Hey! Cool video! I'm El Dr. Gusman, the next version that I'm preparing is going to have a ton of changes, all the Sigrok protocol decoders, better rendering with more samples in screen and support for the pico 2 (and the pico 2 will have three times the ammount of samples and probably will capture up to 200Mhz).
The man himself! Thanks for your work and leaving everything open like this. I'm looking forward to learning more about the protocols. More samples and 200MHz? Sounds great. That's the upper limit of the level shifter chips right?
@@HappyLittleDiodes Yes, that's the shifter's limit, and way over the limit of the pico, to get 200Mhz I have to overclock it to 400Mhz, with the original pico I did not got it stable, but with the pico 2 and a little heatsink seems to work without hikcups :=)
You don't need that much solder paste. Use a toothpick to apply it. When you drag the iron across the pins, add flux over the pins so that the solder balls up more and keeps the bridges from forming. Also put a bit of flux on the solder wick before you use it, it will help the solder be drawn into the wick.
In addition to what other people say about soldering these, I also have a tip: 1. When reworking solder bridges on surface mount chips like these, drag your soldering iron parallel to the pin itself, meaning outwards from the chip and not along its edge. That'll motivate the solder to stay in its lane (on its pad) by taking advantage of cohesive properties (tendency of solder to stick to itself more than other things). Dragging from pin to pin along the edge of the chip is good when you have tons of flux and no excess of solder, as there's less solder on your iron's tip that can bridge the pins. 2. When soldering big chips, it's useful to swirl your hot air gun from further away first to get the flux in the solder paste molten out of the paste and preheat the PCB. That'll both help with the solder adhering by removing oxides, *and* melting the solder more readily and instantly when you do get it hot enough to melt. That's good practice, and similar to reflow ovens where the actual melting stage is very short compared to preheat, soak (flux stage), and the cooldown (controlled solidification. The preheat step is often where a preheating plate comes in, but for small PCB's like these, the hot plate is not worth the space it takes up or the money it costs unless you just really like collecting the tools for it. 3. I think Salae is pronounced "Say-Lay" or "Suh-Lay", similar to the French word for sun, "soleil". It's just like GIF. No one knows how it's pronounced unless a company representative corrects you, and even then it's up for debate!
Nice video about the alternatives to commercial analyser solutions. When it comes to soldering, there are a few things that would help: - Being conservative with the quantity of solder paste, it's easier to add solder to individual pins (with very thin solder). - Wicking will nearly always remove too much - When drag soldering it is better to drag "along the pin", from the inside out for an IC pulling the solder to the outside extremity of the pad rather than from pin to pin otherwise this will affect a whole group of pins (and require wicking). - When making a really great number of boards, a stencil (stainless steel or even plastic) becomes a valid option.
Happy that YT saw fit to put this into my feed today. I've been shopping for a reasonably priced analyzer (sorry us spelling is drilled in hard at this point) so this is some good data. Definitely worth the sub
If you want to use a syringe to apply the solder paste, get some of the plastic, conical dispensing tips. They take less force to squeeze the paste through and you can use a smaller diameter tip. A pneumatic solder paste dispenser will make things even easier. They apply a fairly consistent amount of paste with the press of a button.
@@rocketman221projects and in my view little worry of paste going off or needing to be stored in the fridge. I’ve had a tube in my drawer for a year now and it’s still fine.
Conical tips are a bit of a pain for SMD soldering. With a flat-edged tip like a beveled or chisel tip, solder bridges mostly get sucked onto the iron automatically so you only need to use wick in extreme cases.
Also, for hand soldering, get a microscope. You can get a cheap-ish one from Amscope at ~500$. Cobine with a decent 0.3mm chisel tip and 0402 or even 0201 are manageable.
The Saleae option is expensive, it's true. But it has 16 analogue inputs ! And no limit on capture depth ! I only have an 8 bit one but it's VERY good. However, for wider busses like that z80 on the title page, you might want a few bits more. You probably want at least 14 address bits, 8 data bits, and a few clocks or status bits such as R/W and /ALE. It would be great if you could capture just a few more bits but another option is to get a 1990s analyser like an HP. For much less money than the Saleae you can have HUNDREDS of inputs - even the smallest ones have 34 or 68. And you get 'state' capture where only the actual bus transaction is captured rather than a sample every 10ns. This makes far better use of the capture memory. The loss is mostly bench space :).
Yeah, mostly with the Salea stuff you're paying for the development costs of their (very nice) software suite. But yeah, it's not really affordable for your average hobbyist
I've got an HP 16500B from the mid-late 90s, and yeah having it disassemble ("inverse assemble" in HP speak) a Z80 in real-time was absolutely magical.
The no limit on capture depth only works because it can stream the data to the PC fast enough. Once the sample rate or number of channels gets high enough then streaming the data can become a problem.
@@SomeMorganSomewhere I bought the Saleae after having a hard time with sigrok with several devices. They have hobbyist discounts and small business discounts. Also they will pay you to write articles or produce videos with their product. The software is great and runs cross platform as well. This project looks pretty cool though. The higher end Saleae are really expensive in comparison.
3 หลายเดือนก่อน +2
Hello there! I know everything about this project from the very beginning because Dr Gusman is a friend of mine (and I see him in comments hehe) and I'm very proud of the way you cover it. Thanks and congratullations for this video. Already suscribed.
I used a Logic analyzer back in the day in college and it was one of those things you thought would only ever be affordable by schools and businesses. What a world we live in!
Interface works very well. Smart, simple, easy and free. Set up into a rp2040 of two bucks (maybe less at chinese webstores sales). Up to upgrade when rp2 bugs fixed and improving. I2c protocols and SPI performerd like a charm. Nothing to say but thank you.
I was surprised you didn't use your new solder paste for the headers! :) Putting the trace in a spreadsheet you can relatively easily convert the binary outputs to HEX for example which can be useful when you are trying to decode what is happening. One of these days I need to write a perl script for this and dissemble the output directly. Great video thank you.
I need a logic analyzer, and this looks like a fun project! Enjoyed the video :) With SMD soldering, I usually just apply solder to one side of discretes and a few pins of ICs, use the hot air station to tack them all on, and then hand solder the rest (drag solder the ICs). You can wipe your solder tip off really well and apply additional flux to avoid using wick, too, since the flux will wick extra solder to the clean tip. Components with hidden pins obviously need a different treatment. I will use paste when I have a stencil, of course.
It's really good to see someone doing SMD soldering who isn't obviously already an expert, because it makes me feel like it's actually something I could do myself. Often I just sit and watch in awe and think, "well, ... leave that to someone with an iron will and nerves of steel!" It's also good to see the advice like that of @profdc9501 in the comments.
Drag the iron tip along the length of the pins, rather than along the side of the package. Although actual drag-soldering is done as you show, if there's too much solder there it will just accumulate on other pins as you find. But very often, if you drag in the other (90 degrees) direction, the bead will separate.
You can easily remove a solder bridge by using a cleaned soldering tip that has no solder on it. Drag the solder away from the pin at a 90-degree angle to the pins, making sure not to drag it toward another pin. Clean the tip each time you remove some solder, and repeat this process until the bridge is gone. Hope my explanation makes sense. :)
Excellent video and project, I currently have the "chipo" 8 input logic analyzer from Aliexpress and maybe looking to expand to 24 inputs. this tutorial come at an excellent time. keep the good work!. thank's again
I would probably do a board like that with an iron and a spool of solder. If you're using paste, it can be useful to have a stensil. I wold also recommend a hot plate instead of hot air. The hot air can blow components around like you saw. It also struggles if there's big copper pours. Miniware makes a nice small and cheap one.
Use some decent tacky flux paste before putting the iron on there and clean the tip after every drag it will prevent extra solder from bridging the pins. Dirty tip with solder on it will deposit on other pins so clean it often.
I recently bought the LA2016 from Kingst for €160. 16 channels, 200MHz sampling rate, 50M hardware memory depth, 1G software, 2 x PWM output. I am very satisfied, the software is great.
If you can stretch to the Digital Discovery I would recommend that instead. 40 channels at 100 MHz, 32 channels at 200 MHz, 16 channels at 400 MHz and 8 channels at 800 MHz. 2 Gbit RAM which means 256 M samples at 8 channels and it has 16 channels capable of outputting arbitrary signals at up to 100 MHz, it can output clocks, PWM, protocols, custom signals, etc and can connect to and communicate with interfaces like i2c, uart and SPI and you can write scripts to connect to other protocols or just to control the logic analyser and it’s other functions.
If you want to remove bridges on pins, move the soldering iron away from the chip (down the leg) rather than across the pins. Drag soldering is done better with a chisel tip and you don't need half of that paste!
If you are almost soldering chips one-by-one, then you may as well be using solder wire and an iron, it's probably literally easier because you can fix one point of a chip/passive at a time and then easily solder the rest of the points/legs. Most of the point of using solder paste and hot air is do many at once. I would place all of the paste and all of the chips at once, and have the hot air gun higher up, pointing directly down, to hit larger surface area. Less likely to blow chips around (although definitely still possible) and actually likely to evenly heat and solder everything normally. I've also found aliexpress they are selling little hot plates that are 5.5cm squared, they use a USB-C PD power supply and cheap as chips (~$10-15 USD) and they're not perfect... but they are good enough for also heating up smaller PCBs and possibly doing all the soldering for you (without needing hotair) or preheat the board for you. It could probably do that entire buffer board.
flux would be a helpful addition to your soldering workflow. with a bottle of IPA to clean up the excess will allow you to easily remove the excess solder
To remove bridges, it's better to add flux and drag along the pins, away from the chip. Also, flux on your desoldering braid helps a lot. For chips that big you could just use the iron with the solder paste really - paste has a lower melting point than solder wire, so it's very fast to do it with an iron. It's also better to reflow straight down (I know you're doing it so you can record on camera though). Low airflow, and high heat - I use 400 degrees. I've done a LOT of manual reflow in my time, even though I mostly have my boards assembled at factories. :)
If you’re just soldering the pico down you don’t need those headers unless you want to. Just solder the castellations (little cutouts) to the carrier board.
FWIW with those two capacitors it was less likely the hot air and more likely the surface tension in the solder paste causing them to shift, as others have mentioned, you've got WAAAAAAAY too much solder paste there. And when you have continuous blobs like that small components will tend to float across as the solder paste tries to ball up.
I have the Saleae Logic Pro 8. I tried one of those cheap ones like what you first showed. I found it to be useless. Probably because it wasn't fast enough for my signal, but I can't remember now. The software for the Saleae is great. My only complaint is i didn't get the Logic Pro 16, although I really haven't needed more than 8 channels yet. I'm sure the day will come when I will.
The cheap one is good for ZX Spectrum speeds, but it doesn't default to the fastest sampling rate which I didn't realise until the second time I tried it
To remove bridges on SMD chips, you don't drag along the pins. Instead, clean off the tip completely and then drag away from the chip. That will remove a small amount of solder that gets stuck to the tip. If a bridge is still there, clean the tip again and repeat.
@@HappyLittleDiodes It's true that the castellated holes are for this purpose, but for people not experienced in basic SMD soldering, it may actually be a little frustrating getting the hang of how to solder the castellated holes. Headers is fine if you prefer them! They position very decisively and less goes wrong in that manner.
Very COoL. Thank you for "filming" regular guy soldering.. I have lots of tips (no pun intended) now to apply to my own trials and tribulations. p.s. To add to the Saleae query, I've been pronouncing it: "Say-Lee". But I can't tell you why (really, I can't!). Thank you for the video. Clearly a sub-worthy segment Cheers from So.Ca.USA 3rd House On the Left (please call before stopping by)
I've always stayed clear of SMD's until now but I may give it a go one day.. Would need the hot air station and the paste, but apart from that I'm good to go lol..
@@Retroguyuk75 try it! I was exactly the same when I started out (again, seriously) in electronics and designing my own builds in 2010 or so. Made my own 6809 and then 68000 boards, all throighhole. Eventually I plucked up the courage. By a cheap air station, a syringe of paste, and a practice SMT board. You won’t regret it. Stick with 1206. Eventually you will want a microscope setup, but you can get good ones for less then a hundred pounds.
The traces are lenght mached... But what about the pico board it self? Traces are very different lengths on it. Does it matter? Or do you leave them as is and just try to lenght mach the rest of the system, or you try to take pico boards trace delay into account? Got me thinking - not in this one - but with higher frequencies the different chip pins with their internal connections maybe matter as well..?!
This is definitely not a trace length adjustment. 100MHz corresponds to a 10ns period, 5ns edge change. In this time, current (300,000km/s) travels 3000/1500mm, so the 3mm on the PCB are irrelevant. I rather suspect that these are small coils and therefore a low-pass filter.
@@Furz35 it's absolutely (and obviously) trace length adjustment. It may not be necessary, but that's what it is. p.s. electrical signals do not travel at the speed of light.
*Some electrical signals don't travel at the speed of light. In the case of the em waves on these pcb traces, they're probably propagating at around 1/2 c due to the dielectric constant of the FR-4. However, EM waves have no problem propagating at the speed of light in a vacuum - after all, light itself can be thought of as an EM wave :)
As has already been pointed out, the pico can be soldered directly without pins. There are many YT videos demonstrating drag soldering for SMT devices. Hot air is often too harsh for tiny passives (0603 & 0402) so it may be worth investing in a cheap hotplate. I have a 10x10cm one which works a treat on boards like this. I suspect that the trace length matching does not take into account the different path lengths of the connectors but this is very unlikely to matter at the frequencies used on 8-bit machines. A useful little gadget but unfortunately 24 inputs cannot probe the data, address and control buses at the same time. It would be interesting to see how two can work together to do this.
You can indeed chain two up, but if you only have one you are limited to omitting some address lines. I'll try chaining two at some point, and I'll probably not use headers next time as has been suggested, rather than sticking religiously to the BOM
You don't need to cover every pin. The solder will flow under the contacts. Too much past on small parts (I think the smallest you had there were 0603 imperial? Can cause the parts to float off or "tombstone" which is PITA. You don't need to put resistors, capacitors the right way up but it makes things easier if you do and have to check the values later.
Enjoyed the video very much, although I kept yelling "too much solder paste!" at my computer screen. I had to learn that in the same way, too! What is everyone using for the probes? Any working better than others? I'm about to spring for the latest 6.0 version pcb (I already have a pico 2 in hand!) & start the project.
Don’t use paste unless you have a stencil. The easiest way to do SMD is to tack components on with an iron, leaded solder and lots of flux. Tack one side of the component for small ones, then do the other one. For chips, tack two opposite corners then do drag solder from the corner you didn’t tack on each side. The key to all soldering is flux, good flux and lots of it. The problem you are having is mostly because you don’t have flux. The other key to SMD soldering is to be able to see. I can easily solder 0402 by hand using my microscope. Finally, what I do to make my boards look neat, is go over components with hot air to allow them to reflow and straighten out using surface tension. I almost never use hot air for the initial soldering unless using a stencil and solder paste, but even then, I prefer the iron if possible to ensure all solder actually melts.
14:00 that top capacitor isn't soldered Did you get instructions on the headers? To me it feels more natural to have a male on the base board, to make the level shifter like a proper expansion board.
It's soldered it's just not centered. I only deviated from the BOM with the level shifter board header as I had already purchased probes with female connectors
16:30 Is there a reason you went with headers for the pico? it looks like its ok to solder the pico straight onto the board, even pads for the debug connections too.
I don’t use diy measurement equipment, that’s rule no 1 in embedded development. Especially in professional engineering. We bought two of the Saleae Pro last week, because it’s cheaper than developing it by our own.
He doesn't provide the gerber files, only the kicad projects so you can generate your own using KICAD, to the specific requirements of the manufacturer you want to use (I think Kicad has pre-configured settings for bigger manufacturers?). Pop in to the discord if you like, some members have generated gerbers they can share with you
Quick question, what is the difference between a logic analyzer and an oscilloscope ( apart from the amount of voltage or current they can be safely plugged into )
Here's my understanding. A logic analyser read out if logic signals are low or high (in this case, 0 V or +5 V) - only giving square waves, essentially binary. An oscilloscope shows you waveforms, analogue traces of voltage over time.
I'm both happy and angry that I came across your video. Happy because I'll soon have enough inputs for a logic analyzer, angry because I bought a clone of the old saleae. BTW: you know you can have 5x24 channels? You can connect them together. BTW 2: If you're working on 3.3V, the Pi Pico board alone will be enough.
Dont have your air gun too hot or you will start killing components especialy transistors. I run mine at 235C . If you have access to a 3d printer make a holder for the paste syringe that will allow you to get more pressure and use a smaller nozzle and get better application
06:22 maybe you should get a better soldering iron, even the pinecil (a good option) heats up very fast soldering stations that have the heater integrated in the tip are generally the best
Is a not sigrok driver based analyzer. (Is explained on documentation) that makes this proyect so interesting, and in my opiniln is much friendly user interface than pulseview. And also aviable for linux.
Any problem with using headers to mount the RPi? Direct soldering means if anything goes wrong, the whole thing is eWaste, because you'll never be able to unsolder it.
With Saelae you pay for butter-smooth software. It renders the waveforms with the GPU, thanks to being properly built on top of Chromium. There are plenty such apps that perform abysmally. But they got it 100% right. The software sells the hardware. Once there is a comparable open source alternative, will it be reasonable to put homebrew hardware next to Saelae hardware. At the moment it is not about hardware at all - even though Saelae hardware looks nice and works just as fine as it looks.
@@HappyLittleDiodes I meant the question more for the original solution: if done on a pure MCU, it is really tasking to drive 24 signals at these frequencies. But even a really cheap SOC FPGA can do all this in hardware and you have an integrated ARM MCU with 1-2-4 full cores just read results from memory and do interface and bookkeeping. It can also do things in hardware like matching signals, FFT, decoding serial data or audio/i2s/SPInand other stuff (like what Saelae software does, for example).
@@unixux the RP2040's PIO state machines make it *really* good for automated data transfer to/from IO pins without stressing the CPU itself at all. FPGAs are a PITA for hobbyist projects.
It looks as though the new boards can now be daisy chained to give you 48 channels with 2 Picos. 24 is an annoying number if you want 16 bits of address, 8 bits of data and r/w or a clock signal. I usually sacrifice address lines but then it's harder to read. Be good to see a video on using 2 boards together.. before I build mine and find it doesn't work 😊
WOW... this looks V interesting. Subbed here for sure. I've never used Github before. I assume I can get all the necessary files there.. like gerbers and a BOM?
Yes I believe you can, check the wiki page on the github. It's a bit non-intuitive to browse at first, but look for the wiki page at the top and all the info becomes a bit more easy to navigate
@@HappyLittleDiodes I'm just wondering why the Chinese haven't already trimmed this down to the minimal BOM cost possible and hence why AliExpress is not yet saturated with such clones ...
How much things changed....almost software digital analyser to 100MHz... (i know, i know GPIO are smartasses ;)) In '90 i made 8-bit logic analyser as PC/AT ISA extension card using cache chip from 486 motherboard and plenty (20+) of fast TTLs and few relays.. , RAM chip had about 20ns access time, so my limit was about 25 MHz (with software to visualise ) as my disseratation work at the end of techincal high school of electronics.
@@HappyLittleDiodes Unfortunattely not :( I'm afraid that device itself lost too... One thing possible is run software and do screen capture, but i don't have it at hand. Also schematic and plot file are somwhe od 3.5" floppies, wich i afraid lost in time and space... But seeing it would encourage you to keep doing your projects... i could made only double sided (without holes metalisation ) printed borad, so Vcc and GND were routed with wires ;)
This looks very interesting. I bought a Saleae Logic 16 back in 2013 or so. I still use it regularly. This looks like it could be a big step up. Have you played with the decoders at all? Looking at the Wiki I think I2C, SPI and UARTs are supported? These are fairly essential in any analyser setup. It looks like it's possible to write your own as well, which could be fun and interesting. Thanks for pointing this project out! Should be a great way to debug Speccy problems. :)
Is there someone in the UK that I could get a pair of PCBs from please, or do I have to load up KiCAD and get some Gerbes exported over to JLCPCB or the likes? I have the Github files but there are no Gerbers and there does seem to be an alternative.
MOSI Master Out Slave In connects to TX on pico MiSO Master In Slave Out connects to RX on pico. Think i would use an example SD card sketch to test that its working. According to the pinout, sd card CS to pin 5, MISO pin 4 MOSI pin 3 CLK to pin2.
паяльную пасту феном паять? под микроскопом отлично видна куча неприпаянных шариков,которые и смыть(вымыть из под микры )невозможно ничем и которые потом будут по плате мигрировать и коротить всё подряд. паяльная паста-только для нижнего нагрева-плиты.
Was a pretty nice solder job, until you forgot to add more flux when reflowing with solder iron.. the end joint on level converters is shitty. Nice project tho - will try to mill the pcb as I already have all other components.
Hey! Cool video!
I'm El Dr. Gusman, the next version that I'm preparing is going to have a ton of changes, all the Sigrok protocol decoders, better rendering with more samples in screen and support for the pico 2 (and the pico 2 will have three times the ammount of samples and probably will capture up to 200Mhz).
The man himself! Thanks for your work and leaving everything open like this. I'm looking forward to learning more about the protocols.
More samples and 200MHz? Sounds great. That's the upper limit of the level shifter chips right?
@@HappyLittleDiodes Yes, that's the shifter's limit, and way over the limit of the pico, to get 200Mhz I have to overclock it to 400Mhz, with the original pico I did not got it stable, but with the pico 2 and a little heatsink seems to work without hikcups :=)
@drgusman exciting!
It would be nice to have a version that does not use the micro usb, instead usb-c
@@Kmnri And there are, you can buy chinese clones that use USB-C instead of microUSB :)
You don't need that much solder paste. Use a toothpick to apply it. When you drag the iron across the pins, add flux over the pins so that the solder balls up more and keeps the bridges from forming. Also put a bit of flux on the solder wick before you use it, it will help the solder be drawn into the wick.
Great tips, thanks very much
Aye, flux is your friend here, it makes the solder obey your wishes.
Definitely flux is your missing ingredient here
In addition to what other people say about soldering these, I also have a tip:
1. When reworking solder bridges on surface mount chips like these, drag your soldering iron parallel to the pin itself, meaning outwards from the chip and not along its edge. That'll motivate the solder to stay in its lane (on its pad) by taking advantage of cohesive properties (tendency of solder to stick to itself more than other things).
Dragging from pin to pin along the edge of the chip is good when you have tons of flux and no excess of solder, as there's less solder on your iron's tip that can bridge the pins.
2. When soldering big chips, it's useful to swirl your hot air gun from further away first to get the flux in the solder paste molten out of the paste and preheat the PCB. That'll both help with the solder adhering by removing oxides, *and* melting the solder more readily and instantly when you do get it hot enough to melt. That's good practice, and similar to reflow ovens where the actual melting stage is very short compared to preheat, soak (flux stage), and the cooldown (controlled solidification. The preheat step is often where a preheating plate comes in, but for small PCB's like these, the hot plate is not worth the space it takes up or the money it costs unless you just really like collecting the tools for it.
3. I think Salae is pronounced "Say-Lay" or "Suh-Lay", similar to the French word for sun, "soleil".
It's just like GIF. No one knows how it's pronounced unless a company representative corrects you, and even then it's up for debate!
Nice video about the alternatives to commercial analyser solutions.
When it comes to soldering, there are a few things that would help:
- Being conservative with the quantity of solder paste, it's easier to add solder to individual pins (with very thin solder).
- Wicking will nearly always remove too much
- When drag soldering it is better to drag "along the pin", from the inside out for an IC pulling the solder to the outside extremity of the pad rather than from pin to pin otherwise this will affect a whole group of pins (and require wicking).
- When making a really great number of boards, a stencil (stainless steel or even plastic) becomes a valid option.
Happy that YT saw fit to put this into my feed today. I've been shopping for a reasonably priced analyzer (sorry us spelling is drilled in hard at this point) so this is some good data. Definitely worth the sub
My pleasure, Mike
If you want to use a syringe to apply the solder paste, get some of the plastic, conical dispensing tips. They take less force to squeeze the paste through and you can use a smaller diameter tip. A pneumatic solder paste dispenser will make things even easier. They apply a fairly consistent amount of paste with the press of a button.
Thanks for the tips!
@@rocketman221projects and in my view little worry of paste going off or needing to be stored in the fridge. I’ve had a tube in my drawer for a year now and it’s still fine.
Mechanical leverage to the rescue! If you have a 3D printer, there's also some printed options available for dispensing.
The ras-pi pico can be soldered directly to the board without header pins if desired
I'll keep that in mind for next time :)
The chosen enclosure determines permissible mounting method. Imagine USB socket being blocked by a wall.
Conical tips are a bit of a pain for SMD soldering. With a flat-edged tip like a beveled or chisel tip, solder bridges mostly get sucked onto the iron automatically so you only need to use wick in extreme cases.
I'll try next time! I was being lazy not changing the tip
@@HappyLittleDiodes And flux, don't be affraid of flux
@jankomuzykant1844 hahaha i was gonna say that. Dont use a nozzle either if it's a fresh blank. Heats more evenly i find.
Also, for hand soldering, get a microscope. You can get a cheap-ish one from Amscope at ~500$. Cobine with a decent 0.3mm chisel tip and 0402 or even 0201 are manageable.
The Saleae option is expensive, it's true. But it has 16 analogue inputs ! And no limit on capture depth ! I only have an 8 bit one but it's VERY good. However, for wider busses like that z80 on the title page, you might want a few bits more. You probably want at least 14 address bits, 8 data bits, and a few clocks or status bits such as R/W and /ALE.
It would be great if you could capture just a few more bits but another option is to get a 1990s analyser like an HP. For much less money than the Saleae you can have HUNDREDS of inputs - even the smallest ones have 34 or 68. And you get 'state' capture where only the actual bus transaction is captured rather than a sample every 10ns. This makes far better use of the capture memory. The loss is mostly bench space :).
Interesting stuff, thanks. Yes the analogue capability is very useful indeed you are right there. For my use (8 bit computers) - maybe less so
Yeah, mostly with the Salea stuff you're paying for the development costs of their (very nice) software suite. But yeah, it's not really affordable for your average hobbyist
I've got an HP 16500B from the mid-late 90s, and yeah having it disassemble ("inverse assemble" in HP speak) a Z80 in real-time was absolutely magical.
The no limit on capture depth only works because it can stream the data to the PC fast enough. Once the sample rate or number of channels gets high enough then streaming the data can become a problem.
@@SomeMorganSomewhere I bought the Saleae after having a hard time with sigrok with several devices. They have hobbyist discounts and small business discounts. Also they will pay you to write articles or produce videos with their product. The software is great and runs cross platform as well. This project looks pretty cool though. The higher end Saleae are really expensive in comparison.
Hello there! I know everything about this project from the very beginning because Dr Gusman is a friend of mine (and I see him in comments hehe) and I'm very proud of the way you cover it. Thanks and congratullations for this video. Already suscribed.
Thank you for your efforts and we are all looking forwards to the next version, I'll definitely build one when you are ready to release it
I used a Logic analyzer back in the day in college and it was one of those things you thought would only ever be affordable by schools and businesses. What a world we live in!
Check out the new video on this device, just released today
@@HappyLittleDiodes Lol, I'm glad you reminded me. I started watching it, then was diverted to the older one, then got sidetracked... Oh, Shiny!
I use a tooth pick to apply very small amounts of solder paste. Also I usually use the iron directly, not hot air.
Great tips I'll give it a try thanks
@@storhemulen decent flux can clear almost any bridge. But there’s many techniques, that’s a great thing about it.
I’m also new at this. Thanks for showing your process. It makes me feel so much better
Thanks that's nice to hear
Interface works very well. Smart, simple, easy and free. Set up into a rp2040 of two bucks (maybe less at chinese webstores sales). Up to upgrade when rp2 bugs fixed and improving. I2c protocols and SPI performerd like a charm. Nothing to say but thank you.
I was surprised you didn't use your new solder paste for the headers! :)
Putting the trace in a spreadsheet you can relatively easily convert the binary outputs to HEX for example which can be useful when you are trying to decode what is happening. One of these days I need to write a perl script for this and dissemble the output directly.
Great video thank you.
This is a fantastic video! Rare to see something like this. Thanks. Appreciate it.
I need a logic analyzer, and this looks like a fun project! Enjoyed the video :)
With SMD soldering, I usually just apply solder to one side of discretes and a few pins of ICs, use the hot air station to tack them all on, and then hand solder the rest (drag solder the ICs). You can wipe your solder tip off really well and apply additional flux to avoid using wick, too, since the flux will wick extra solder to the clean tip. Components with hidden pins obviously need a different treatment. I will use paste when I have a stencil, of course.
It's really good to see someone doing SMD soldering who isn't obviously already an expert, because it makes me feel like it's actually something I could do myself. Often I just sit and watch in awe and think, "well, ... leave that to someone with an iron will and nerves of steel!" It's also good to see the advice like that of @profdc9501 in the comments.
I'm happy to hear that, it is intimidating and I have made a mess of it on this channel before, but you won't get anywhere unless you have a go at it!
Drag the iron tip along the length of the pins, rather than along the side of the package. Although actual drag-soldering is done as you show, if there's too much solder there it will just accumulate on other pins as you find. But very often, if you drag in the other (90 degrees) direction, the bead will separate.
Thanks!
Flux will help a lot with that
You can easily remove a solder bridge by using a cleaned soldering tip that has no solder on it. Drag the solder away from the pin at a 90-degree angle to the pins, making sure not to drag it toward another pin. Clean the tip each time you remove some solder, and repeat this process until the bridge is gone. Hope my explanation makes sense. :)
Great tip (no pun intended) - I'll try it next time
Excellent video and project, I currently have the "chipo" 8 input logic analyzer from Aliexpress and maybe looking to expand to 24 inputs. this tutorial come at an excellent time.
keep the good work!. thank's again
I would probably do a board like that with an iron and a spool of solder. If you're using paste, it can be useful to have a stensil. I wold also recommend a hot plate instead of hot air. The hot air can blow components around like you saw. It also struggles if there's big copper pours. Miniware makes a nice small and cheap one.
Nevermind. I didn't notice it was two sided. The hot plate won't work so well.
Use some decent tacky flux paste before putting the iron on there and clean the tip after every drag it will prevent extra solder from bridging the pins. Dirty tip with solder on it will deposit on other pins so clean it often.
I recently bought the LA2016 from Kingst for €160.
16 channels, 200MHz sampling rate, 50M hardware memory depth, 1G software, 2 x PWM output.
I am very satisfied, the software is great.
I'll check it out thanks
If you can stretch to the Digital Discovery I would recommend that instead. 40 channels at 100 MHz, 32 channels at 200 MHz, 16 channels at 400 MHz and 8 channels at 800 MHz. 2 Gbit RAM which means 256 M samples at 8 channels and it has 16 channels capable of outputting arbitrary signals at up to 100 MHz, it can output clocks, PWM, protocols, custom signals, etc and can connect to and communicate with interfaces like i2c, uart and SPI and you can write scripts to connect to other protocols or just to control the logic analyser and it’s other functions.
@@conorstewart2214 200-250€ is a fair price for the performance. You don't even get an adapter for oscilloscopes for that...
Is the software good?
Nice patient, informative video. Someday I will screw up the courage, to ruin a lot of PCB's learning to do this!
If you want to remove bridges on pins, move the soldering iron away from the chip (down the leg) rather than across the pins. Drag soldering is done better with a chisel tip and you don't need half of that paste!
Brilliant video! I have, probably, no use for one of these being a luddite 😂 but it doesn't stop me wanting to build one
I'd recommend it! It's interesting if not useful
If you are almost soldering chips one-by-one, then you may as well be using solder wire and an iron, it's probably literally easier because you can fix one point of a chip/passive at a time and then easily solder the rest of the points/legs.
Most of the point of using solder paste and hot air is do many at once. I would place all of the paste and all of the chips at once, and have the hot air gun higher up, pointing directly down, to hit larger surface area. Less likely to blow chips around (although definitely still possible) and actually likely to evenly heat and solder everything normally.
I've also found aliexpress they are selling little hot plates that are 5.5cm squared, they use a USB-C PD power supply and cheap as chips (~$10-15 USD) and they're not perfect... but they are good enough for also heating up smaller PCBs and possibly doing all the soldering for you (without needing hotair) or preheat the board for you. It could probably do that entire buffer board.
flux would be a helpful addition to your soldering workflow. with a bottle of IPA to clean up the excess will allow you to easily remove the excess solder
To remove bridges, it's better to add flux and drag along the pins, away from the chip. Also, flux on your desoldering braid helps a lot. For chips that big you could just use the iron with the solder paste really - paste has a lower melting point than solder wire, so it's very fast to do it with an iron. It's also better to reflow straight down (I know you're doing it so you can record on camera though). Low airflow, and high heat - I use 400 degrees. I've done a LOT of manual reflow in my time, even though I mostly have my boards assembled at factories. :)
If you’re just soldering the pico down you don’t need those headers unless you want to. Just solder the castellations (little cutouts) to the carrier board.
FWIW with those two capacitors it was less likely the hot air and more likely the surface tension in the solder paste causing them to shift, as others have mentioned, you've got WAAAAAAAY too much solder paste there. And when you have continuous blobs like that small components will tend to float across as the solder paste tries to ball up.
I have the Saleae Logic Pro 8. I tried one of those cheap ones like what you first showed. I found it to be useless. Probably because it wasn't fast enough for my signal, but I can't remember now. The software for the Saleae is great. My only complaint is i didn't get the Logic Pro 16, although I really haven't needed more than 8 channels yet. I'm sure the day will come when I will.
The cheap one is good for ZX Spectrum speeds, but it doesn't default to the fastest sampling rate which I didn't realise until the second time I tried it
To remove bridges on SMD chips, you don't drag along the pins. Instead, clean off the tip completely and then drag away from the chip. That will remove a small amount of solder that gets stuck to the tip.
If a bridge is still there, clean the tip again and repeat.
Excellent video and cool project!! Thanks for sharing
Thank you! Cheers!
No need to attach the headers - thats what the castellated connections are for on the board
Ahh I was just following the BOM
@@HappyLittleDiodes It's true that the castellated holes are for this purpose, but for people not experienced in basic SMD soldering, it may actually be a little frustrating getting the hang of how to solder the castellated holes. Headers is fine if you prefer them! They position very decisively and less goes wrong in that manner.
peace be upon you sir
Just a good habit while connecting probes: always begin with ground.
Cheers!
Very COoL. Thank you for "filming" regular guy soldering.. I have lots of tips (no pun intended) now to apply to my own trials and tribulations. p.s. To add to the Saleae query, I've been pronouncing it: "Say-Lee". But I can't tell you why (really, I can't!). Thank you for the video. Clearly a sub-worthy segment Cheers from So.Ca.USA 3rd House On the Left (please call before stopping by)
Avalonia for the win!
When removing solder bridges try dragging in the direction of the pin instead of at right angles.
I've always stayed clear of SMD's until now but I may give it a go one day.. Would need the hot air station and the paste, but apart from that I'm good to go lol..
@@Retroguyuk75 try it! I was exactly the same when I started out (again, seriously) in electronics and designing my own builds in 2010 or so. Made my own 6809 and then 68000 boards, all throighhole. Eventually I plucked up the courage. By a cheap air station, a syringe of paste, and a practice SMT board. You won’t regret it. Stick with 1206. Eventually you will want a microscope setup, but you can get good ones for less then a hundred pounds.
@@lawrencemanningawesome thanks for the info Lawrence, I will take your advice and give it a go 😀👍
Thank You!
Pico 2, with double the RAM will be nice upgrade.
Yes it sounds like it
And more pio and larger version(60pin rp2350B) have more ADC
The traces are lenght mached... But what about the pico board it self? Traces are very different lengths on it. Does it matter? Or do you leave them as is and just try to lenght mach the rest of the system, or you try to take pico boards trace delay into account?
Got me thinking - not in this one - but with higher frequencies the different chip pins with their internal connections maybe matter as well..?!
Good point! Maybe the traces take that into account, maybe it's negligible, I'm sure Mr Gusman would be happy to answer
This is definitely not a trace length adjustment. 100MHz corresponds to a 10ns period, 5ns edge change. In this time, current (300,000km/s) travels 3000/1500mm, so the 3mm on the PCB are irrelevant.
I rather suspect that these are small coils and therefore a low-pass filter.
@@Furz35 it's absolutely (and obviously) trace length adjustment. It may not be necessary, but that's what it is. p.s. electrical signals do not travel at the speed of light.
*Some electrical signals don't travel at the speed of light. In the case of the em waves on these pcb traces, they're probably propagating at around 1/2 c due to the dielectric constant of the FR-4. However, EM waves have no problem propagating at the speed of light in a vacuum - after all, light itself can be thought of as an EM wave :)
As has already been pointed out, the pico can be soldered directly without pins. There are many YT videos demonstrating drag soldering for SMT devices. Hot air is often too harsh for tiny passives (0603 & 0402) so it may be worth investing in a cheap hotplate. I have a 10x10cm one which works a treat on boards like this. I suspect that the trace length matching does not take into account the different path lengths of the connectors but this is very unlikely to matter at the frequencies used on 8-bit machines. A useful little gadget but unfortunately 24 inputs cannot probe the data, address and control buses at the same time. It would be interesting to see how two can work together to do this.
You can indeed chain two up, but if you only have one you are limited to omitting some address lines. I'll try chaining two at some point, and I'll probably not use headers next time as has been suggested, rather than sticking religiously to the BOM
You don't need to cover every pin. The solder will flow under the contacts. Too much past on small parts (I think the smallest you had there were 0603 imperial? Can cause the parts to float off or "tombstone" which is PITA. You don't need to put resistors, capacitors the right way up but it makes things easier if you do and have to check the values later.
Cool video thank you, where did you get the probes please?
Just cheap ones from Amazon
Enjoyed the video very much, although I kept yelling "too much solder paste!" at my computer screen. I had to learn that in the same way, too! What is everyone using for the probes? Any working better than others? I'm about to spring for the latest 6.0 version pcb (I already have a pico 2 in hand!) & start the project.
Thanks 👍
Sinclair ZX80 there is some good history :)
very interesting!
I'd like a set of the logic probe PCB's if possible.
Drop me an email please
happylittlediodes@gmail.com
I decided to build the Jason Yang PCB that's in the repository, but need to know what is in the BOM for this board?
I can't find it myself, but you could try contacting via their GitHub github.com/JasonYANG170
Don’t use paste unless you have a stencil. The easiest way to do SMD is to tack components on with an iron, leaded solder and lots of flux. Tack one side of the component for small ones, then do the other one. For chips, tack two opposite corners then do drag solder from the corner you didn’t tack on each side. The key to all soldering is flux, good flux and lots of it. The problem you are having is mostly because you don’t have flux.
The other key to SMD soldering is to be able to see. I can easily solder 0402 by hand using my microscope.
Finally, what I do to make my boards look neat, is go over components with hot air to allow them to reflow and straighten out using surface tension. I almost never use hot air for the initial soldering unless using a stencil and solder paste, but even then, I prefer the iron if possible to ensure all solder actually melts.
14:00 that top capacitor isn't soldered
Did you get instructions on the headers? To me it feels more natural to have a male on the base board, to make the level shifter like a proper expansion board.
It's soldered it's just not centered.
I only deviated from the BOM with the level shifter board header as I had already purchased probes with female connectors
Hot air gun. What brand are you using if you don’t mind to share it and how happy you are.thx
I'm using a YIHUA 85D, bought from Amazon
16:30 Is there a reason you went with headers for the pico? it looks like its ok to solder the pico straight onto the board, even pads for the debug connections too.
I was following the BOM from the wiki on the GitHub but yes it looks like you're right and I've learned something today
You need some skinny John Wick and add some flux to him. Get a chisel tip or a new soldering iron.
I have a chisel tip, I'm so used to the cone tip that I didn't change it, but I will next time I'm dragging
4:50 Confirmed the diode is little, but is it happy?
I hope so!
I don’t use diy measurement equipment, that’s rule no 1 in embedded development.
Especially in professional engineering.
We bought two of the Saleae Pro last week, because it’s cheaper than developing it by our own.
Yes that's a given! For hobbyists though this is ideal
@@HappyLittleDiodes sure!
Is there a complete kit with part ? for some apparent reason I keep see the plain boards .
No you'll need to source the parts yourself
I'm new to KiCAD - are there Gerber files available for this?
He doesn't provide the gerber files, only the kicad projects so you can generate your own using KICAD, to the specific requirements of the manufacturer you want to use (I think Kicad has pre-configured settings for bigger manufacturers?).
Pop in to the discord if you like, some members have generated gerbers they can share with you
Quick question, what is the difference between a logic analyzer and an oscilloscope ( apart from the amount of voltage or current they can be safely plugged into )
Here's my understanding. A logic analyser read out if logic signals are low or high (in this case, 0 V or +5 V) - only giving square waves, essentially binary. An oscilloscope shows you waveforms, analogue traces of voltage over time.
@HappyLittleDiodes noted, thank you.
Pretty cool!
Where are the PCB-s? If I decide to buy ones - what to do?
Email me I'll help you get one
I'm both happy and angry that I came across your video. Happy because I'll soon have enough inputs for a logic analyzer, angry because I bought a clone of the old saleae.
BTW: you know you can have 5x24 channels? You can connect them together.
BTW 2: If you're working on 3.3V, the Pi Pico board alone will be enough.
Yes I'm going to do a bit on chaining then up soon! Ordered some more PCBs
@@HappyLittleDiodes did you saw modified PCB? They have TVs diodes for protection.
What paste are you using?
Chipquik SMD291AXT4
Dont have your air gun too hot or you will start killing components especialy transistors. I run mine at 235C . If you have access to a 3d printer make a holder for the paste syringe that will allow you to get more pressure and use a smaller nozzle and get better application
06:22 maybe you should get a better soldering iron, even the pinecil (a good option) heats up very fast
soldering stations that have the heater integrated in the tip are generally the best
I use a Weller which already pushed my budget, it heats up fast, just the hot air station heats up in just a few seconds!
Can this analyzer HW works with pulseview under linux ?
@@marekdorobczynski1727 I'm going to have to redirect you to the creator and the GitHub on that one!
Is a not sigrok driver based analyzer. (Is explained on documentation) that makes this proyect so interesting, and in my opiniln is much friendly user interface than pulseview. And also aviable for linux.
Any problem with using headers to mount the RPi? Direct soldering means if anything goes wrong, the whole thing is eWaste, because you'll never be able to unsolder it.
That seems like a good advantage to using headers. No problem at all from my experience building this
With Saelae you pay for butter-smooth software. It renders the waveforms with the GPU, thanks to being properly built on top of Chromium. There are plenty such apps that perform abysmally. But they got it 100% right. The software sells the hardware. Once there is a comparable open source alternative, will it be reasonable to put homebrew hardware next to Saelae hardware. At the moment it is not about hardware at all - even though Saelae hardware looks nice and works just as fine as it looks.
Why not use an fpga, probably with better rates and signal integrity ?
That could be a solution, I don't think you could re flash it though? May be more difficult from a driver/usb side too
@@HappyLittleDiodes I meant the question more for the original solution: if done on a pure MCU, it is really tasking to drive 24 signals at these frequencies. But even a really cheap SOC FPGA can do all this in hardware and you have an integrated ARM MCU with 1-2-4 full cores just read results from memory and do interface and bookkeeping. It can also do things in hardware like matching signals, FFT, decoding serial data or audio/i2s/SPInand other stuff (like what Saelae software does, for example).
@@unixux the RP2040's PIO state machines make it *really* good for automated data transfer to/from IO pins without stressing the CPU itself at all. FPGAs are a PITA for hobbyist projects.
Hi , does it have protocol decode ?
Have a scroll through the wiki it has some examples screenshots of what the software can do
@@HappyLittleDiodes thank you so much
@@thevoidedwarrantyyep. SPI, I2C n serial prtcol.
awesome
It looks as though the new boards can now be daisy chained to give you 48 channels with 2 Picos.
24 is an annoying number if you want 16 bits of address, 8 bits of data and r/w or a clock signal. I usually sacrifice address lines but then it's harder to read.
Be good to see a video on using 2 boards together.. before I build mine and find it doesn't work 😊
I wonder why not 32 channels with Pico 2, when the larger version is available.
I'm definitely gonna try that out when I get another built up
I'm not sure how the software would work, I assume you can run two instances at once and just have a simple trigger on the "second" instance
WOW... this looks V interesting. Subbed here for sure. I've never used Github before. I assume I can get all the necessary files there.. like gerbers and a BOM?
Yes I believe you can, check the wiki page on the github. It's a bit non-intuitive to browse at first, but look for the wiki page at the top and all the info becomes a bit more easy to navigate
Why do you need length matching ? :D
This is only for >500MHz or even more
I suppose it doesn't hurt to future proof that design for potential improvements!
Imagine how much smaller (and cheaper?) this would be if it would just integrate the rp2040 chip directly with everything else on a single PCB ...
It definitely could be, I'm not of the design process that's occurred but this is how it has ended up, I'm sure there are good reasons
@@HappyLittleDiodes I'm just wondering why the Chinese haven't already trimmed this down to the minimal BOM cost possible and hence why AliExpress is not yet saturated with such clones ...
How much things changed....almost software digital analyser to 100MHz... (i know, i know GPIO are smartasses ;))
In '90 i made 8-bit logic analyser as PC/AT ISA extension card using cache chip from 486 motherboard
and plenty (20+) of fast TTLs and few relays.. , RAM chip had about 20ns access time, so my limit was about 25 MHz (with software to visualise )
as my disseratation work at the end of techincal high school of electronics.
That's great! Do you have any photos??
@@HappyLittleDiodes Unfortunattely not :( I'm afraid that device itself lost too... One thing possible is run software and do screen capture, but i don't have it at hand.
Also schematic and plot file are somwhe od 3.5" floppies, wich i afraid lost in time and space...
But seeing it would encourage you to keep doing your projects... i could made only double sided (without holes metalisation ) printed borad, so Vcc and GND were routed with wires ;)
@@AK-vx4dyIt would be a badass recicling project!
This looks very interesting. I bought a Saleae Logic 16 back in 2013 or so. I still use it regularly. This looks like it could be a big step up. Have you played with the decoders at all? Looking at the Wiki I think I2C, SPI and UARTs are supported? These are fairly essential in any analyser setup. It looks like it's possible to write your own as well, which could be fun and interesting.
Thanks for pointing this project out! Should be a great way to debug Speccy problems. :)
I haven't looked at the decoders, no, but I assume they are functional. Hoping to find some time to delve deeper into it
This version contains SPI, I2C, UART and Parallel decoders, but the next version that I'm preparing contains all the Sigrok decoders :)
❤❤❤❤❤❤❤❤❤❤neci god
If you buy a 40pin IC clip you can have address and data setup and ready to go 😀
I'm definitely going to check that out, that's sounds like a very useful invention
Is there someone in the UK that I could get a pair of PCBs from please, or do I have to load up KiCAD and get some Gerbes exported over to JLCPCB or the likes? I have the Github files but there are no Gerbers and there does seem to be an alternative.
Come join the discord channel, we may be able to help
@@HappyLittleDiodes mm, discord is new to me. How do I get to it please?
@pepperm16 try following the link in the video description, it should help you sign up, if you're willing to sign up to it that is
@@HappyLittleDiodes Ahhh, cracked it I think, maybe...Ta
MOSI Master Out Slave In connects to TX on pico
MiSO Master In Slave Out connects to RX on pico.
Think i would use an example SD card sketch to test that its working.
According to the pinout, sd card CS to pin 5, MISO pin 4 MOSI pin 3 CLK to pin2.
When your pc gets so sick that the cursor turns green! 😅
I like green ;)
паяльную пасту феном паять? под микроскопом отлично видна куча неприпаянных шариков,которые и смыть(вымыть из под микры )невозможно ничем и которые потом будут по плате мигрировать и коротить всё подряд. паяльная паста-только для нижнего нагрева-плиты.
I think this is more an issue with my skill than the method, if I had a hot plate it would be an improvement although it might take longer
Was a pretty nice solder job, until you forgot to add more flux when reflowing with solder iron.. the end joint on level converters is shitty. Nice project tho - will try to mill the pcb as I already have all other components.
Still learning ;)
My God! Why on earth would you solder to those pin headers? Why can't you solder the Pi-Pico straight to the board?
You can do that, I was just following the BOM, next time I won't
🍄🟫
🎷