Pogo Pins: Attractive Distraction or Ultimate Wiring Solution? - DIYson Lamp Build Log #9
ฝัง
- เผยแพร่เมื่อ 26 ก.ย. 2024
- Welcome back to another DIYson build log, the series where I gradually overcome my fear of speaking on camera while also trying to build an open-source version of the Dyson Lightcycle task light. I really want to make this wiring alternative work, so I spent an unreasonable amount of time doing a proof-of-concept experiment to test its viability. I ended up learning a lot exploring this solution so I hope you enjoy seeing the process as well :)
Current DIYson and DIYson Express project repository and parts list: github.com/ste...
Big thanks to Jake Brownson and his beautiful aluminum carriage: / jake.brownson
My Favorite Filament: amzn.to/3HiAHnv
Subscribe to the channel: / @stevenbennettmakes
Hello! Hope you all enjoy the return to the full DIYson. Lot’s of experimentation and iteration in this one!
(Thanks for your patience if you noticed the re-upload, had to fix an audio issue.)
didn't even notice the audio issue :D
Sooooo much better than the drag chain. Love where this is heading
Hi, If you add Excentric Nuts to the carriege holder you'll solve the play. Check out any info regarding Excentric nuts on 3d printers, that's how you eliminate woble from the moving parts, and if it works in the 0.01mm range for 3D printers I guarantee it will work on this project.
Have a good one.
Interesting build! One thing that occurred to me was that you have the pins mounted *past* the three guidance wheels. If you mount them opposite the center wheel, instead, you should have *MUCH* less trouble. (The project mentioned at 15:53 shows this.)
i noticed the little sneak peak of the diyson super track in your previous vid. it was exactly what i was expecting and im so happy you went this route! :) i cant wait to make my own
Careful on relying on a capacitor. Of course, I would add one, it will help - but when testing, if you don't have a load similar to the light on the circuit, it will appear to fix it more than it really will. Like in your test case (at least as you showed it), there is effectively no load on current path, so the capacitor is smoothing out the 'signal' much more than it will with a light attached.
Capacitors don't fix the source of the problem, electrolytic caps also die after 5000hours of usage.
Another problem I can think of right now would be sparks - if the power disconnects and regains connection the sparks could "destroy" the copper layers and pins over time (that's a problem without the capacitors as well but they sure would make it worse)
maybe using something like motor brushes would be a good idea - just a thought though
@@lyqide8123it would fix all problems, you can use thé graphite from a soft pencil, graphite is very conducting and self lubricating, it's thé same material as used in dc motors, "Carbon Brush", also slide potentiometer use thé same principe 😁
@@RoyvanLierop and can be replaced
@@lyqide8123 Motor brushes might be a really good idea. Some even come spring loaded. Not sure if you can get them this small though.
Jake is now a certified DIYson legend. Hope he makes another one for the final version.
Agreed.
I think a solution - or an approach that makes other solutions more likely to succeed - is to move the placement of the pogo pins. At the moment they are mounted away from the point the arm is pivoting (somewhere between the rollers) - which magnifies the misalignment. If you get the pins mounted closer to where the arms pivot - the length of travel of the pogo pins may be enough to keep contact during lamp adjustments (which may also explain the positioning of the original power block @ 15:53 - at the point that the arms pivot around). Just some thoughts.
I came down into the comments to mention this if noone else had. Glad to see I'm not the only one with this thought.
I also thought of adding springs to ensure the contact independently of the depth, solving the other problem!
true, putting the mins closer to the centre of rotationwould help
I was thinking about this too, and it got me wondering about the relationship between the geometry of the rollers and the location of that pivot point. I’m curious how the first carriage design would fare…
I had the same idea. When aligning the pins inside the bracket, you would need to take it off for changing the pins, but since there are no wires between the base and the led it might be ok. In the center the pins might not loose contact that much.
Also relocating the supertrack to the lamp from the top the the inner side would compensate for dust buildup.
Its a great project. i love it.
Hi!:) You could use carbon brushes, like in electric motors. Those are fairly long lasting considerint how many time a motor turns before brushes need to be changed. They are "self lubricating" and somewhat selfadjusting, because of the graphite itself can be seen as a lubricant.
This is a better design. The metal pins will ruin the PCB in a relatively short time. He could use pencil graphite which is readily available in about the right size for drafting pencils.
Even better would be to design a small metal wheel on a spring.
most rotating speaker array sirens use a collector ring and brush assembly and those use carbon brushes like motors and they work almost all the time
Also they are very cheap, you can get a pair of power tool carbon brushes for under $5
@@court2379 Came here to say this. Pins will destroy the copper tracks.
I don't know if the dust and the look of the smears they will cause on the tracks aren't dealvreakers though
Loving this series, your iteration and attention to detail is really inspiring! One potential alternative for the pogo pins could be some off the shelf carbon brushes intended for use with a small dc motor. These are spring loaded and maintain electrical contact to the spinning part of a motor so keeping contact with a linear smooth track should be no issue for them, they may also offer a little more surface contact area which could help with any slack or twisting in the assemblies. Looking forward to future videos!
Yes carbon brushes were my first thought. POGO Pins aren't really designed for any sort of friction, usually used for static connections between boards for testing.
The problem is that 1) They have to be replaced over time and 2) they put carbon dust everywhere, which is probably not ideal for a clean office space full of technology.
@@lbgstzockt8493 yeah the dust and wear isn't ideal, but it's not moving too much to produce a lot of dust. Plus it's better that inexpensive brushes wear down and need replacement, rather than the copper on the track wearing due to friction with the pogo pins
Also of note, carbon brushes as used in motors can last for a couple years of use on something that spins thousands of rpms. The likelihood of them wearing out quickly on this lamp is very minimal. To coincide with that, the movement of this is slow in comparison so any carbon dust that could arise would not be flung anywhere and could be easily wiped off if it gets thick enough to bother you.
Agreed - one risk that I see with the pogo pin is that over time they will wear the contacts on the track. The roller pins would be better as they create less friction.
That moment when you have a whole set of things you want to test, with neat possible solutions in mind and you end up fixing it on the first try
I can't get enough of this project, for real!... Your commentary ties it all together very nicely, and I love seeing and listening to you go through your creative process.
Thanks Kyle! Glad you appreciate seeing the process 👍🏻
Great solution to the wiring issue. Sounds like it will work. Must admit I didn't like the wire chain solution, but figured it's the best way considering the patent limits. But this really works from the aesthetics point of view. a DIY solution for an extremely tough issue. Won't be surprised if you get a nasty email from Dyson now that you've basically achieved their own design, but without patent infringement.
No nasty emails so far🤞🏻
Getting yourself very prepared to solve a complex problem and fixing it first try is so funny. This is one of the best build series I've ever seen. Can't wait to see more.
TH-cam just reccomended me this project, and I gotta say, I love this project and can't wait to binge all of the other build logs! About testing the pogo pins, one way you could've tested it is by putting either a small voltage (~0.5v) or a high frequency signal (around 10KHz sine wave) on the rails, and connect the pogo pins to a 3.5mm audio jack, then to a line in of a PC or a audio recorder. you can then hear or see (in Audacity or other software) the failure as drops in the signal, which sounds like when you turn a headphone jack in a bad plug, but you have much more resolution than with the method you were using.
With that resolution, you can characterize the drop-outs by crackles, very short drop-outs that're almost unavoidable with moving pogo pins, pops, much longer drop-outs from a bad connection, or actual drop-outs. With some data analysis, you could characterize those automatically and find out what solution is actually the best!
Keep in mind that If you do go the route of putting just a voltage on the pins, you'll have to remove the AC coupling capacitor from your line in on your sound card. Luckily, cheap USB sound cards with line in plugs are plentiful to buy online, and they're easy to hack to remove DC coupling!
I’m really enjoying the series! I think a design aspect of the light cycle that may be worth testing here is just the location of the connection, positioning the pogo connection between the two v wheels will get them closer to the axis that the arm is twisting on. The closer you get to that axis the less apparent the rotation should be.
I was thinking the same thing. Try putting the pogo pins on the sliding plate that connects the vertical and horizontal extrusions. If the pins are at the center of rotation it should defeat the twisting misalignment.
Furthermore, if done well, it could completely remove the clip currently holding them in place!
Yes, this thread should be pinned up
super cool ideas and developments again this time, very well done for the thorough testing you did, that seemed like a ton of work.
the tidiest approach i can think of to address the issues you had would be to place the pogo pin connector(s) inside the carriage as central as possible to the various wheel contact points, so deflection is minimised. if you can find a double-ended pogo pin connector, a single connector bridging the tracks at the intersecting point of the two arms would be extremely tidy with virtually no pivot to amplify the deflection.
this is the solution that came to my mind as well, double-ended pogo pin at the intersection and super track moved on the other side of the rail. If no double-ended pobo pin available you can 3d a solution and just solder the pogo pins
I love the aesthetic of the black supertrack PCB and black aluminum extrusion.
"No need to panic, I can create my own problems." 😂
You could try something similar to slot car track - instead of pogo pins, they use little metal brushes to stay in contact with the track. Brushed electric motors also use, well, brushes.
I really hope he sees this comment. This would really solve the problem in such a simple tried and true way.
Stephen, your videos are great! As soon as you upload a new one, I drop everything and go watch it.
Your content is on par with that of StuffMadeHere and Alexandre Chappel. Keep up the good work!
Wow thank you so much, Zohar, it's so motivating to hear that.
100% agreed
This build is amazing
Couldn’t agree more! Question Steven, do you happen to have an Instagram as well where you share progress on your various projects? I’d love to see more of your content in addition to the long form. Regardless keep up the amazing work, you’ve definitely inspired me to take on some new projects!
Awesome as always! One solution to the pins could be to have them pressed down with a spring so when they move too far away from each other the spring keeps pushing the pins down.
Indeed! Float the entire pin mount using a compression spring to maintain contact with the T-slot. Perhaps chamfer the edges of that piece to fit the profile more 'snugly'.
Perhaps also check how little force is required to maintain contact to reduce friction wear on the track?
Haha, I really felt the "No need to panic, I can create my own problems!"
I like your voice and I can see myself listening to the audio from this at bedtime to calm me down and help me drift into sleep.
Totally my favorite series on TH-cam. I’m so excited for upcoming videos. Great work and so much inspiration for my upcoming industrial design internship!!
YES! I have been designing my version of this project with the idea of using a PCB and pogo pins for the build... glad I wasn't the only one! The diference in my design comes down to the position of the PCB. I put the PCB in the side facing the carriage, making it directly perpendicular to the other track and allowing for a simple hole in the carriage that I put 2 wires through to pass the electricity.
Really enjoying watching this series of build logs that make the lamp better and better with each episode along with your thought and idea. Really admire your strike for perfection here. Will definitely build a copy of your lamp when I find some time
Glad you're enjoying these, thanks for watching :)
The supertrack solution came immediately to my mind as i saw your dragchain solution but with two seperate tracks.
same here, also I suggest to use a carbon brush like motors have in rotors. a bit bigger and wearness will be forever in this particular case.
@@josedavid6400 At least I would try to use some larger contacts. Because I think the wear of the point load is a bit too much.
@@Hassla agree!
Just discovered this series last night and binged it all today. I am so impressed, this is a remarkable project. Can’t wait to keep up with it now!
I'd re-orient the horizontal PCB-power strip to either the top or bottom of that rail (rather than the side), which means gravity can assist with keeping pressure on the pogo pins. Wait, you did that! Ha. Good update :P
I'd be interesting in seeing how the pins and tracks wear over time, and in particular how/whether the resistance of the contacts changes. That's where I'd expect the most power losses to occur. The power loss will change with light output power (P = I²R), so definitely worth measuring.
Good set of iterations this video!
I clicked the video initially because the thumb looked like a zero cable solution for a 3D printer and I was intrigued! I'm glad I found a much more interesting project. I'll build one when you find a great design checkpoint.
I was one (of probably many) that suggested a solution like this. A quick follow up to increase the reliability is to use the extrusion itself for the ground connection and use a thicker single line for the + wire. Or for a quick modification, use both for +, if one looses contact, you got built in redundancy. Also add an isolation (paint) to the side of the pins to avoid rail contact. And finally take a look at supercaps. They are tiny but store a lot of energy, even if the contacts got interrupted, they will power the lights for probably a second or two. Also your type of measurement of the reliability is flawed. Without a load, the voltage is more or less meaningless and not a good indicator of contact quality. If you use a multimeter in resistance mode, you got more details, or add the load (the light), then analog voltage will be fine.
I was also thinking that using the extrusion as the ground plane and just a single track with perhaps a leaf spring contact on the moving axis’s may be more reliable
I'm so glad you went for this design change. I suggested it in the comments of one of your earlier videos and I'm so glad you got it to work.
Try placing the pogo pins between the bearing wheels, in which case, the relative deflection between the PCB rails and the pogo pins, would be minimal. Really neat design.
I love the way you show the iterative process of getting from v0.1 of a part to v5. That's one of the most interesting parts and some people just gloss over it, barely showing v1 and straight out put v5 into the spotlight.
Also: "Now the problem is gone?! No need to worry, I can create my own problems!" - spoken like a true engineer xD
Reminds me of these Tracks you build at home which you can drive small cars along. They always have these woven copper contacts maybe something as simple as that will work for you to
I really like the idea, it looks incredibly good with those pcbs in the tracks, although there is one issue that comes to my mind, usually the pcb gold coating is extremely soft and thin when not explicitly bought with "hard gold" finish. Even then the traces might wear quickly and the underlying copper can oxidize which will lead to increasing resistance and worsening connection.
One thing to look into could be the "brushes" from DC motors, they use exactly that principle to transfer the electricity to the right coils. Those are made to intentionally wear over time but still provide good lifetime. IIRC those pads are made from graphite.
Really love the Project so far!
Maybe the tracks can be carbon tracks instead of the regular gold-plated copper? As it is applied in faders and potentiometers…
@@el_micha good idea, but then you cant order it as a usual pcb i assume
@@user-wi5vi7dd3z Probably not, true.
Yep, L, I see your comment now after I posted re: "hard gold" option. I think this is minimum, but the wiper material should be softer thank POGO pins too.
1. Your power transfer tracks are basically just straightened slip rings - which suggests a good source for a good bit of well-ironed-out R&D. I seem to recall multiple (redundant) contacts made of single bent-metal 'springs' with high conductivity coating. 2. as something of a designer, I keep wanting to see your carriage plate's corners be more concentric with the wheels at each corner. That tends to keep an even material thickness between the holes for mounting screws and the outer edge of the plate, and tends to make things appear a little more intentional and harmonious. 3. Great stuff. Really enjoying this series. I'm about about to get my first 3D printer and might need to take a crack at your battery-powered one
Been following this series and right from the first video I've asked myself why use those tracks-chains ..
I have a couple of suggestions:
- Super-track center slot : a 95% (100% movement length?) slot to allow clipping the track to a 3d printed "inverted T" holder. This would increase isolation between tracks and make usage of other suggestions, such as carbon brushes, or slotcar brushes type connections.
- Placing the connection on underneath of the beam. Gravity is your friend and would work on favor of maintaining contact between the track and the connections.
- place a diode in series with the capacitor and place them both at the "head" of the horizontal beam. The diode will help avoiding inadvertently return of short-circuit current to the track.
Amazing project! Well done on the date collection and iteration. It does seem that your pin mount position is the main issue here and could be solved by moving the pins so their travel direction is matched but the "rotation" direction of the horizontal arm. I would think that mounting the pins on the top or bottom of that rail and moving it closer to the pivot point (in the middle of the carriage) would help fix your issue. (which you touched on a little but were talking about mounting it directly to the arm rather then just changing the mounting position.) You could also add some additional spring force/travel distance to your pin assembly by designing in some flex/spring into the 3d pin housing itself. Again, great work! keep it up
Just wanted to comment that this is what I thought as well. We are supporting the track on the top and bottom, so adding the rail and pogo pins there might elimate some of the issues you're facing. I'm loving the series and seriously considering building my own DIYson in due time! Definitely think that this rail solution is a lot cleaner than the wires.
Another really fantastic and impressive build log. I really enjoy the way you're putting together all the different aspects of engineering yourself, and coming up with great ideas to solve problems and test the solutions to really get into the details of what's going on. Can't wait for the next one.
Thanks Graham! So glad you're enjoying these 🙂
I completely understand your feelings about designing the pin holder, those sorts of small designs allow 3d printing to more or less fulfill the almost magical perception it has in the eyes of the broader public. Truly rapid iteration is possible due to the minimal time it takes to print such small designs, allowing you to be able to make very minor changes one at a time until you achieve perfection in a timely manner that wouldn't be feasible at any other timescales.
Great video, Steven! To build on your pogo pin idea, and the motor brush idea someone else mentioned, you could use a pair of solid contact points (or brushes) attached to a 3D printed compliant arm that presses against the PCB. That way you could control the total deviation that the contacts can tolerate without relying on the stroke of the pogo pin. Depending on how you design the compliant arm, it could push down on the PCB and flex side to side to account for the most extreme deformations you showed in the video.
This is an awesome project! I was actually theorizing my own DIY Dyson lamp a few years ago, it nice to see you execute it so well :) One idea I had was to use two 1020 aluminum extrusion pieces with a 1mm insulation layer in between, and use the two together to make a 2220 aluminum rail. One side for negative and the other for positive DC power. Would use the bearings as power transfer which obviously isn't optimal but would work for low power draws. Would also need to remove the anodization on both the v slot wheels and extrusion slots.
Great update. The pcb track really looks like a good idea But. You will have to start thinking about wear and tear. Just like brush motor, you should try to find pogo pin or you should try to make your own pogo pin from a softer material making the pin wear but not the track. The pin should be easyer to replace and cheaper. Also if not yet already done, for the final pcb (if still done with a pcb) should be gold plated so the copper doesn't oxyde with time.
Here are my tips. Good job and don't stop. Will probably build one when I change desk.
Hey mate, the lamp looks good. In normal applications you'll see tracks exactly like what you've got, but the contacts are normally phosphor bronze leaf springs. Either a turn back design, which needs to be very carefully engineered, or, for something that will work a lot better for you, just a straight leaf spring. You can bake in a LOT more travel so that you get a good long length of contact and a very consistent pressure which will eliminate the variation due to contact pressure you're getting. The second thing is that you'll have a point of rotation which is about the front roller in your setup, put the leaf springs over that and that'll eliminate lateral motion due to rotation. I hope this helps.
Love to see the progress made. This episode hit kinda close to home, as at work i build test fixtures for pcbs, using pogo pins.
My guess for the problem only existing on the horizontal arm, not the vertical, is that you attached the pins in between the guide rollers on the vertical arm, and outside on the horizontal. This could lead to more movement of the arm itself.
Second, in the video the wheels seem to be not that tight on the arm, i think readjusting here could also make a difference.
Great work. It's very nice to show how detailed designs must be and how much work and testing it's required to get a nice functioning product. Thanks for sharing
Watching your videos is just so much fun. I like that you try to perfect something rather than churning out weekend projects although I totally get the appeal of this too. Have you considered making two separate wider lanes? Maybe one at the top and one at the bottom. This would give the pins more tolerance. It would also add to the cost of course but maybe its neglectable. Another solution could be a clip of sorts around the horizontal arm attached to the carriage? You could then mount the pins to this pins so they hopefully move together with the arm. I am stoked for your side project by the way!
I'd replace the pogo pins with long leaf-spring copper contacts. Lower contact pressure to reduce wear issues, but a much longer "hight" travel to handle the contact block moving away from the track. As used in my 40-year-old VW campervan, for the horn push contacts in the middle of the steering wheel. Proven and long-lasting solution :)
Really enjoying your iteration process. These points have probably been made many times (and you might have reached them already on your own) but...
- Wipers would probably work better than pogos. Because pogos are perpendicular to the track, getting ones with longer travel will mean they'll become more likely to bind as they get pushed off-centre when the track moves. But wipers can have much more travel and won't bind. They are used in potentiometers, faders, slip-rings - anywhere that current needs to cross a friction point.
- Mutliple wipers (I'd suggest a pair each side of the torsional axis) improve your odds of one being in contact at any one time. Your idea of mounting to the connected axis might solve this anyway, but I'd still suggest using multiple wipers so you're never reliant on a single contact point. Most wipers also have a split "foot" so they effectively give you a doubling, but more wipers is more gooder. Camera lenses use this trick a lot. Also wipers (and multiple wipers) can have a LOT more surface area, which will become very important when you...
- Test under load. Your test rig is measuring voltage, but at near zero current. This means that if your contact suddenly introduces say 10 ohms, you might not measure any significant voltage drop, but if it was under load the light would go out (figures are examples only, point is that not testing under load means you're not testing the thing that matters). You could just measure the voltage at the input to the lamp head instead (use a voltage divider if you need to keep the pin inputs in range) and that should tell you what you need to know.
- Related to that, your capacitor helps smooth the voltage drops, but you would need a bigger capacitor as your load increases. With very little current, the cap can hold the voltage up easily, but you'll need a pretty big cap to power the light for the same situation. You could use a supercap but it's not a terribly elegant solution - I'd expect you should be able to get the mechanical part of it solved instead - but a bit of filtering is rarely bad :-)
- Longevity, DC and corrosion. DC and AC behave differently at contact points. This is why a given relay will have a much lower DC rating than it's AC rating. DC tends to arc - which in the short term might be good for keeping the circuit going, but longer term will cause carbon buildup and pitting on the track. I think you might also be more likely to have corrosion issues with DC due to electrolysis (and other effects?) which I believe that AC tends to have fewer issues with. They deal with this in copper phone lines where the battery voltage applied causes corrosion to build up at every termination, but it tends to be kept at bay by the AC pulses during dialing or ringing (I'm not 100% on the specifics there, but it results in problems on ADSL lines when they no longer get used for normal phone calls). I don't know if / how much that might become an issue for your design over time - it might be only theoretical, and the scrubbing action of the contacts might prevent it ever being an issue. You could send the power as AC (either at the mains 50/60Hz after stepping down to a safe voltage via a transformer) or you could have a chopper circuit create a pulsed or shaped AC and rectify it at the lamp head. Or, just solve the problems you *know* you have before worrying about the ones I'm making up!
Looking foward to your next installments!
Really loving this series! The improvement with capacitor will not translate when the power hungry light source is added. So do repeat the test with the actual load before finalizing the solution.
All the best! Looking forward to the next iteration
How about a bigger capacitor or even a Gold cap - as used in bike rear lights, for making it stay lit when you stop at traffic lights?
Yay, I posted this idea a few weeks back. I guess great minds think alike 😜
I had another idea for your pogo pin problem but making the clip slide inside the track to keep your alignment is better. That and you can alway add more bigger caps!
You could combine the best of roller and standard pogo pin by using a two part pin holder to place a small spring (think ballpoint pin spring) behind the roller pin. An issue you'll run into with a plain pogo pin, is that because you have a direct sliding motion across the track, over time you will wear through the conductor and possibly the pogo pin as well; a rolling contact would help quite a bit there. Also, since you're already building your own pin assembly, you don't need to keep that same .100" spacing, and can have a wider pitch, along with widening the traces (and perhaps reducing the trace clearances) to create a larger allowable angular deviation in the arm position.
Came across your project on this video, I watched it back from episode 1 and I am really impressed by your talent ( both at designing and storytelling ) and I really want to build one for myself now.
Regarding the pogo pins, a solution could be to use 2 sets of pins for the horizontal axis.
I don't have your assembly to test that out, but to me, when you rotate the light, the horizontal axis twists around the vertical one right at the carriage, so one side of the horizontal axis goes away from the pins, while the other side comes closer to it.
So if you were to mount 1 set of pins to each side, when 1 set of pins loses contact, the other gets compressed against the track. This also adds redundancy in case a set of pins would get misaligned.
At lesat that's how I see it, but I might be... Off track
Hey Steven, just wanted to say that I love this series! Following your journey of thought and discovery, and not to say self reflection and learning, is highly fascinating. And as a prototype machinist, those CNC carriages came out great, I'm sure Jake would be up for making more of your final design, you just gotta finalize it :P
BTW, those tracks are called busbars, and you could spring load the holder that holds the pins, at least that's how it's done in industrial applications.
I'm utterly enthralled with this series, and you explain your thought process and discoveries in a clear and logical manner which I really enjoy. Also that Excel magic was whack. Smart man. Keep it up! Good luck, and looking forward to the next installment.
Hi Steve, Huge fan of this series. I have a couple Elegoo Saturn Resin Printers that I would like to volunteer my printers for any of your resin printing desires/needs. Secondly, I'd like to speak to you as someone with experience in both the Lighting Industry (previous role) and the Furniture Industry (current role) (No, I don't work for Dyson, they're a competitor in some spaces.) that some of what your encountering with your electrical issues are mechanically related. First, I believe some of the "Rotational" issues that you're incurring are when you're trying to either move in an XZ direction, or XYZ direction, has to do with mechanical play/slop. Your aluminum extrusion is not securely fashioned to your base. There is movement between your base and the beginning of the extrusion, creating an opportunity for the vertical arm to rotate in a plane that is not originally intended based on what you believe is rotation motion in say, the XZ plane or YZ plane, when you're basically moving the light in a flat circle, or arc, if you're looking at it. I believe that some of this could be fixed if your extrusion was recessed into a tight-fitting cavity, with outer dimensions matching the square of your base, to keep it aesthetically pleasing. Second, I believe you're seeing some of that slop in the horizontal axis because you're using static v-slot rollers. I don't' recall if one of your previous builds you mentioned using an eccentric nut in order to tighten down the contact between the upper and lower wheels, but Dyson decided to overcome part of that challenge by adding the spring load to the caster, maintaining constant tension on the 3 points of contact, regardless of the change in moment/weight created by having the end of the arm with the light on it, further out. or closer. I think if you were to return to some of the mechanical aspects of your design that are introducing an accumulation of inaccuracies, I genuinely think that if you could take care of a big chunk of what your encountering on the electronics side. Also, I applaud you on thinking about cost, and ease of assembly, which some of us in the industry refer to Design For Manufacture and Assembly (DFMA), however, quality can't suffer when cutting costs. Engineers also like to say "Fast, good, cheap, Pick 2" Let me know what you think, feel free to DM me for more info about the resin printing.
Really like your attention to detail, seeing the iteration, and exploring a range of possible solutions to each problem that arises - with full video coverage. Also great to see extracts from your notebooks as part of the record keeping to remember everything with extended projects. Thank You!
Interesting that the Dyson mounts the pickup to the horizontal bar on the underside mid-way between the bottom two wheels thus reducing twist (?) rather than on the side. [One option could be to add a 4th wheel, so two above and below, to reduce unwanted movement. Might not detract from the minimalist goal too much, for a better solution.
Really inspiring build log. Looking forward how the pogo pins will hold up, or rather the pressure of the pin contact. Though from my experiences with power rail energy transfer its the contact patch that gets worn out quicker, that is with copper contacts. I like the adjustment possibility of the pin holder, great job!
What also matter is how far the pogo pins sit away from the horizontal and vertical axis. When applying leverage on the horizontal axis, it's obvious that the pogo pins loose contact. :)
Wow, I binged the whole series today expecting a bit more closure since it started last year... Either way, amazing documentation of the process. I think where I also felt the need to maybe make my on version of the lamp after seeing it (the OG Dyson) and its price tag, I'm beginning to think I should just spring for it based on my newfound admiration for the engineering. Great series overall!
I suspect that the Dyson Lightcycle might use a single flex PCB that's laid into the vertical and horizontal channels during assembly. A couple tips for you on this design:
1. Even with hard gold plating, the movement of those pogo pins is going to wear the track PCBs out pretty quickly. It might also create conductive dust in the process, leading to shorts, leading to .. bad things. You might try something like the Mill-Max rolling ball pins, which have a similar design to a ball-tip pen, but you'll need to drop the tolerances on the flatness of the track PCBs to make them work, since they want to be under a very specific amount of compression for free rotation. Good news is that they're tested to something like 65 miles of rotation before failure. They're also beryllium-copper AFAICR, which might have better corrosion and oxidation characteristics than the pogo pins you're using. Even better would be some carbon-tipped brushes. Haven't yet found a manufacturer in the US - but there are a bunch in china. Check out Alibaba ;)
2.. The wires you're using are way too small for the amount of current you're pushing through them. You should probably bump the voltage to drop the current - but you'll need to add in some protection so you don't cause a fire, etc. Look into adding a device called an "eFuse" - you'll have to design a PCBA around it, but the eFuse will detect and react to shorts quickly, and prevent a dead or partial short from becoming an ongoing hazard. Exposed contacts are a recipe for disaster if they're not done correctly.
3. If you're using a big-ass capacitor to deal with interruption of power to the light head, remember that it can release all of its energy very very quickly - it could hurt you or someone else, possibly even by melting metal, etc. As you get more into the electronics side, you might consider some protection against energy flowing from the capacitor back out onto the track PCBs/contacts. I'll let you figure that one out ;)
4. You should be able to replace the track PCBs and the ball/pin assembly with flex PCB(s) that are constrained within the channel - similar to the way you were working with your cable channels earlier. You might need a deeper channel - so no guarantees - but I think it would probably work if you're careful.
Lmk if this is useful - I'd love a shout-out! Good luck with the project. Looks super fun.
Very cool! My first thought is that instead of pogo pins you could use a small piece of flat metal as a "wiper" like in a potentiometer. This would allow a much greater range of contact and probably reduce the friction force on the supertrack. I've seen this design on air conditioning switches for cars which are built to withstand thousands of uses, so I don't think durability would be too much of a concern either.
I was looking for this type of comment! A potentiometer, or even better a fader, is very similar to the application on the rails… Maybe carbon tracks are even better than copper tracks?
subbed. this is my favourite series on youtube right now
I love this solution and I think looking to slot car tracks for inspiration might create some improvements
One suggestion. Use the aluminium structure as the ground, and the PCB track as the V+ (assuming it is 5V) to power the circuit.
Hey Steven! Great update and a clever solution, but the "supertrack" is basically a full length of exposed wire that could be touched by someone unaware or the user forgetting or not looking and adjusting the lamp. This makes it an electrocution hazard and thus in current form unfit for the market on account of consumer safety. I'd suggest some sort of chain cover that goes around the profiles or reverting back to previous iteration with cable chains. If you chose to go back to cable chains might I suggest reduction of slack by using a similar mechanism such as the one in vacuum cleaners to wind the cable back. Cheers!
Two ideas.
1. Instead of two exposed lines next to each other on one the same side of PCB i would suggest two sided PCB with tracks on both sides. PCB mounted perpendicular to v-slot. On carriage pogopins pointed at each other. PCB would end up between them.
2. Instead of PCB use flexible metallic strips with some isolation between them. Strips need to be a little bit longer than aluminum profile. Strip would lay in v-slot and at carriage it would be guided out of v-slot in order to have a possibility to touch both sides of that strip. Kinda like cover strips on linear servo actuators.
Definitely go with 2x2 or 2x3 for the pogo pins whether or not you use a capacitor. The more secure electrical continuity will be worth it especially as the track wears in and is subjected to atmospheric contamination. Cool project!
If anyone’s interested, there are pogo pins specifically made to be slidable. They are constructed similarly to a ballpoint pen, ultimately making contact with a small metal ball, with freedom to spin whilst rolling along it’s PCB track.
What could help a lot with the pins disconnection is putting the pins inside the carriage, in the middle of the guiding wheels, close to the axis of rotation of the arm. Maybe even designing the top wheel of the carriage with an excentric cam that would allow to control the fit of the wheels on the rails to reduce the rotation further.
Do that thing you said at the end, that's the solution. Have the pogo carrier attached both to the beam and the carriage, each with degrees of freedom, so the beam maintains 2D position/orientation relative to the supertrack, and the carriage drives position along the supertrack exclusively. I think the easiest option is to make a T-nut sort of thing that rides inside the T-slot, but that is going to throw a wrench into the supertrack PCB, and it'll become sensitive to the differences between specific styles of aluminum extrusion.
A more involved solution is put something ferrous behind the whole track, and have the carrier stick to it with magnets, riding in the face of the slot. Maybe home depot sells a perfectly sized piece of flatbar. But then you need to design a carrier which will gladly slide across the face of a PCB without causing wear to the PCB, and between two anodized aluminum walls without marring or wearing away. You do get to use magnets though, everybody loves magnets.
This might help:
The pogo pin holder can reach further into the cavity where the lanes are. If designed wide enough they'd provide a barrier for the rail enclosure to bounce so far around ( stops pogos swinging across the lanes).
This is awesome. I didn't expect it to go this way, and I think there could be another way, like keeping the wire length the same and using a pulley-like mechanism for the wire to adjust. Keep up the good work man.
i love the new approach for the power delivery, i didn't like at all the old solution, it messes with the whole tidiness of the lamp, which is a big part of its attractive
Have you considered graphite on a springs instead of the pogo pins? Copper on copper introduces another durability issue, which this "consumable pin" could solve. It could be installed at an angle to give it length and at this power even 0.8mm mechanical pencil tips could work. It gives you control over the ditance the contacts can move aaaaand a lot of headache. Also wider tracks and even using the aluminium body as one track could work. Love seeing your videos, it's always a treat to sit down and see you work through problems.
Make a 3-dimensional "super track" out of two PCBs, with one long track and seat those inside the aluminum extrusion. Think of a triangle shape, where your pogo pins will be on either side of the crest of the triangle.
The ridge will help keep the pogo pins in contact by resisting twist/vertical movement.
Just a suggestion: for a diy project, getting a pcb like that for the supertrack built and delivered could be a hassle for some of us in hard-customs countries (Argentinian here). I thought that maybe an approach with copper tape using both sides of the aluminum bar (top and bottom) could work the same way and we don't have to get a pcb built for it. Maybe it works for you or maybe it don't, either way i hope to keep seeing your devlogs, they are amazing!! keep up the good work!
Instead of the PCB-tracks, you could use a ~5x5mm brass bar and mount it in the track with 3d-printed spacers and screw insulators.
Put on on either side of the extrusion and you have a beefy, long lasting track.
This enables you to use beefier pogo pins or brass springs for better contact.
If you are worried about corrosion: My last car had a similar system on the steering wheel for the horn. Just a brass track and a brass spring. Not a wire spring, just a springy piece of sheet metal like the ones you find as earth pins on EU power outlets, specifically CEE 7/3. It worked fine for over 20 years until the nut that holds the steering wheel got loose and the wobbling steering wheel bent the spring before I got it fixed.
The ultimate flex would be to gold plate the sliding surface which isn't that hard but still a lot of effort.
You could also use (kapton-) tape as an insulator.
Hell yeah for a solution like this. In case pogo pins don't work brushes might do the trick, like in electric rail car toys. I really hope you find a way to make the supertracks work, this is the way I'd go with my DIYson when I get to it!
I think you could tuck a ribbon cable into the v slot groove, it might even be able to fold over itself in that space. Not sure about lifespan with that bend radius though.
As far as your pogo pin problem, one possible improvement is to move the pogo pin mounts closer to the centre of the carriage, where torsion results in less deflection seeing as they will be closer to the pivot point of rotation. The v wheels also seem like they could be a lot tighter.
Failing that, make the tracks wider so the pins don't slip past the edges. If there's no room you could made two wider tracks and put them in two separate slots in the extrusion. Or use one extra wide track and the extrusion itself as the return path (the anodizing will need to be removed where the pin contacts).
The simple fix to the rotation of the horizontal arm is to have the pogo pins pass through a spacer (possibly T shaped) that moves along the rail within that gap with that spaced being equal to the width of that track opening. That limits the side to side movements of the pogo pins within the track.
To get a better approximation of how much a capacitor will help you need to have a connected load (LED).
Also, it would be much better to move pogo pins in the center of rotation of the axis (between all three rollers) and not off-axis.
Not sure if copper tracks will be durable enough. You could design them to be nickel plated (this is a standard process and is normally done at no additional cost by the PCB manufacturer).
Love your product design workflow. Keep up the great work.
Another potential solution would be putting some wipers under the wheel. Given that the wheels ride within the V-groove, there should be some amount of toleranced space between the wheel and the power rail PCB, while also having much tighter wiper to rail tolerances given the constraint and slight pressure of the wheel and rail. Should be fairly simple, leaf spring wipers for any deviation with a holder that partially encompasses the wheel and the linear surface of the rail, which then can attach to the carrier plate around the wheel's axle bolt with a screw on each side.
Think about it like a tension pulley within a serpentine belt, where the pulley applies force to keep the belt within tension, and in the case of V-belts will help with locating the belt via tension. In this case, the adjustable wheel is what applies tension, the wiper is the part that's in continuous contact such as a belt, and the aluminum extrusion with the power rail PCB would act as the impartial contact surfaces such as various other pulleys within a serpentine belt.
Though there's also another dimension to this, the amount of rotation the rail itself experiences, which causes an issue with the pogo pins. But considering the rail will pivot at the highest point of tension between the wheels and rail, this becomes a non-issue with a wiper under the tensioning wheel. Another fix to this would just be more tension, as this rotational movement comes from slop within the setup; keep in mind that people use these these kinds of components on various CNC machines that need dimensional accuracy, the slop stems from assembly and not the components themselves. Other than higher tension, you could also use some perpendicular wheels within the rail to better locate the rail if there is any slop between the extrusion profile and wheel profile, similarly a shim of high lubricity that interfaces with the perpendicular face could suffice for this purpose as well; you could even put the wiper or pogo pins here, as this circles back to constraining the components under a toleranced area.
Personally, I would probably go with the latter. Put a perpendicular locating shim of high lubricity within the carriage, and pass the wiper/pin solution into it. Doing this within the carriage can also hide any wires, the wiper/pin set could also be soldered externally and dropped in, depending on distance and tolerances this could also be a floating part that's captured geometrically via a hole in the carriage. For example, use a 2x2 pogo pin setup, wire specific pins accordingly, put a spring between the two ends, and capture the pin set with a square hole. This could also make shimming fairly adjustable as well, just pop the rail of and adjust, as the shims themselves would be independent from everything else outside of a retaining mechanism on the carriage. Simple solution, adjustable, tightens tolerances, and hides the connection, while also not modifying existing design too drastically, nor using highly specialized components that can also be easily assembled.
This series has been really inspirational, thanks for all the work you've done!
damn the light is finally coming to it's final stages
Great improvement!
I would suggest also measuring the current, as a capacitor is essentially a short when it is discharged.
I do not know the rating of those pogo-pins, so please be aware - otherwise you might get sparks and fried tracks (in the future).
Also mounting the tracks downwards would be optimal imho, to void shorts from dropped metal (unlikely, but it could happen) and dust build-up causing increased resistance, possible charring of the dust particles and eventually shorts/fire.
If you put the pogo pins right in the center of rotation for both axis (would be between the rollers) you could minimize the movement and separation
EDIT: There are a handful of comments with the same thought... should've scrolled a bit further down
You may consider flexible PCBs 👍 You can get rid off cable mess this way.
You can also use magnetic rollers in order to stabilize both axis. You can install a thin steel bar to the other side of the alu profile. You just need to use two neodymium magnet rollers per axis. Another advantage of this method is easier and faster assembly.
Very nice project!!!
Great job!!!
You can use more pins so there is more contact surface like potentiometers or vacuum cleaners with retractable power cable...
Cool project. Having a battery included with the electronics next to the LED could maybe help solve some of these issues. In that case the tracks are only used for charging, and won't need to be 100% reliable. Batteries increase volume and weight though.
Steve! Add borders around your “super-track” contacts, this will stop your pogo pins from ever leaving them!
This is a super awesome soluition. I did think of something like this when you did the drag chains (which I did not like at all lol) but couldn't really see a viable way. I had not thought about just having a long simple pcb made up - that is ingenious
I hope you get it to work. This slop in the carriage/arm should not be there, the eccentric nuts should be able to tighten that out - but i guess there can be a trade-off , tighten it too much and it gets less buttery smooth to move.
Using a capacitor here is a very good idea, though beware that when loosing connection and they begin to discharge, they can cause a very large voltage spike if not properly accounted for and protected. Most lights and electronics can handle this for a short while, but it can mean the difference between an average electronics lifespan being measured in months or in decades, keep up the improvements though!!
This is by far the coolest project. I'm hoping to make my Diyson express soon. And I'm planning to also add one more carriage. It will mount a 2020 extrusion to the side of my desk, and allow the whole thing assembly to move along a track on my desk.
Awesome! I'm working on a hardware booster kit for the Express as we speak so keep an eye out for that in the coming days :)
I got part way through the last upload and then you removed it as I was watching it 😂
Haha sorry about that, I'm bad at this
and I was wondering why the comments didn't load :D
@@StevenBennettMakes No need for apologies, you've got a loyal fan base that loves your work. I ended up watching the whole video again. Anyway. Keep up the good work, can't wait to see what's next
@@G2bb101 You rock, thank you!
Can't wait to see this type of tech to make it into a 3d printer for cables setup
At first I thought that printing and gluing in pin separator channel insert to keep the pins in place would work, but then I realized that the wobble would still be there and that would bend or break the pins. The quickest idea I could come up with for the wobble would be to use 8 Exoslides. The DIY solution might be to add a third v-wheel axis opposite the Supertrack. Possibly thinning the wheels on each side to they ride at the bottom of the channel instead of just the v-groove.
Great work. Looking forward to the next upload.
I have to admit that my interest in building one for myself dropped when the drag chains were added initially.
These new supertracks seem promising and I am (selfishly) hoping you can make them work, so I don't have to "settle" for the DIYson Express 😆
Regardless, super great work, and clean editing, I love these build log videos!
Been looking forward this episode for some time :) really cool stuff.
The supertrack idea is great. I wonder how clean/conductive the copper will remain after a few months? When you raise the horizontal arm, can you see where the majority of the flex (that separates the pogo pins from the track) is coming from? The CNC block may not flex as much or maybe its the pins holding the rollers that are flexing or maybe the tension on the rollers or the material the rollers are made from.
The way you are analysing each problem is so logical.... I love it. Looking forward to seeing the next video. Cheers
Lovely tracks! I think some improvement can be also done by having multiple pogo pins in parallel touching the same track
Everyone's talking about different pogo pins and pcb designs. But to me it looks like your carriage just isn't rigid enough. I'd try and make the arm more rigid to counter the moment, and thus the separation. This would completely negate the need to super engineer the pogo pins, and you'd be able to use the shorter throw ball pogos, which will cause less wear over time.
Also you could put the pcbs in the tracks facing each other, so the pogos would be perpendicular to both the vertical and horizontal rails. This would be a bit more cumbersome to design in, but would look much cleaner imo.
You can add pins on both sides of the carriage also adding a fourth roller will help with the twisting visible in the video.