Win-7?" Nope, haven't upgraded yet from XP. Man what a journey to get it "working".
Not sure how I found this channel, but you sir are a BEAST.... not just your lift abilities either. I don't need any of what you are showing, and I doubt my talent will ever get there, but I appreciate you
I can actually provice some insight into the statement, that "most of industry systems are WIN7 currently". We make equipemt that is used in industries like car manufacturing and others (usually also "production line" types of setups). "The industry" in general is incrediably avers to any sort of change, and often doesn't update anything, ever (no matter if there is support available or not), unless something actually doesn't work and needs the update (like the firmware for some device). Now it's important to understand that all systems I've ever seen or heard of are either air gapped, meaning it's physically impossible to connect to them from the outside, or heavily firewalled in multiple stages. Obviously, they can't connect to the outside either. Windows-Updates are either not possible, or have to use a WSUS server (Windows Server Update Services), which basically hosts the files for windows updates locally. Also I do know of at least some companies that do get updates, since you can still basically pay microsoft so you're still getting support for Windows 7, and from what I remember that will be available until ~2024 or something? Likely increasing in cost during that time, but I don't know that. As some more background, most companies that have these production lines contract other companies to design, build and assemble them. At some point they are delivered and there's an inspection, it's then deemed completed and functional. The company that accepted the "cell" is now responsible for keeping it working, if they update a firmware or install windows updates, that's on them. If you do update something and then something doesn't work, you call the company that made it they just say "well it worked fine with the versions we tested, look you signed off on it here *points to signature*". If a real production line stands still it can eaily cost multiple thousands of dollars PER MINUTE. So having to stop it for any sort of problem introduced by a windows updade can cost real money real fast. Keeping the systems truly isolated from the outside is just easier. One more Note: my terms may be very far off the "industry norm", my primary language isn't english so I just don't know the correct terminology, but tried to translate as well as I could.
I completely agree. Being a system administrator, updates are the most likely cause of problems on a system. If it's working, do not do the update. This is especially true on system that are not connected to the internet. All the issues James was referring to about companies getting hacked are security flaws that are discovered in old system that ARE connected to the internet.
Great insight, thanks. For myself, when building my CNC rig I specifically chose Win 7 so that I didnt have to deal with updates. End of life? That just means it should be more stable, I'll take it!
In an automotive plant even the local engineering team doesn't want to deal with their IT department. Better to not connect the automation to the network than deal with IT's restrictions, and glacier like response. Unfortunately that means some automation has cellular modems as a back door into the equipment for remote service.🤪
This does look impressive and gives me confidence to consider this motor for my applications. The ONLY downside is I am not fond of ordering direct from China and hopefully they will see the value in this motor to pickup some stock for it so guys like me are not on the fence about ordering from out of country.
This is the first of your video's I have found. I loved the relaxed vibe and the way you have done and explained your testing. I'm going to go be and watch from the start.
Aaah prolific, that explains everything. I've thrown out all my prolific cables and am only using FTDI based USB to serial cables and I haven't had any weirdness with serial since.
This brings back memories. When I was doing BIOS engineering 9 years ago we always had issues with Prolific USB to serial cables. And we had the same conclusion FTDI controllers were better. I guess nothing has changed.
@Lassi Kinnunen 81 not my circus, not my monkeys. We're talking about companies who not only sell you some piece of analytical equipment, but also an underspecced computer to communicate with it - at a hefty markup, no less. They show zero interest in removing technical debt.
@Lassi Kinnunen 81 issue we have is security settings in windows 10 dont allow sufficient access unless running in admin mode. IT wont give us admin mode... we end up having to take the computers controlling our equipment off the network and running windows 7.
at least for win xp, it's much easier to do low level stuff with it. generally it'll work on 10, but sometimes drivers are written abusing non standard behaviour of an os, or deprecated APIs, and so upgrading might brick the machine, you'll also lose any support, and generally you don't want to mess with a $100k machine even if it would work.
Useful Excel Tip: When you added a new row and it broke the line of your other motor graphs, you can make Excel ignore the gap in data by putting "=na()" in the unused rows. When Excel sees that in the middle of graph data, it ignores the point.
Goes to show: when you buy a high end motor (or anything else, really), you're not paying just for the motor but also for the support and documentation. Most of us with less patience than you would probably have figured we wasted $100.
I own 2 clear path motors. Even though I am not running them for use in a lathe . The technicians were great at helping me choose a motor and I managed to install my X and Y axis on a system. One thing to note is that the original motors I had were standard servos with 80 in/oz and I replaced them with the SDSK ClearPath ones that had 620 in/oz. During the replacement I found that the older couplers were 1/4 to 1/4 for in and out shafts. I had to get new ones because the ClearPath motors had 3/4" shafts. Now I do not run my motors at 100% due to the larger shafts and higher torque. I didn't want to snap those 1/4" shafts on the ball screw. My upgrade cost me 1600 dollars. Worth every penny as they make no noise like the old ones did with the motor whine and harmonic feedback. and they are precise in the system. Did you ever talk to the people at Tenknic to gain their perspective for your application? I am sure they would love to hear from you and suggest a motor to use. BTW I am using my motors on a Cartesian robot that I modified to work as a mill system.
Your RPM torque testing methodology seems to assume that the motor cuts off immediately when it's out of position range so implicitly the torque measured at a given RPM setting is the torque for that RPM. That's true for the clearpath, depending on the tracking error limit setting. THIS motor looks like it's slowing down to nearly stopped long before it cuts off, so you're measuring the torque at nearly stopped each time (unless I'm missing something) Edit: I think you'd need to test with an external encoder attached to validate the RPM/position is actually tracking the commanded movement
@@bookyle21 sure, ultimately what matters is whether it works adequately for the application. I'm just questioning the empirical measurements for my own point of reference. I've spent the past month evaluating motor options for a project that needs torque, position tracking, and reasonable RPM so these videos have been pretty relevant. This (or maybe it has a big brother?) looked like an interesting option
Yes. The motor actually slows down as load is aplyed and increased and breaks when almost stopped (e.g. for 200rpm at 16:56min). This system is keeping the torque constant by reducing speed until it no longer can maintain that torque. Not usable for position+speed CNC control.
I don't know, the system that you originally show with the automation technologies motor package really is a superior setup. I bought exactly what you recommended and it really does work great. Took nothing to set up and never had to talk to any customer support about anything. I'm only speaking up here because I just wonder if making a recommendation on a motor with only a price difference of 30 or $40 really makes sense. I truly think your first choice was solid for 99% of the people out there, including me. Of course I'm probably a little biased but, there it is....
It was working for me, but it was closer to stalling than I would like. I've never had it stall while actually cutting, but I'd be nervous cutting 8 TPI threads for a chuck adapter plate, for example. It would probably be fine. Probably.
Yeah I'm doing really light threads, 1032 1024, some half inch stuff so I really don't have a problem. It's funny because automation direct has a bigger motor for a little bit more money as well
@@tgirard123 yeah, I tested it in the previous video. It works fine I still like the servo for its high speed performance and silence. This servo isn't awesome, but it's cheap and works fine in this application.
Great video! Well laid out. Your initial challenges were set out in tabs and ready for presentation, and matter-of-factly without frustration. Presentation of the motor was done very well. Excellent job on the video. Content was very valuable as well. Helped me understand the strengths and shortcomings of the motor... and how to get the most out of it... without needing to go through that pain myself. Really appreciate your work on this one and the guidance. Thanks.
This Old Tony did a neat closed stepper refit to his MAHO mill. I started to look at the package he used and it seemed really cool in that no limit switches were needed to rehome. The stepper locked with a brake and position was kept alive in the encoder back half of the stepper even when OFF.
I bought one of these, being careful to mention who sent me. Very fast shipping from Germany indeed. I'm happy to report that it seems to accept 3,3V logic without complaint. Which is nice since, when I started looking into this 1,5 years ago, before my first born kind of got in the way, I decided on using an Arduino Due as the mainboard.
He wasn't lying. I work for a top CNC machine tool manufacturer and we run windows 7 on all of our machines (mostly embedded, though). None of our machines connect to the internet but are able to access shared network folders. It drives IT guys into a tizzy, especially when we tell them that they need to use SMB2 network protocol.
Hi James. A similar servo to yours is the JMC iHSV57 180W servo. It comes on top with Enable I/O, same pricing and better tunability it seems. I am running them on a CNC router. Auto tuning up to 19 without any shrieking and rock solid stiffness. With a bit of work you can also tune them manually, to further reduce position error during run. In your set-up, I would recommend to change the control mode to "0". "4" is not assigned. Also, I would change the microsteps to at least 1000, not 500. The build-in encoder has a natural resolution of 1000 steps/rev. But then, for a leadscrew maybe absolute precision is not so important... In any case, great video
Since it's geared down 6:1 to the leadscrew, 500 steps/rev at the motor already equates to 3000 steps/rev at the screw. Changing it to 1000 at the motor would make it 6000 at the screw, which is needless precision, and also limits the max speed due to the pulse frequency the controller can generate.
I raised the base of the G0752 lathe an additional 1.25 inches above the chip pan. This allowed me to mount the integrated servo motor below the internal gearbox. All the motor connections are easily accessed from the front of the lathe.
New subscriber! Amazing test procedure! The only problem is, I don't understand many of the things you discussed during your testing, LOL. I am always learning something on TH-cam and you're one to look forward to. Thanks for taking the time to share your knowledge.
@@Clough42 In my situation, my time available is limited. Over the years, I have been burned by various 'in my budget' purchases that wasted so much of my time that I was never able to get my planned project done. The 'in my budget' parts took over and became the project. It is a tough decision sometimes, but on this one I feel that the fiddly software/firmware/support along with what appears to be a modestly performing PID control - the money saved would not be worth the frustration. Certainly other people have different goals in mind were these would be a great fit. My experience with ClearPath has been so easy that I don't really even have to think about them. The time saved goes into other areas of the project.
In my opinion (and experience!) the extra cost of Teknic servo system is well worth it: excellent documentation, phone support, and problem-free running for a long, long time.
Great analysis as always Cough42 !!! Another off-topic comment on versions of Windows in industry ... In the Life/Safety world everything has to pass the UL certification and its equivalent in other countries before it can be used. It's a very long process to test all the possible combinations of computers and field devices and could take years. If you own a large apartment complex with hundreds of smoke detectors, complex ventilation and sprinkler controls, and visual and audible notification systems installed in the days of XP it would be very expensive to upgrade. Our company kept laptops with all the various flavors of Windows and its compatible version of the different vendors' software locked in a closet. Our technicians would grab the appropriate one to use to service our customers' systems based on what they were running. When I retired several years ago, we still had a few customers who required us to keep an XP laptop in service.
Wow that was an amazing break down. Your like a mechanical Dr. Thank you for taking the time and sharing your experiences. That motor may save me what $300 switching from the clear path.
@@gyrogearloose1345 Took me three weeks and probably 15 hrs to finally get the setting to populate in the software. I needed to change the step amount. 4000 steps were way to high for an Arduino. I was able to get them down to as low as 100 steps I think. Very fast at that resolution. I had to buy the cable with shipping is a $30 cable. I had access to two different NEW USB-serial cables which wouldn't work. I dug into the issues deep enough I was reprogramming the chip in the USB-serial cable and never could get them to pull the settings from the controller. Also changing them out with various other Null modem cables and serial cables. The $100 price tag is misleading at best. You have to have their cable so call it a $130 dollar motor. Besides adjusting the settings in the controller the motor worked good. Worked as expected with the Arduino, quite, and powerful.
In industry I've come across computers running anything from DOS to Win NT to Win 10 and pretty much everything in between. You have to remember, almost all of these computers are dedicated machines, they are never allowed to connect to the internet, usually not allowed on the company's intranet either, the office computers are totally different and separate. As for your serial port problem, since there are no USB to serial adapters that are UART compliant you are going to run into issues, some will work for one thing but not another, none of them will work for everything. If you can, get away from the USB all together and go with a serial card, most newer laptops support an Express card, a slightly smaller version of the PCMCIA card, if you have a desk top use a PCIe serial card, it will work on any serial connection
Yes. Brand New 2021 Amada Punch presses and Laser Cutters are equipped with Windows XP embedded. Fanuc Servo Drivers Runs on XP. Amada has software that integrates with Fanuc controller to make an easy-to-use interface and graphics.
And a few oddballs too… like the Amiga that was running a school district’s entire HVAC system, or the Commodore 64 that was running a punch press. I personally have in my garage a rather large CNC knee mill that had a control system as big as a refrigerator, and was based on a S100 bus system with an 8080 processor (not 8086, the first IBM PC’s processor, but before that) and used punched paper tape as it’s only storage system. It didn’t even have a CRT screen; just a single line 7-segment LED display for entering G-codes and numeric data. --------- One hopes that machines are not connected to the internet (especially with old software that isn’t supported anymore); but all too often short term convenience wins over security… especially when technically ignorant management gets involved. --------- Sometimes you can get software that’s expecting a classic 16550 UART serial port to work with a USB-to-serial adapter by running it in a virtual machine (VMWare, VirtualBox, Parallels, etc). The VM software catches the accesses to the 16550’s IO port addresses, figures out what the machine control software is trying to send or receive, and sends that to the USB serial adapter.
In spec usb-rs232 devices are available, question is if you're getting one or a clone that isn't. The other issue is lots of devices aren't going to be in spec any more either
@@darkwinter6028 It's not so bad if you connect all your systems via "Sneaker Net". While it won't stop trojans etc, it will severely limit the damage they can do.
@@Agnemons ‘sneaker net’ *is* not connecting the machines - they’re not on the internet. If somebody is really gunning for you though, they’re going to get through - Stuxnet seems to have been delivered to the Iranian nuclear refining plant by a USB thumb drive dropped in the parking lot; from there it managed to hop it’s way into the centrifuges’ PLCs, and subtly mess with them such that they were ruined. Windows had a “feature” for a while called Autorun; where whenever any volume (CD, USB drive, etc) was mounted, it would look for a (usually hidden) specially named .exe and (as the name would indicate) run it. This can be disabled with a registry key setting; this should be done on any machine that runs Windows.
Hi James, first I have to thank you for the great videos, just amazing info and well done! So, after watching your video, i decided to give these servos a try for my new 3D printer, ClearPath are too expensive for just an experiment. As you said, the stiffness wasn't near comparable to ClearPath, so I decided to change some parameters... If you look at the parameter Pr-0.04, (Ratio of inertia) is quite high, and unfortunately is in percentage, mine was 600, i noticed that yours is 800, that is a lot for such a small rotor, so I reduced it to 40, then set the Pr. 0.03(Stiffness of machine to 19, saved, and restarted the system, to my great surprise, the servo is now very stiff, nothing comparable as was out of the factory or with the parameter set to 11 or 12. The motor is silent and relatively warm(35 degrees Celsius), nearly impossible to move by hands, is just rock solid. Running the graphic chart, you will notice that the encoder following error is then reduced to a linear line same as the position line . Just my experience here, I thought I share it here with you. All the best, and keep the good videos coming. Cheers from the UK🍻 .
I had just bought that same motor today and ran across your video as a review. Turns out I have the same lathe (G6020 retrofitted with a VFD) and am looking to make a power-feed for it. So thanks for all the information!
"It's a leadshine driver inside" - but Leadshine has a lot of ranges of servo controllers. I've seen drivers with almost identical names but a complete different look and they also need complete different software for tuning. It seems like some of those are like OEM parts and for those the software is not even on Leadshine's website. Still - Leadshine servo's and drivers are very nice to work with, easy to tune. WARNING: for those who want to order one, make sure you get the genuine StepperOnline BLDC servo motor. I have seen other motors that look exactly the same but I discovered (in time) that these are closed loop stepper motors with integrated driver with the exact same housing. Always check for the title or description to make sure this is a brusheless DC (of BLDC) motor.
James. LOVE your channel, everything you do is so interesting and useful. I am installing this exact same setup on my lathe, I am using a 3:1 ration on the pulley. Exact same servo motor. I'm confused about what I should set my MICROSTEPS and my RESOLUTION to in the code. ??
Hi, I've come across some really ancient Windows versions running the media entertainment on airplanes. Not that risky from hacking I guess but boy did it keep crashing! BobUK.
Over here we running in our industry Win 2000 systems, of course in an isolated network. But we do a big upgrade this year to the latest version, the development of the new software did take almost 2 years.
This motor is also available from a few others on Aliexpress. I found it sold by TopCNC for $83 right now. There are also similar specced ones in a nema 23 and nema 17 with up to 4000rpm.
Yes the companies are not upgrading their systems ever, and barely any receive security updates once they're deployed. At where I used to work some of the CNC equipment used Windows 2000 and Pentium III's. A lot of the newer stuff still used Windows 7 and they're updated yearly only, during the anual service, they pay the fee for long term support for Windows 7. If Windows 2000 machines are still in use, wouldn't surprise me that they're still deploying windows 7 machines.
Windows 2k? Thats cute, laser cutting guys where I cut my metal still using Win3.11, huge machine, 3 computers, 3kW of lasers the whole shebang. Or a wire EDM machine one shop had, I don't even know what its running, looked like something from the PDP/VAX era.
@@VEC7ORlt The laser cutter we had was comparably modern, it has windows 7.Some of the Swiss machines had DOS on them though, those ones are quite old.
People are using old systems because they are reliable. If there was too much problem with them they would have been replaced a loong time ago. And if they were working for the past 20 years, chances are they will do so in the next 20 as well. You just need to stock up on floppy disks / punch cards / stone tablets whatever. :)
@@kistuszek The whole "if it ain't broke don't fix it" only applies to electronic and mechanical things, it falls through with software and firmware because security bugs and vunerabilities do not stop it from working, but they leave the system exposed to exploitation and in some cases sabotage. If these systems are completely offline that's not so bad, it's harder to get malware on them, but some of these machines are connonnected, either directly or indirectly, to the internet and in those situations it's absolutely imperitive they're constantly kept up to date. A blast furnace in Germany was destroyed beyond repair after it got hacked. In that case however the actual weakness was humans for the most part, but the control computers where still network connected to the internet anyways, so these systems can and sometimes are connected to the internet, it's not always possible to air gap, or the air gap is only skin deep. Malware can also be loaded by phsyical media too, if you're that hell bend on causing havock there are easier ways though.
Prolific based USB to serial adapters have been very hit and miss in my experience. FTDI ones are more expensive but have been much more reliable for me.
There are tons of fake FTDI and Prolific chips. Worse yet, FTDI has intentionally released poisoned drivers that will intentionally inject garbage into your circuit if it detects a fake chip. Older versions of FTDI's driver would actually brick the fake chips.
best greetings from Austria. I tested in the last week a JMC Servo 200w on my CNC. I spent more than 10 hours to find the right parameters. The JMC is also not stiff enough. But I am still not satisfied. Nice to see that even a professional does not manage so easily. Best regards Guido
As someone in IT who works with automotive and diesel diagnostic systems plus welding and industrial automation systems, absolutely Windows 7 is still in use a lot and there are plenty of companies who, just before the pandemic hit and I was researching what can be upgraded, still said oh no don't upgrade to Windows 10 our software isn't compatible yet. And I have also experienced issues with certain USB to serial adapters.
I really love how much detail you put into your video, I wish I can wrap my head around all this. I have a metric HLVH that need a easy way to do different pitch Imperial, 2 leads etc and there's no really good alternative out there than I custom solution.
I know this is a little late, but might be interesting: "Microstep: The number of pulses required for one revolution of the motor. For some special users, it can use Pr0.09 (1st numerator of electronic gear) and Pr0.10 (Denominator of electronic Gear) to set required pulses number, they are available when Pr0.08 is set to zero."
I have finally found the guy ; " Clough 42 " that's undoubtedly qualified to build the device , Kelly Johnson always wanted at the early Skunkworks . Kelly insisted upon hydraulic controls for everything & the only thing his engineers could not give him was a Hydraulically Operated Radio ! Clough 42 is eminently qualified to take this project on & I am sure we can build a few more Tomcats with the new radio if this project is started right away
Great video! You have a certain logical/analytical way of going about a problem that makes watching your videos a pleasure. Now that you have a Clearpath servo it would be great to see you compare that to one of DMMs DYN2 combos 😃
"I hope that's not true" I work at Visa. I don't deal with ATMs myself, but talking in the break rooms, I can tell you that supposedly a majority of ATMs use windows xp.
@Lassi Kinnunen 81 jackpotting is still fairly hard as you'd need access to the network, since most ATMs don't allow you to go out of the software with the available controls. also it's windows xp Embedded and is sometimes still updated if there's a serious enough threat
I worked for large British supermarket chain that untill about 5 or 6 years ago still used a DOS application for there back office stock ordering system.
I am a tech and yes it is pretty true. I have had an entire fuel company running on win 98SE in 2010. A local but regional supplier and it was like an act of congress to get them to upgrade at all. These companies buy systems by the pallet and only spend on the most stable and long term at that moment they can get in bulk. It is also sometines the fault of the databases or the software being behind the new OS changes by 5 yrs sometimes.
James, Impressive testing regime and parameters you aimed for. You did good Bud !!!! I might suggest you employ a mechanical hold for the torque line, that will eliminate flex from your fingers of the line slipping through them and allow a hard stop setting mechanically which can be the same for each and every motor tested.....Uniformity in testing is key !!!! Again excellence as is the norm for all that you do. THANKS !!!
It's also worth trying reducing the number of wraps a little, if it's hard to stop the braking torque from running away, and your not having to pull too hard. On traditional dyno brakes (the other extreme) it's usually only half a wrap, and although this requires a lot of tension, it's much more stable.
@@Clough42 they've got some exciting stuff starting to come out in the 3d printing world, high speed stepper based and closed loop steppers for 3d printing
@@Clough42 And welcome to the internet, anyone who buys the good good stuff is desperate to protect it's reputation as the best and the rest of the masses simply consider whatever is the easiest to setup as the best because anything that requires brainpower to make work is obviously junk. If you are willing to experiment and tinker you can really save a lot of money and often get a better product if you are willing to tinker, tune and test these cheaper chinese products.
There are tons of USB/Serial converters out there that use Prolific clones instead of the real chip. Also, newer Prolific drivers check the chips for authenticity before they will initialize. I have worked with steppers and bldc's for many years and prefer bldc's. Great vid. Thanks...
Thank you for the video! A BIG Company in Europe i worked for, is running their robotics on Windows 7 (not even embedded) thats how they got it from the robotics companies- i guess thats a "standard" i was very surprised
Hi James. Can an updated WiringDiagram.JPG be created and uploaded to the Github to support this new motor. Currently there is just the KL-5080H configuration shown. Thanks....
Your small error due to pulling on the string can be removed by adding a simple lever. Make a lever same length as the diameter of your drum. Low friction pivot in the center. Wrap string around drum and attach to both ends of the lever. Attach force gage to end of lever. Create tension in the string by raising the pivot up relative to the center of the drum. Tension in the string due to the lifting of the pivot will be equal and opposite and not show up on the gage. Start the motor and torque will be accurate on the gage. Make your drum hollow and put in some water. Power loss that generates heat will boil the water and keep from burning up your string.
@@Clough42 Yes, that's what we had in the power lab in college. 5hp ac motors with pony brakes. 19 year old kids burning up canvas belts. In the lab near the card punch machines.
@@mattwerner523 nice. And if I remember correctly, full size prony brakes for testing tractors had wood brake shoes with running water over them to keep them cool.
During the stall testing it seemed no matter what the initial chosen rpm, the stepper slowed to the same rpm before stalling. Meaning you recorded the same torque number because it was always at the same rpm number. During the actual lathe work did you try varying the "stiffness" number from the default "11" and did it make any real world difference, beside hum? I have read elsewhere Windows 10 does not like Prolific chips and FTDI chips are preferred. USB drivers can be changed to get the prolific working but next windows 10 update it stops working again.
I wondered why the iHSV57 wasn't in the first test as imho it's one of the best Nema23 motors you can get for that price. These seem to be a copy but the parameters are definitely different. For the iHSV57 i keep the proportional gain on max. As long as all other parameters are set low i can't get it to oscillate and it's not soft like in the video. I can't turn it by hand until it stalls. For tuning i use the open source iHSV-servo-tool (you can find it on Github). I didn't even bother with the other options as i only read bad things about them and don't use a windows system. Some people are arguing that the test is not correct as the motor is slowing down. I don't think that it's a problem. It's just a question of allowable following error which seems to be very large in the default settings. The torque value doesn't really change as the motor starts bogging down. If you set a more realistic error it would stop and go to alarm immediatly but still show the same torque. One advantage is that you don't need a high voltage (60V or 80+V) power supply to get torque at high speeds. I use a 42V (i read that you should leave some headroom for back EMF so no 48V PSU) server power supply which you can get dirt cheap on eBay up to 3kW. I paid 19€ for one. That's much cheaper than buying several of those 60V chinese PSUs. At the moment closed loop Nema34 12Nm Sets cost about the same and are probably a better deal considering their much smaller penalty when not tuned properly. Tuning is a hassle and it's even way worse when you have to deal with bad software and chinese-english parameter descriptions. I wouldn't recommend these for a beginner or someone who wants a plug and play solution. Going down that route gives you the reward of a very quiet machine without the annoying stepper beeping noise though.
Btw the iHSV has an enable input and also an in position output which shows if the commanded position is reached. I'm not sure if they really do what they say as i never tried them but i'm sure at least if the enable didn't work i would've heard about it.
Yeah, I learned about that motor from the comments on the last couple of videos. I think it works be possible to devote an entire channel to testing motors.
I can't fit a NEMA34 where I need this motor. Those motors usually have much higher inductance, so the torque falls off faster, but I haven't looked closely at the curves or run the numbers on effective torque once the reduction ratio is reduced.
@@Clough42 They claim around 5.5mH which doesn't seem too excessive but definitely a high voltage psu is a must. Size was a constraint for me too. Depending on what torque curves to believe in they servos have better performance above around 800rpm (geared 3:1) but in a much smaller package. Also they weren't always that cheap. Maybe they're just clearing their warehouses for the newer generation models.
I just stumbled across your channel and I'm glad I did. You are an excellent teacher, thank you and keep up the great work! I just subbed. I have a set of ClearPath servo's in stock waiting for me to upgrade my homebrew CNC Router. I'll have to check into these motors for work apps. We automate test fixtures in an R&D lab for med devices. We usually just use plain old steppers since we don't need a lot of torque. I like the integrated driver on these units and the price.
I was a bit confused when you said the pully was hot - it's bound to be hot as you're using it as a friction surface to measure the torque - this could affect the results as the heat from that would affect the motor (even though its a very low temperature)
Most industry computers are running Windows 7. I work at an ethanol plant, and we have two networks: one for business and one for plant control. Plant control computers are running Siemens PCS7 on Windows 7. Siemens doesn't even offer software solutions on 10 yet.
I think increasing the damping (by adding the friction of all the bearings, ballscrew, etc to the system) will help it use the higher stiffness setting a lot. Even just putting a couple motor brushes rubbing against the shaft could help the system a lot.
The controller has PID, PI, and P loops for position, velocity, and current. Tuning these would likely solve all of the issues, but with the current state of the manual and software, this is beyond the reach of most hobbyists.
@@Clough42 Have they done anything since then to remedy this in the manual and software? I'm considering buying one of these, but a little hesitant because of these issues
Good stuff...even for a Power Electronics dude (as I'm getting old, I become more interested in Mechanic...weird !). 🤪 I think the driver's programmer did what we commonly do for power devices : give it some overload temporal tolerance...counting on some thermal inertia that any physical system exhibits. Basically, instead of using instantaneous power (torque x speed) to protect the motor, you just keep running a temporal integral of power (energy), discounts some dissipation factor (also estimated)...and compares to a threshold value (based on thermal inertia). That gives overload tolerance.
Most servo systems need to be tuned once connected to the final mechanism. I assume they'll get the software sorted eventually. High-dollar systems like ClearPath do it well.
I figured it would be nice to get a positive comment here and there, so if you want my opinion, I think you're extremely thorough in testing, and honestly, there aren't many people who take as much time, and who you can tell that they are actually passionate about what they're doing, and who are so meticulous in every step and I am that way as well, so I really appreciate it. I have watched a bunch of your videos as I modified my very similar M1099 Lathe, and my MR-1 Gantry mill, and built my high-temp 3-D printer, and your channel has given me some really good insight on these modifications right when I needed it. I appreciate you, and what you're doing is the punchline! Chris...
4:17 that is not necessarily a disadvantage from servo motors. This just indicates either a very low control bandwidth, or excessive integral gain if a PID controller is used. I.e. I think this is mainly due to poor control system tuning/performance. Although maybe since they specifically call it 'stiffness' they are emulating a stiffness on the servo by using impedance control. Although honestly that wouldn't make much sense to implement for a position controller that doesn't interact with humans.
I just gotta say your comments at the end of the video are so true, spot on! It really bothers me that so many people will comment about EVERYTHING they feel you did wrong, whether they're correct or not. However, probably 99.99998% of the ones that leave those comments will never go but all those motors, film thier "Correct" testing, and then put it up for the community! I really commend everyone in the community that does do things like this to make it easier for us COMMONERS and/or NOOBS. Because of a few of my favorite "Machinist TH-camrs" I went and bought a series 60 Monarch Lathe, in fact it's the same one another pretty famous TH-cam machinist has, don't want to name drop here, his channels name is the one that has a play of word from "it's always sunny in Is Philadelphia"... Anyways, thank you for all your help and videos, we very much appreciate it!
I’ve been on the edge of my seat for any and all of the ELS videos 🤗
Some companies should put you on their payroll, since you’re fixing a lot of errors and saving other consumers!!!
Thank you sir for everything you’ve done on the ELS! Not only has it been awesome to watch you develop it in each video, but it’s taught me about steppers and servos!
Hi James. Another masterclass. This time in how to deal with a manufacturer that clearly was in need of some skilled and knowledgeable help. Maybe they had looked your channel up first and were hoping for this level of support. In any event, very nicely done. It's a pleasure watching you wade through the morass! Cheers. BobUK.
Agree with Stepper Online - I've seen some even running windows XP or NT in my industry. Companies paid 10's of thousands for a tool, they won't change it if it works, or worse still, runs a real risk of causing it to no longer work
I really appreciate your testing methodology and the fact that you continue to improve on your equipment recommendations. This is why I feel confident recommending your ELS on all of the home machining forums that I belong to. As a long-time owner of a 9x20 lathe, I am a big proponent of the ELS, mainly to overcome the need for change gears and the back gear mod that is almost a requirement for this class of lathe. I especially appreciate that you continue to improve your recommendations with the goal of providing the best operation at an affordable price. Not all of us hobby machinists can afford the cost of a Clearpath solution so it's nice when someone provides an affordable alternative more in line with my budgetary constraints.
I was impressed with your persistence! I deal with software/hardware issues on a daily basis and something new like that which presented so many up front challenges would have been tossed well before getting to where you got to. Well done to you sir on a what seems like a good find.
Holy crap! Who's got the time for all this malarky? You are a man of endless patience, on your course of investigation and explication.
For the rest of us - go with Teknic Clearpath. It works. It keeps on working. Nothing else to say.
It's overpriced. Leadshine is just as good. You obviously have to tune it and the challenge is getting past the chinglish in the manual.
I commend your patience and persistence. I would have bailed many emails ago. "Are you running Win-7?" Nope, haven't upgraded yet from XP. Man what a journey to get it "working".
"XP4Eva" :)
Not sure how I found this channel, but you sir are a BEAST.... not just your lift abilities either. I don't need any of what you are showing, and I doubt my talent will ever get there, but I appreciate you
Every time i watch your videos i am impressed with your level of accuracy and intelagence, you never cease to amaze me!
I bought a bunch of these motors because you posted this video, thank you.
I can actually provice some insight into the statement, that "most of industry systems are WIN7 currently". We make equipemt that is used in industries like car manufacturing and others (usually also "production line" types of setups). "The industry" in general is incrediably avers to any sort of change, and often doesn't update anything, ever (no matter if there is support available or not), unless something actually doesn't work and needs the update (like the firmware for some device). Now it's important to understand that all systems I've ever seen or heard of are either air gapped, meaning it's physically impossible to connect to them from the outside, or heavily firewalled in multiple stages. Obviously, they can't connect to the outside either. Windows-Updates are either not possible, or have to use a WSUS server (Windows Server Update Services), which basically hosts the files for windows updates locally. Also I do know of at least some companies that do get updates, since you can still basically pay microsoft so you're still getting support for Windows 7, and from what I remember that will be available until ~2024 or something? Likely increasing in cost during that time, but I don't know that.
As some more background, most companies that have these production lines contract other companies to design, build and assemble them. At some point they are delivered and there's an inspection, it's then deemed completed and functional. The company that accepted the "cell" is now responsible for keeping it working, if they update a firmware or install windows updates, that's on them. If you do update something and then something doesn't work, you call the company that made it they just say "well it worked fine with the versions we tested, look you signed off on it here *points to signature*". If a real production line stands still it can eaily cost multiple thousands of dollars PER MINUTE. So having to stop it for any sort of problem introduced by a windows updade can cost real money real fast. Keeping the systems truly isolated from the outside is just easier.
One more Note: my terms may be very far off the "industry norm", my primary language isn't english so I just don't know the correct terminology, but tried to translate as well as I could.
I completely agree. Being a system administrator, updates are the most likely cause of problems on a system. If it's working, do not do the update. This is especially true on system that are not connected to the internet. All the issues James was referring to about companies getting hacked are security flaws that are discovered in old system that ARE connected to the internet.
Great insight, thanks. For myself, when building my CNC rig I specifically chose Win 7 so that I didnt have to deal with updates. End of life? That just means it should be more stable, I'll take it!
If it ain't broken.... 😉
In an automotive plant even the local engineering team doesn't want to deal with their IT department. Better to not connect the automation to the network than deal with IT's restrictions, and glacier like response.
Unfortunately that means some automation has cellular modems as a back door into the equipment for remote service.🤪
I would like to say that you have a better grasp on the english launguage than most people who were born here. ;-)
This does look impressive and gives me confidence to consider this motor for my applications. The ONLY downside is I am not fond of ordering direct from China and hopefully they will see the value in this motor to pickup some stock for it so guys like me are not on the fence about ordering from out of country.
This is the first of your video's I have found. I loved the relaxed vibe and the way you have done and explained your testing. I'm going to go be and watch from the start.
You're in for a surprise. :)
Aaah prolific, that explains everything. I've thrown out all my prolific cables and am only using FTDI based USB to serial cables and I haven't had any weirdness with serial since.
True. Ftdi is the golden standard. Even better to have optoisolated data lines. Really helps with noisy industrial stuff.
@@Trancelebration I mostly use serial for HAM Radio stuff but a cable with optoisolators would be great. Hmm.... Time to go shopping.
Yup. FTDI cables are all that we use for PLC programming and other data functions. I also use them with my Ham gear
Morons, all.
This brings back memories. When I was doing BIOS engineering 9 years ago we always had issues with Prolific USB to serial cables. And we had the same conclusion FTDI controllers were better. I guess nothing has changed.
I know for certain that Win7 is still very much in use in laboratory machines, most control software for medical analysis is unable to run on Win10.
Win7 is certainly not commonplace in industrial applications, they were just blowing smoke. Everyone these days has upgraded to Windows XP.
@Lassi Kinnunen 81 compatibility.
@Lassi Kinnunen 81 not my circus, not my monkeys. We're talking about companies who not only sell you some piece of analytical equipment, but also an underspecced computer to communicate with it - at a hefty markup, no less. They show zero interest in removing technical debt.
@Lassi Kinnunen 81 issue we have is security settings in windows 10 dont allow sufficient access unless running in admin mode. IT wont give us admin mode... we end up having to take the computers controlling our equipment off the network and running windows 7.
at least for win xp, it's much easier to do low level stuff with it.
generally it'll work on 10, but sometimes drivers are written abusing non standard behaviour of an os, or deprecated APIs, and so upgrading might brick the machine, you'll also lose any support, and generally you don't want to mess with a $100k machine even if it would work.
Useful Excel Tip: When you added a new row and it broke the line of your other motor graphs, you can make Excel ignore the gap in data by putting "=na()" in the unused rows. When Excel sees that in the middle of graph data, it ignores the point.
Awesome tip!
@@Clough42 Great job on the video, really cool find on that motor. The Clearpath motors are really nice for CNC stuff, but they are pretty expensive!
@@jgvicke How much would the Clearpath motor tested here cost, do you happen to know?
I'm using Clearpath motors on my CNC machine but this video shows your process in selecting motors and I find it very educational. Thank you.
Goes to show: when you buy a high end motor (or anything else, really), you're not paying just for the motor but also for the support and documentation. Most of us with less patience than you would probably have figured we wasted $100.
Just stopped in for peak, like the setup, sure beats having to swap gears. Patience is a strength, sticking with it paid off. Cheers!
I own 2 clear path motors. Even though I am not running them for use in a lathe . The technicians were great at helping me choose a motor and I managed to install my X and Y axis on a system. One thing to note is that the original motors I had were standard servos with 80 in/oz and I replaced them with the SDSK ClearPath ones that had 620 in/oz. During the replacement I found that the older couplers were 1/4 to 1/4 for in and out shafts. I had to get new ones because the ClearPath motors had 3/4" shafts. Now I do not run my motors at 100% due to the larger shafts and higher torque. I didn't want to snap those 1/4" shafts on the ball screw. My upgrade cost me 1600 dollars. Worth every penny as they make no noise like the old ones did with the motor whine and harmonic feedback. and they are precise in the system. Did you ever talk to the people at Tenknic to gain their perspective for your application? I am sure they would love to hear from you and suggest a motor to use. BTW I am using my motors on a Cartesian robot that I modified to work as a mill system.
Your RPM torque testing methodology seems to assume that the motor cuts off immediately when it's out of position range so implicitly the torque measured at a given RPM setting is the torque for that RPM. That's true for the clearpath, depending on the tracking error limit setting. THIS motor looks like it's slowing down to nearly stopped long before it cuts off, so you're measuring the torque at nearly stopped each time (unless I'm missing something)
Edit: I think you'd need to test with an external encoder attached to validate the RPM/position is actually tracking the commanded movement
Seems like a valid concern for other applications like a CNC conversion. But would probably be ok for the ELS project
This is an excellent observation and I was thinking the same too. It appears the torque being measured is at stall.
@@bookyle21 sure, ultimately what matters is whether it works adequately for the application.
I'm just questioning the empirical measurements for my own point of reference. I've spent the past month evaluating motor options for a project that needs torque, position tracking, and reasonable RPM so these videos have been pretty relevant. This (or maybe it has a big brother?) looked like an interesting option
Yes. The motor actually slows down as load is aplyed and increased and breaks when almost stopped (e.g. for 200rpm at 16:56min). This system is keeping the torque constant by reducing speed until it no longer can maintain that torque. Not usable for position+speed CNC control.
Yeah, I think this is right. For this application, it's good enough.
Thank you so much for your generosity. I’ve ordered your boards etc. and will attempt to set up an ELS on my lathe.
I don't know, the system that you originally show with the automation technologies motor package really is a superior setup. I bought exactly what you recommended and it really does work great. Took nothing to set up and never had to talk to any customer support about anything. I'm only speaking up here because I just wonder if making a recommendation on a motor with only a price difference of 30 or $40 really makes sense. I truly think your first choice was solid for 99% of the people out there, including me. Of course I'm probably a little biased but, there it is....
It was working for me, but it was closer to stalling than I would like. I've never had it stall while actually cutting, but I'd be nervous cutting 8 TPI threads for a chuck adapter plate, for example. It would probably be fine. Probably.
Yeah I'm doing really light threads, 1032 1024, some half inch stuff so I really don't have a problem. It's funny because automation direct has a bigger motor for a little bit more money as well
@@tgirard123 yeah, I tested it in the previous video. It works fine I still like the servo for its high speed performance and silence. This servo isn't awesome, but it's cheap and works fine in this application.
Great video! Well laid out. Your initial challenges were set out in tabs and ready for presentation, and matter-of-factly without frustration. Presentation of the motor was done very well. Excellent job on the video. Content was very valuable as well. Helped me understand the strengths and shortcomings of the motor... and how to get the most out of it... without needing to go through that pain myself. Really appreciate your work on this one and the guidance. Thanks.
Great video! Seeing the entire process of setting everything up and how the motor performs in actual cut is very helpful
This Old Tony did a neat closed stepper refit to his MAHO mill. I started to look at the package he used and it seemed really cool in that no limit switches were needed to rehome. The stepper locked with a brake and position was kept alive in the encoder back half of the stepper even when OFF.
Yeah, but the servos were donated to TOT and cost well over $1000, closer to 2k. If ya got it, spend it.
Those weren't steppers and a 1M sub YT channel can probably afford a few servos
I bought one of these, being careful to mention who sent me. Very fast shipping from Germany indeed.
I'm happy to report that it seems to accept 3,3V logic without complaint. Which is nice since, when I started looking into this 1,5 years ago, before my first born kind of got in the way, I decided on using an Arduino Due as the mainboard.
He wasn't lying. I work for a top CNC machine tool manufacturer and we run windows 7 on all of our machines (mostly embedded, though). None of our machines connect to the internet but are able to access shared network folders. It drives IT guys into a tizzy, especially when we tell them that they need to use SMB2 network protocol.
Hi James.
A similar servo to yours is the JMC iHSV57 180W servo. It comes on top with Enable I/O, same pricing and better tunability it seems. I am running them on a CNC router.
Auto tuning up to 19 without any shrieking and rock solid stiffness. With a bit of work you can also tune them manually, to further reduce position error during run.
In your set-up, I would recommend to change the control mode to "0". "4" is not assigned. Also, I would change the microsteps to at least 1000, not 500.
The build-in encoder has a natural resolution of 1000 steps/rev. But then, for a leadscrew maybe absolute precision is not so important...
In any case, great video
Since it's geared down 6:1 to the leadscrew, 500 steps/rev at the motor already equates to 3000 steps/rev at the screw. Changing it to 1000 at the motor would make it 6000 at the screw, which is needless precision, and also limits the max speed due to the pulse frequency the controller can generate.
I raised the base of the G0752 lathe an additional 1.25 inches above the chip pan. This allowed me to mount the integrated servo motor below the internal gearbox. All the motor connections are easily accessed from the front of the lathe.
New subscriber! Amazing test procedure! The only problem is, I don't understand many of the things you discussed during your testing, LOL. I am always learning something on TH-cam and you're one to look forward to. Thanks for taking the time to share your knowledge.
Thanks! I now know why I love my ClearPath motors so much. There is zero chance that I will even try these.
If they're in you budget, they're great motors.
@@Clough42 In my situation, my time available is limited. Over the years, I have been burned by various 'in my budget' purchases that wasted so much of my time that I was never able to get my planned project done. The 'in my budget' parts took over and became the project.
It is a tough decision sometimes, but on this one I feel that the fiddly software/firmware/support along with what appears to be a modestly performing PID control - the money saved would not be worth the frustration.
Certainly other people have different goals in mind were these would be a great fit. My experience with ClearPath has been so easy that I don't really even have to think about them. The time saved goes into other areas of the project.
I've been using teknic servos for a lot of my automation jobs MC and SD. when you get over the price they can't be beat.
In my opinion (and experience!) the extra cost of Teknic servo system is well worth it: excellent documentation, phone support, and problem-free running for a long, long time.
Dude is a flipping genius. GENIUS I TELL YOU!!
Great analysis as always Cough42 !!!
Another off-topic comment on versions of Windows in industry ... In the Life/Safety world everything has to pass the UL certification and its equivalent in other countries before it can be used. It's a very long process to test all the possible combinations of computers and field devices and could take years.
If you own a large apartment complex with hundreds of smoke detectors, complex ventilation and sprinkler controls, and visual and audible notification systems installed in the days of XP it would be very expensive to upgrade.
Our company kept laptops with all the various flavors of Windows and its compatible version of the different vendors' software locked in a closet. Our technicians would grab the appropriate one to use to service our customers' systems based on what they were running.
When I retired several years ago, we still had a few customers who required us to keep an XP laptop in service.
Thanks again for your dedication that you put into your testing and and your videos.
Wow that was an amazing break down. Your like a mechanical Dr. Thank you for taking the time and sharing your experiences. That motor may save me what $300 switching from the clear path.
I'm interested to learn how you got on with Leadshine. Any comment?
@@gyrogearloose1345 Took me three weeks and probably 15 hrs to finally get the setting to populate in the software. I needed to change the step amount. 4000 steps were way to high for an Arduino. I was able to get them down to as low as 100 steps I think. Very fast at that resolution. I had to buy the cable with shipping is a $30 cable. I had access to two different NEW USB-serial cables which wouldn't work. I dug into the issues deep enough I was reprogramming the chip in the USB-serial cable and never could get them to pull the settings from the controller. Also changing them out with various other Null modem cables and serial cables. The $100 price tag is misleading at best. You have to have their cable so call it a $130 dollar motor. Besides adjusting the settings in the controller the motor worked good. Worked as expected with the Arduino, quite, and powerful.
@@robgib3100 OK! Fingers crossed it holds up.
In industry I've come across computers running anything from DOS to Win NT to Win 10 and pretty much everything in between. You have to remember, almost all of these computers are dedicated machines, they are never allowed to connect to the internet, usually not allowed on the company's intranet either, the office computers are totally different and separate. As for your serial port problem, since there are no USB to serial adapters that are UART compliant you are going to run into issues, some will work for one thing but not another, none of them will work for everything. If you can, get away from the USB all together and go with a serial card, most newer laptops support an Express card, a slightly smaller version of the PCMCIA card, if you have a desk top use a PCIe serial card, it will work on any serial connection
Yes. Brand New 2021 Amada Punch presses and Laser Cutters are equipped with Windows XP embedded. Fanuc Servo Drivers Runs on XP. Amada has software that integrates with Fanuc controller to make an easy-to-use interface and graphics.
And a few oddballs too… like the Amiga that was running a school district’s entire HVAC system, or the Commodore 64 that was running a punch press. I personally have in my garage a rather large CNC knee mill that had a control system as big as a refrigerator, and was based on a S100 bus system with an 8080 processor (not 8086, the first IBM PC’s processor, but before that) and used punched paper tape as it’s only storage system. It didn’t even have a CRT screen; just a single line 7-segment LED display for entering G-codes and numeric data.
---------
One hopes that machines are not connected to the internet (especially with old software that isn’t supported anymore); but all too often short term convenience wins over security… especially when technically ignorant management gets involved.
---------
Sometimes you can get software that’s expecting a classic 16550 UART serial port to work with a USB-to-serial adapter by running it in a virtual machine (VMWare, VirtualBox, Parallels, etc). The VM software catches the accesses to the 16550’s IO port addresses, figures out what the machine control software is trying to send or receive, and sends that to the USB serial adapter.
In spec usb-rs232 devices are available, question is if you're getting one or a clone that isn't. The other issue is lots of devices aren't going to be in spec any more either
@@darkwinter6028 It's not so bad if you connect all your systems via "Sneaker Net". While it won't stop trojans etc, it will severely limit the damage they can do.
@@Agnemons ‘sneaker net’ *is* not connecting the machines - they’re not on the internet.
If somebody is really gunning for you though, they’re going to get through - Stuxnet seems to have been delivered to the Iranian nuclear refining plant by a USB thumb drive dropped in the parking lot; from there it managed to hop it’s way into the centrifuges’ PLCs, and subtly mess with them such that they were ruined.
Windows had a “feature” for a while called Autorun; where whenever any volume (CD, USB drive, etc) was mounted, it would look for a (usually hidden) specially named .exe and (as the name would indicate) run it. This can be disabled with a registry key setting; this should be done on any machine that runs Windows.
Its cool they sent you one.... Says a lot about both parties
What a crazy deep test you made !
I'm glad you stuck with it. Looks like a great motor system if the specs fit your needs.
Hi James, first I have to thank you for the great videos, just amazing info and well done!
So, after watching your video, i decided to give these servos a try for my new 3D printer, ClearPath are too expensive for just an experiment.
As you said, the stiffness wasn't near comparable to ClearPath, so I decided to change some parameters...
If you look at the parameter Pr-0.04, (Ratio of inertia) is quite high, and unfortunately is in percentage, mine was 600, i noticed that yours is 800, that is a lot for such a small rotor, so I reduced it to 40, then set the Pr. 0.03(Stiffness of machine to 19, saved, and restarted the system, to my great surprise, the servo is now very stiff, nothing comparable as was out of the factory or with the parameter set to 11 or 12. The motor is silent and relatively warm(35 degrees Celsius), nearly impossible to move by hands, is just rock solid.
Running the graphic chart, you will notice that the encoder following error is then reduced to a linear line same as the position line .
Just my experience here, I thought I share it here with you.
All the best, and keep the good videos coming.
Cheers from the UK🍻 .
Awesome video and props to the company for sending it in for review.
I had just bought that same motor today and ran across your video as a review. Turns out I have the same lathe (G6020 retrofitted with a VFD) and am looking to make a power-feed for it. So thanks for all the information!
Cool. I hope it helps.
"It's a leadshine driver inside" - but Leadshine has a lot of ranges of servo controllers. I've seen drivers with almost identical names but a complete different look and they also need complete different software for tuning. It seems like some of those are like OEM parts and for those the software is not even on Leadshine's website.
Still - Leadshine servo's and drivers are very nice to work with, easy to tune.
WARNING: for those who want to order one, make sure you get the genuine StepperOnline BLDC servo motor. I have seen other motors that look exactly the same but I discovered (in time) that these are closed loop stepper motors with integrated driver with the exact same housing. Always check for the title or description to make sure this is a brusheless DC (of BLDC) motor.
19:45 the pulley is hot from the friction between the string and the pulley. That is where most of the energy pushed into the motor ended up.
James. LOVE your channel, everything you do is so interesting and useful. I am installing this exact same setup on my lathe, I am using a 3:1 ration on the pulley. Exact same servo motor. I'm confused about what I should set my MICROSTEPS and my RESOLUTION to in the code. ??
Clough, Great video. Sorry about my comments at the Summer Bash. Take care. Bill
Wow! Nice! I may have to do this
Thanks for your work, and good engineering.
J. Hicks
I wish I had your electronics know how but still fun to watch.
90% of our cnc machines are Win7, the others such as our nesting and plasma table are WinXP.
Hi, I've come across some really ancient Windows versions running the media entertainment on airplanes. Not that risky from hacking I guess but boy did it keep crashing! BobUK.
True even the atms from my bank run on Windows xp,
I run my plasma table with Win10. 64 bit. it came with win7 32bit.
I upgraded from W10 to XP ( and I do mean UPGRADED ).
@zomgthisisawesomelol He's dead, Jim.
Over here we running in our industry Win 2000 systems, of course in an isolated network. But we do a big upgrade this year to the latest version, the development of the new software did take almost 2 years.
This motor is also available from a few others on Aliexpress. I found it sold by TopCNC for $83 right now. There are also similar specced ones in a nema 23 and nema 17 with up to 4000rpm.
Yes the companies are not upgrading their systems ever, and barely any receive security updates once they're deployed. At where I used to work some of the CNC equipment used Windows 2000 and Pentium III's. A lot of the newer stuff still used Windows 7 and they're updated yearly only, during the anual service, they pay the fee for long term support for Windows 7. If Windows 2000 machines are still in use, wouldn't surprise me that they're still deploying windows 7 machines.
Windows 2k? Thats cute, laser cutting guys where I cut my metal still using Win3.11, huge machine, 3 computers, 3kW of lasers the whole shebang.
Or a wire EDM machine one shop had, I don't even know what its running, looked like something from the PDP/VAX era.
@@VEC7ORlt The laser cutter we had was comparably modern, it has windows 7.Some of the Swiss machines had DOS on them though, those ones are quite old.
People are using old systems because they are reliable. If there was too much problem with them they would have been replaced a loong time ago. And if they were working for the past 20 years, chances are they will do so in the next 20 as well. You just need to stock up on floppy disks / punch cards / stone tablets whatever. :)
@@kistuszek The whole "if it ain't broke don't fix it" only applies to electronic and mechanical things, it falls through with software and firmware because security bugs and vunerabilities do not stop it from working, but they leave the system exposed to exploitation and in some cases sabotage. If these systems are completely offline that's not so bad, it's harder to get malware on them, but some of these machines are connonnected, either directly or indirectly, to the internet and in those situations it's absolutely imperitive they're constantly kept up to date. A blast furnace in Germany was destroyed beyond repair after it got hacked. In that case however the actual weakness was humans for the most part, but the control computers where still network connected to the internet anyways, so these systems can and sometimes are connected to the internet, it's not always possible to air gap, or the air gap is only skin deep. Malware can also be loaded by phsyical media too, if you're that hell bend on causing havock there are easier ways though.
Prolific based USB to serial adapters have been very hit and miss in my experience. FTDI ones are more expensive but have been much more reliable for me.
There are tons of fake FTDI and Prolific chips. Worse yet, FTDI has intentionally released poisoned drivers that will intentionally inject garbage into your circuit if it detects a fake chip.
Older versions of FTDI's driver would actually brick the fake chips.
@@joels7605 CP210x are the best
@@joels7605 Interesting. Used tons of chinese ft232rl chips. Probably fake. Never had a problem.
best greetings from Austria. I tested in the last week a JMC Servo 200w on my CNC. I spent more than 10 hours to find the right parameters. The JMC is also not stiff enough. But I am still not satisfied. Nice to see that even a professional does not manage so easily.
Best regards Guido
Dude, You are the Man! Thanks for sharing your genius with us.
This was a very impressive, informative and highly enjoyable presentation! Thank you from your new subscriber, Rolf.
As someone in IT who works with automotive and diesel diagnostic systems plus welding and industrial automation systems, absolutely Windows 7 is still in use a lot and there are plenty of companies who, just before the pandemic hit and I was researching what can be upgraded, still said oh no don't upgrade to Windows 10 our software isn't compatible yet. And I have also experienced issues with certain USB to serial adapters.
What a wealth of information. Very nice
Thanks James. You have a remarkably broad education.
Again a really really good video! Awesome to really know about this motor!
I really love how much detail you put into your video, I wish I can wrap my head around all this. I have a metric HLVH that need a easy way to do different pitch Imperial, 2 leads etc and there's no really good alternative out there than I custom solution.
I watched this with a pot of Black Rifle Coffee. A great way to spend a Sunday morning! Totally thorough!!
I know this is a little late, but might be interesting:
"Microstep: The number of pulses required for one revolution of the motor. For some special users, it can use Pr0.09 (1st numerator of electronic gear) and Pr0.10 (Denominator of electronic Gear) to set required pulses number, they are available when Pr0.08 is set to zero."
My small CNC setup has been running XP for about a decade now. No internet, no problems. My CAD machine runs win10.
Thanks for another fantastic episode!! Well done!
I have finally found the guy ; " Clough 42 " that's undoubtedly qualified to build the device , Kelly Johnson always wanted at the early Skunkworks . Kelly insisted upon hydraulic controls for everything & the only thing his engineers could not give him was a Hydraulically Operated Radio ! Clough 42 is eminently qualified to take this project on & I am sure we can build a few more Tomcats with the new radio if this project is started right away
Great video!
You have a certain logical/analytical way of going about a problem that makes watching your videos a pleasure.
Now that you have a Clearpath servo it would be great to see you compare that to one of DMMs DYN2 combos 😃
Very interesting. Great video!
"I hope that's not true"
I work at Visa. I don't deal with ATMs myself, but talking in the break rooms, I can tell you that supposedly a majority of ATMs use windows xp.
about 2 years ago, i found a atm of a german bank running Win98 😨.
I must admit, it was quiet tempting 🤣
@Lassi Kinnunen 81 jackpotting is still fairly hard as you'd need access to the network, since most ATMs don't allow you to go out of the software with the available controls.
also it's windows xp Embedded and is sometimes still updated if there's a serious enough threat
they run ATMs at win XP because 7 and 10 have so many CIA backdoors and are more unsafe than xp sp3
Very good review, good solid information and real life testing supporting the preliminary tests! Awesome work!
I worked for large British supermarket chain that untill about 5 or 6 years ago still used a DOS application for there back office stock ordering system.
I am a tech and yes it is pretty true. I have had an entire fuel company running on win 98SE in 2010. A local but regional supplier and it was like an act of congress to get them to upgrade at all. These companies buy systems by the pallet and only spend on the most stable and long term at that moment they can get in bulk. It is also sometines the fault of the databases or the software being behind the new OS changes by 5 yrs sometimes.
James,
Impressive testing regime and parameters you aimed for. You did good Bud !!!! I might suggest you employ a mechanical hold for the torque line, that will eliminate flex from your fingers of the line slipping through them and allow a hard stop setting mechanically which can be the same for each and every motor tested.....Uniformity in testing is key !!!! Again excellence as is the norm for all that you do. THANKS !!!
It's also worth trying reducing the number of wraps a little, if it's hard to stop the braking torque from running away, and your not having to pull too hard. On traditional dyno brakes (the other extreme) it's usually only half a wrap, and although this requires a lot of tension, it's much more stable.
I've personally been using stepperonline steppers for 3d printing for the last 2 years, gotta say they're one of the better stepper vendors out there.
I've been told over and over that they're junk, but they've been running just fine in my CNC mill for years.
@@Clough42 they've got some exciting stuff starting to come out in the 3d printing world, high speed stepper based and closed loop steppers for 3d printing
@@Clough42 And welcome to the internet, anyone who buys the good good stuff is desperate to protect it's reputation as the best and the rest of the masses simply consider whatever is the easiest to setup as the best because anything that requires brainpower to make work is obviously junk. If you are willing to experiment and tinker you can really save a lot of money and often get a better product if you are willing to tinker, tune and test these cheaper chinese products.
The hardware seems like a great deal, if only the software and support could match.
There are tons of USB/Serial converters out there that use Prolific clones instead of the real chip. Also, newer Prolific drivers check the chips for authenticity before they will initialize.
I have worked with steppers and bldc's for many years and prefer bldc's.
Great vid. Thanks...
This one initializes and works great with other devices. I'm pretty certain it dates from after the FTDI wars.
Thank you for the video!
A BIG Company in Europe i worked for, is running their robotics on Windows 7 (not even embedded) thats how they got it from the robotics companies- i guess thats a "standard" i was very surprised
Hi James. Can an updated WiringDiagram.JPG be created and uploaded to the Github to support this new motor. Currently there is just the KL-5080H configuration shown. Thanks....
I always enjoy your presentations. Thanks for sharing.
Fantastic testing and analysis!
Your small error due to pulling on the string can be removed by adding a simple lever. Make a lever same length as the diameter of your drum. Low friction pivot in the center. Wrap string around drum and attach to both ends of the lever. Attach force gage to end of lever.
Create tension in the string by raising the pivot up relative to the center of the drum. Tension in the string due to the lifting of the pivot will be equal and opposite and not show up on the gage. Start the motor and torque will be accurate on the gage.
Make your drum hollow and put in some water. Power loss that generates heat will boil the water and keep from burning up your string.
What you're describing is essentially the way a real prony brake works.
@@Clough42 Yes, that's what we had in the power lab in college. 5hp ac motors with pony brakes. 19 year old kids burning up canvas belts. In the lab near the card punch machines.
@@mattwerner523 nice. And if I remember correctly, full size prony brakes for testing tractors had wood brake shoes with running water over them to keep them cool.
During the stall testing it seemed no matter what the initial chosen rpm, the stepper slowed to the same rpm before stalling. Meaning you recorded the same torque number because it was always at the same rpm number.
During the actual lathe work did you try varying the "stiffness" number from the default "11" and did it make any real world difference, beside hum?
I have read elsewhere Windows 10 does not like Prolific chips and FTDI chips are preferred. USB drivers can be changed to get the prolific working but next windows 10 update it stops working again.
I totally agree. I think the results are not comparable.
I was real sceptical. Glad you carried through with it and made this finding!
I wondered why the iHSV57 wasn't in the first test as imho it's one of the best Nema23 motors you can get for that price. These seem to be a copy but the parameters are definitely different. For the iHSV57 i keep the proportional gain on max. As long as all other parameters are set low i can't get it to oscillate and it's not soft like in the video. I can't turn it by hand until it stalls. For tuning i use the open source iHSV-servo-tool (you can find it on Github). I didn't even bother with the other options as i only read bad things about them and don't use a windows system.
Some people are arguing that the test is not correct as the motor is slowing down. I don't think that it's a problem. It's just a question of allowable following error which seems to be very large in the default settings. The torque value doesn't really change as the motor starts bogging down. If you set a more realistic error it would stop and go to alarm immediatly but still show the same torque.
One advantage is that you don't need a high voltage (60V or 80+V) power supply to get torque at high speeds. I use a 42V (i read that you should leave some headroom for back EMF so no 48V PSU) server power supply which you can get dirt cheap on eBay up to 3kW. I paid 19€ for one. That's much cheaper than buying several of those 60V chinese PSUs.
At the moment closed loop Nema34 12Nm Sets cost about the same and are probably a better deal considering their much smaller penalty when not tuned properly. Tuning is a hassle and it's even way worse when you have to deal with bad software and chinese-english parameter descriptions. I wouldn't recommend these for a beginner or someone who wants a plug and play solution. Going down that route gives you the reward of a very quiet machine without the annoying stepper beeping noise though.
Btw the iHSV has an enable input and also an in position output which shows if the commanded position is reached. I'm not sure if they really do what they say as i never tried them but i'm sure at least if the enable didn't work i would've heard about it.
Yeah, I learned about that motor from the comments on the last couple of videos. I think it works be possible to devote an entire channel to testing motors.
I can't fit a NEMA34 where I need this motor. Those motors usually have much higher inductance, so the torque falls off faster, but I haven't looked closely at the curves or run the numbers on effective torque once the reduction ratio is reduced.
@@Clough42 They claim around 5.5mH which doesn't seem too excessive but definitely a high voltage psu is a must. Size was a constraint for me too. Depending on what torque curves to believe in they servos have better performance above around 800rpm (geared 3:1) but in a much smaller package. Also they weren't always that cheap. Maybe they're just clearing their warehouses for the newer generation models.
Hah Called it! you were trying to test one! Excellent, don't know if you caught the other comment suggesting these but nice one. Feeling a bit smug.
Very interesting video. It was nice meeting you in person at the Bash.
I just stumbled across your channel and I'm glad I did. You are an excellent teacher, thank you and keep up the great work! I just subbed. I have a set of ClearPath servo's in stock waiting for me to upgrade my homebrew CNC Router. I'll have to check into these motors for work apps. We automate test fixtures in an R&D lab for med devices. We usually just use plain old steppers since we don't need a lot of torque. I like the integrated driver on these units and the price.
I was a bit confused when you said the pully was hot - it's bound to be hot as you're using it as a friction surface to measure the torque - this could affect the results as the heat from that would affect the motor (even though its a very low temperature)
That motor looks appetizing James. Wow did not expect a servo like this for 100$. Nice job on the testing.
Most industry computers are running Windows 7. I work at an ethanol plant, and we have two networks: one for business and one for plant control. Plant control computers are running Siemens PCS7 on Windows 7. Siemens doesn't even offer software solutions on 10 yet.
Win 7 would be a blessing for some companies. I have worked at companies with machines using Win 3.11 on their industrial laser cutters.
I think increasing the damping (by adding the friction of all the bearings, ballscrew, etc to the system) will help it use the higher stiffness setting a lot.
Even just putting a couple motor brushes rubbing against the shaft could help the system a lot.
The controller has PID, PI, and P loops for position, velocity, and current. Tuning these would likely solve all of the issues, but with the current state of the manual and software, this is beyond the reach of most hobbyists.
@@Clough42 Have they done anything since then to remedy this in the manual and software? I'm considering buying one of these, but a little hesitant because of these issues
I would go with the Ethernet motors and controllers that this old Tony used. Worked flawlessly.
Or save your money and go with the RS series drivers instead of the Ethercat drivers so you can do step+dir/analog input.
Good stuff...even for a Power Electronics dude (as I'm getting old, I become more interested in Mechanic...weird !). 🤪
I think the driver's programmer did what we commonly do for power devices : give it some overload temporal tolerance...counting on some thermal inertia that any physical system exhibits.
Basically, instead of using instantaneous power (torque x speed) to protect the motor, you just keep running a temporal integral of power (energy), discounts some dissipation factor (also estimated)...and compares to a threshold value (based on thermal inertia). That gives overload tolerance.
Most servo systems need to be tuned once connected to the final mechanism. I assume they'll get the software sorted eventually. High-dollar systems like ClearPath do it well.
I figured it would be nice to get a positive comment here and there, so if you want my opinion, I think you're extremely thorough in testing, and honestly, there aren't many people who take as much time, and who you can tell that they are actually passionate about what they're doing, and who are so meticulous in every step and I am that way as well, so I really appreciate it. I have watched a bunch of your videos as I modified my very similar M1099 Lathe, and my MR-1 Gantry mill, and built my high-temp 3-D printer, and your channel has given me some really good insight on these modifications right when I needed it. I appreciate you, and what you're doing is the punchline! Chris...
I used nanotec integrated motor driver nema 34 for a few years. Genius.
4:17 that is not necessarily a disadvantage from servo motors. This just indicates either a very low control bandwidth, or excessive integral gain if a PID controller is used. I.e. I think this is mainly due to poor control system tuning/performance.
Although maybe since they specifically call it 'stiffness' they are emulating a stiffness on the servo by using impedance control. Although honestly that wouldn't make much sense to implement for a position controller that doesn't interact with humans.