As a professional hardware developer, i've been using the tag-connect for probably the last 10 years, and what i really dislikes it is the issue that i need the extra space for the clamping pins, so i need to use their back clamping accesories that always fails. i loved how you used the torque generated by the pins to lock the device in place, simple yet brillant!
Been doing similar design and hacking since the late 1970's and early 1980's, mostly with the Commodore 64, then microcontrollers with PICs and others. So, let's go back through the design. A original, B the large 3D version, and C the revised 3D print. First, study B, then C, then A. Recognize that similar to C being more simple but a better redesign of B, actually A is the simpler, better design overall. Go back to A, and make it more easily manufacturable than either B or C. Go to 4:44 or so in the video, to have the picture for reference. As it is, the wires are separated, stripped and soldered into the pin header cups. That's extremely tedious and time consuming, we're going to redesign it all and make it a breeze to assemble. First, make the cylinder 3 times taller. This will make it much less tedious to handle. It will still be mostly the bottom 1/3rd as the working tool head, the upper 2/3rds will be mostly about making it easier to assemble and attach to the tool. So start with a solid 3 high cylinder. Have your center hole, and from the top still recess a 2h hole for the screw head, so you can still use the same screw for the center, just with a hole for the scredriver to get down to it. No need to use a 3x longer screw. Now for the drill. Don't make a hole. Make a U shaped slot into the cylinder, for the drill bit to press into. Narrow the upper part of the U, so the drill snaps down into the bottom of the U in the correct place. Assembled in moments by pressing it in, instead of trying to force it through a hole. The cable can be made FAR more easily. Instead of separating the wires, use .1 header cable or similar, and just cut it straight across without separating it. Then use a razor blade on both sides, to shave away only the outer part of the insulation on each side. So you will end up with all conductors exposed and in place, with the remaining plastic between them separating the conductors. You can then tin all of the wire ends at once, just sliding your iron and solder down the row of conductors. On the upper 2/3rds of the cylinder, make an IDC cable recess, so the cable lays against the cylinder and becomes part of the cylinder form. This will hold all of the conductors in place at once for soldering to the pins, and makes it ridiculously faster compared to individual separate wires. And of course for the pins, you make U channels so they pop right in like the drill bit, with correct spacing around the cylinder side so they match the IDC cable conductor spacing. You pop in your pins, cut then scrape both sides of the cable, lay the cable against the cylinder for positioning, and then solder all of the connections to the pins in one operation. And remember, there are springs in the pogo pins. Forget the pogo pins, and use small nickel plated springs directly, it should be fine for low duty operations like this. There are a lot of other design ideas for this, but YT hates much longer comments so I'll have to put it in another eventually.. Adjust the cylinder for ridges in the upper part, and you can use zip ties around it to hold the cable tightly for strain relief, or glue etc to the cylinder. TLDR: You can use relatively standard 1980's style .1 IDC cable making techniques to make constructing the first tool version even easier than the later versions, and probably better overall. Or even use .05 IDC cable if you want to go smaller.
I've worked in Engineering Development for nearly all of my career and I'd say that it's pretty common to develop something that works and then go back and simplify and refine the design for the final product. You have to do the proto first to understand the limiting factors and assess your design concepts. It's an iterative process. Nice work, great video and your English is excellent. Thank you.
Notching the alignment pin slightly at the board depth would cause it to lock into place once the rotational force was applied by the pogo pins, removing the requirement to use a harder steel drill bit.
Yeah, me too. And most people here say that having pogo-pins is not a big deal. Maybe I should just slightly polish my first design, or maybe even make a small pcb to mount pogo-pins to it.
Small rigid PCB, locking alignment pin and pogo pins would be perfect. There's a lot of engineering in a pogo pin, and it helps to provide a reliable connection.
Very nicely made! My compliments. The original one left the most usable area on the board. And I don't really see an issue with it. So I personally would use that. Btw... the use of the mounting holes as part of the connector is brilliant!
Interesting to see. My suggestion is that you use those 5x Through-holes in a line. Adjust the diameter to make the (square cross-section) pins lightly loose. The pins are supported by plastic sleeves that hold them apart, but allows independent and minor flexional play. Reaction force can come from a guiding hole (and could position anywhere), or just use position near a border of the PCB for that, with a flat “hand” holding from the border to the bottom side.
Cool project. The locking principle is also used in woodworking. There the clamp, called a holdfast, goes into a hole in the workbench and is hit with a wooden mallet to lock it. To release it you hit the back of the holdfast.
For an improved version... I'd consider some spring-loaded pogo pins soldered into a custom pcb as the programmer adapter. I'd use 2x pcb's, one spaced apart from the other, to make it more stable.
@@Positive_Altitude You're welcome, I hope it goes well. Looking forward to seeing it 😊 I got my pogo pins from adafruit; they're about 1mm diameter, but you'd likely find similar on digikey=mouser. Look around, some have different heads for different purposes.
Oh, this is a really cool idea. Honestly this video have popped up a couple of times already, but just skipped thinking you discovered tag connect. But I was completely wrong. Good job!
What I’d do, is take your original design but solder the pogo-pins into a custom PCB, instead of gluing them into a 3D print. Maybe an extra thick PCB, or two stacked PCBs. I don’t think you’d have an issue of solder wetting inside the pogo-pin, but I’ve never used them before. It might also be an option to heat-set the pogo-pins into a 3D print, like you would with a threaded insert. What I was doing for my UPDI AVRs was to just mount some 0.1” header strips inside a clothes-peg. Two pins on one side, and one pin in the middle of the other side. It clamps onto the PCB, and since it’s three pins it is self-levelling, though not self-aligning. But I found that the peg ends were flexible enough to bend a bit even if the pins were slightly different lengths, so I’m pretty sure it would work fine for 4 pins too. Just not 5, but hey a lot of STM32 uses don’t actually need the reset pin anyhow. And those new CH32V003 MCUs only need three pins to program.
Awesome work, the hobby wires you were talking about are called DuPont terminal connectors. It was really nice to see all the work and attention to detail, and looking forward to more!
this is GENIUS. the only downside i could think of is if a flanged bolt is used to mount the board, it could short the power to ground. but then you just don't use a flanged bolt lol
Hello Dima, luckily I found your channel today and subscribed. I am impressed by your comprehensive knowledge and skill in so many related fields. You are an important role model for me. Thank you very much for this easy-to-understand explanation in a very pleasant English language. Best wishes from Vienna.
I've been dealing with this for 25 years. Just use a row of 5 holes, SIL spaced at 0.1". Then the connector you use a row of SIL pins, get the long reach pins about 1" long, and bend a couple of the pins a little bit. The bent pins hold it in place with the spring tension of the long pins themselves, plenty secure enough for program/verify operation. You can put the holes anywhere on the PCB, and even stack them as 2x3 etc they dont have to be one line. They also make ideal probe points later for multimeter or scope testing.
You can get the ground from the center screw and move the 3.3v to the alignment pin. The alignment pin could just be an unbent connector. This leaves three pads/spring peaces.
1. I would use the male pins with their housing. 2. Insert them in flexible holders in the body. 3. Prepare matching holes with traces to the required signals in the PCB around the center. 4. Compress the pins so they align with the holes, insert them then release them so they make contact and hold the fixture. This can be made smaller and can waste no PCB area.
If you have a number of through holes arranged in a circle (one for each pin in the programming pin) and then make a holder that is springy radially. The radius of the hole placement on the board is slightly larger or smaller than the radius of the placement of the pins in the connector. This gives the locking force when mounting the connector. One could have either slip-ring locking even harder.
really like the flexure approach. excellent use of creativity and 3d printer! looking at the design, my opinion (but i haven't used it) would be to put 1/2 of the pins on the bottom layer. this would reduce the square area required and be more stable. think alligator clip that uses the square corner of the PCB for alignment. Still, really impressed with a working prototype that looks clever and usefull!
The first thing to consider is if you even need that many pins. When I program STM32 boards all you really need is 3 pins, swdio, swclk and gnd. You don’t really need the others, so if you can cut your pin count down then it may make it even easier to design your programmer and take up even less board space. Another idea would be some sort of clamp or peg that goes onto the corner of the board, it would need the backside clear (although maybe not) or you could put contacts on the back side too and it shouldn’t need any sort of spring mechanism on the pins themselves meaning they don’t have a larger footprint than the pads since the spring would be in the clamp.
this is really cool! the “jamming a metal rod against the side of a hole” mechanism reminds me a lot of the old woodworking tool called a “hold-fast”, which is basically a ?-shaped rod of metal (literally shaped like a question mark, can definitely find on youtube) that you hammer to hold a piece of wood in place while you’re working on it. there’s maybe some inspiration there; i’m not sure.
It's an amazing idea, very very cool. I love designing stuff like this with 3D printing, each iteration brings you closer :-) Thanks for showing this project!
That's a really cool idea and process! It occurs to me the footprint looks to end up about the same, maybe slightly larger, than a standard JST SH connector.
Get rid of the pins, just tin a mm or 2mm of bare wire and model a cup for the wire to sit into on each spring. That would remove the bulk of each pin housing and allow you to bring the angle of the spring inwards a bit. If you’re concerned about longevity you could snip the pins down and solder wire to just the tip.
I know I'm late to the party, but I liked the first simple solution you used to validate the idea. What if you created just the "head" of the unit and used female connections in it to plug into? You could get rid of the bulk of the wire head-block that takes up a lot of the space. That way the wires are separated too.
Make a pcb small mimicking the pads and holes , make a plastic pad to cover the new pcb, with the connector will be your plug, in the ailment could also be plastic, and in the middle make a thread and put a plastic screw from the bottom of the pcb. If the pads are in the back of the pcb will be much better same idea , but the screw will also hold the pcb.
Think of a tiny, fancy, drink umbrella. Now, imagine the paper cover removed. You have a tiny umbrella skeleton. You will be 3D printing this. In the 3D print, print a thin (3mm) bolt for the umbrella's handle. Make pin holders on the ends of the bare clothe rib supports. I think this need not be anything other than a hole in the end of the ribs. The Dupont programming pins will go down, vertically, through these holder holes. Insert the bolt handle vertically through the PCB hole. On one of the holders, insert a non-electrical alignment pin. Align the pins to the pads. Align the alignment pin with its hole. Add a nylon nut to the bolt. Tighten down the nut, by hand (only,) to add pressure on the pins onto the pads. Now, the springy part sort of becomes the jig's body and the rest of the body is sort of eliminated. Hope that makes sense. If it works, please publicly post it (its print files) as open source, give me due credit and send me 10 copies. Hope it works. Hope it helps. Cheers.
Really innovative idea and thanks for sharing. My biggest concern would be how to do in-circuit programming if installed in some housing/enclosure? The point of using a mount hole as part of the connector means that would be “populated” with an actual mount screw on the board, if you are talking about a “production” or already installed unit you are trying to troubleshoot or upgrade. However, maybe this could actually “kill two birds with one stone”, if you rely upon _something_ acting as a nut or mounting stud underneath the connector. Engaging the center screw/bolt into something on the back of the pcb mount hole could actually help with alignment, correct contact depth & tension, and eliminate the (some) need for reliance upon connector torsion for tension. This could then be used both out of enclosure (where you 3d print some floating stud (mount post, maybe with heatset threaded insert, etc), or if mounted in an enclosure, you can remove the enclosure’s screw from the mounting location, and then the screw from the connector would use the existing mounting boss/stud in the enclosure for security.
I didn't take a close look at your drill bit, but some drill bits are centerless ground, which leaves microscopic grooves around the shank, try to spin your new rod in a drill and sanding it with a 320 sandpaper and see if it behaves differently. I am enjoying your channel, keep up the good work!
Doing it in a corner is a clever approach but adds a certain complexity. As @SystemsPlanet says, creating an small array in one edge can be more effective. You can design an upside down "F" shaped clamp and a mirrored contact mini PCB as a start point for a minimum clearance clamp.
I tried to make a design that does not require access to the board edge. My servomotor boards are mounted in a tight space so I wanted to be able to program them without removing them from the device.
Very clever, I have never done any 3D printing design so Im really impressed with the detail you can achieve. I still prefer the spring loaded pins to a bent metal pin, the spring loaded pins are not that hard to get considering the reliability of the connection.
Yeah, modern printers are amazing. There is great progress compared to 5 years ago. It doesn't require lots of experience and hours of tuning any more, even some cheap 300$ printers can produce excellent results right off the box. 3D printing is very simple now. I think I will get back to the spring loaded pins idea, since everyone likes it more :)
At first I thought that it was the screw, but the gap between the screw and it's hole edge is much bigger, so the screw could probably do that too, but it requires more tilt.
Very cool! I wonder if you could use the binding action of pins in through holes to make the five jumper wires just directly plug into a board. Of course through holes might not be something you want in your design... I had a project using a PIC MCU (also a 5 wire interface) where I'd left a 2.54mm header footprint, but instead of populating it I ended up just plugging the bare 5 pin header into the female connector of the PICKIT 2 programmer to make a "male pin PICKIT programmer", shoving this into the empty footprint, and torquing the whole thing sideways to cause all the pins to make binding contact with the holes. You would think this would be a terrible connection. I was expecting that one pin would not connect, or that it would produce intermittent or vibration sensitive connections. But it worked great, I think I got through probably 50 programming cycles with one failure due to contact intermittency. Of course this was a tiny application with a 5 second programming cycle. It was easy to hold the programmer steady and pushed to the side for that long. But I wonder if you could design a 3D printed housing for 5 jumpers which just positioned them to enter 5 through holes, and then pushed them all sideways? Or, you could design it so that the jumpers were angled inwards like yours already are, without bending the tips of the pins for surface mount pads: then the pins could enter holes that are more closely spaced than the 2.54mm pitch but of larger interior diameter at an angle, and maybe there could be a similar binding action. Maybe through holes are just too much real estate for this application since they use both sides of the board and all inner layers. Anyway, super neat implementation, really liked the reflection and commentary you did on the design process!
If I were you I would go with a clips design. Think a small clips that sandwiches the corner of board both front and back. It can be slightly spring-ly or not at all. All you need is some notches aligning with the holes. Maybe one side of the clips can have positive cylinder and the corresponding side can have negative cylinder.
from a technical point of view - quite well done. from a video publishing point of view - also well for the body - nice mix of presentation speech and intermixed object shots. i would just recommend to give a bit more sort of a teaser in a visual sort at the beginning. and maybe sort of a 5seconds logo and/or animation in form of a recognizeable opening scene for your channel to make regular watchers comfy... 🙂
4:40 going back to this, what if you used a two-piece core to mount the pogo pins instead? So you'd be able to just seat the pogo pins into the main body, clip in the rest of the body over them to lock them in-place, (with them half-exposed on the outside, like the Raspberry Pi Pico's exposed metal contacts) and then you can easily solder the wires onto the pogo pins from the outside. As for latching, I'm not sure if an FDM machine could do it, but what about two plastic-latches on extended stems? (alternatively, if you embedded a small magnet into the design then the "clip" could simply be placing another magnet on the opposite side of the board)
How about moving the contacts completely off the PCB by making a sacrificial break-away area on the opposite side of the screw hole? You could also make a double-sided break-away protrusion from that corner which fits an 8P SOIC test clip. Those are extremely cheap! I have received them free with programmers that only cost a couple dollars. Just make the “mouse bite” edges castellated with the whole thing being the same width as an SOIC-8.
This is probably the best solution. A commodity connector with compliant pins that has a very well defined footprint and is available from many vendors and will be long into the future. Mill a slot into the board near the edge for 4 of the contacts and ground it and then have 4 cut through plated holes on the edge of the board that match the other 4 contacts. As long as one is getting boards made at a fab where milling the slot has low cost there is no reason to shy away from it The clip is self supporting, a flat bit of ribbon can be soldered to the 4 contacts and you can add a 5th ground contact so a permanent connection can be made for an expansion interface or during development. Could be a DIL-8 test clip as well if bigger is better.
Personally i like ones you can buy. works reliably and easy to replace. Have used them many years. i mean tinkering is fine if you doing it for fun. But if you doing it for work, no company will accept board with DIY thing. Maybe few yers later another engineer needs to work on software patch. They have no time to mess around with "can be made out of scrap" designe. Worker hours are valuble, so if worker cost say 30 USD/hour, you can buy ready made thing for that money.
For the alignment pin you are 'jamming' via tilt, that's a lot like what fixture tables use in welding. To avoid needing to bend the wire, tilt the printed hole to add a preload! Very clever and happy accident though figuring that out.
Another idea would be is print the alignment pin but have a locking clip where the screw is like the expensive clip. Instead of bending the pin CA glue the wires and they will have some play like the spring loaded pins.
If you really want to use this kind of mechanism, I wouldn't make 5 individual springs, but instead focus on just a single spring on the drill bit side. If it doesn't work with a piece of steel rod, it's probably because the surface is too smooth for the tiny contact area - so maybe roughen it up a bit (or make a dent or hook) or just take another material that has more fiction. (And since we're dealing with 3D printing... Layer lines can have their benefits too if you know how to use them, you know? ;D) I would take a step back though and think about what the actual goal is: Press pins firmly against the contacts. Why not go with a clothespin-style design where you clamp it on from the edge and a single spring that puts pressure from the bottom and the top holds everything into place? If you're worried that not all pins make perfect contact, print everything in 90A TPU, leave tiny comb-like gaps in the pin arms that act like mini springs (keep in mind that you probably won't need more than 1mm action anyway) or incorporate o-rings / washers made of rubber, and if necessary increase the clamp pressure.
Сверло в качестве "ключа" - топ) Обычно ставлю B6B-PH разъемы (они есть с гнутыми в шахматном порядке пинами, чтобы без пайки держались, их можно не припаивать). Либо можно двухрядные с 1.27 мм шагом, но сами пады на плате сделать аналогично в шахматном порядке, чтобы с натяжением разъем вставлялся. И целесообразно добавлять либо SWO, либо банально UART для удобной отладки прошивки.
While I do like the design, I think you've pidgeon-holed yourself with your initial decisions, i.e. the choice of reusing the corner screw hole. IMO, the answer has been there all along - it's "DIY electronics" and PCBs, which means that one can make initial provisions that would make for a simpler and more useful end design. For example: you disregarded the edge-aligned/based type of a fastener on account of that your previous motor project would not be able to accomodate it. But _you_ control & _can_ acommodate _it_ (a different debug connector) in yoir future board revision! :) For example, for the board edge kind, a simple 2-3 milimeter deep inset region along one edge of the PCB would be enough to accomodate a lower "jaw" that would lay below the PCB. On the PCB - contacts. As part of the debug connector: ordinary pins glued together (assuming original commector casing, or alternate) and tensioned towards the jaw via any kind of various means. And bonus: keep the steel pin, just offset it a bit deeper into the PCB, past the contacts, as a stabilizer and position lock. Alternative, simpler solution: Instead of pads, use those cuttable round-holed IC row socket holders. Connector? Laterally spread-out regular connectors with some sort of locking mechanism (tiny metal jaws that lodge under the plastic of the IC row connectors?) All this avoids the use of the unreasonably expensive pogo pins... although, when used on the probing connector, cost isn't really that much of an issue.
I like your comment. Indeed I locked myself too much to some decisions. The initial idea of reusing the mounting hole looks great, but actually it creates some constraints to the board design: there should be a hole, the connector should be located where the mounting point is. I wanted to make the connector very small and flexible, so it could be used in any design and now I see that these decisions actually act against it. There were few eye opening comments like yours and some other. I will try to make the following versions quite different and hopefully much better.
@@Positive_Altitude Just don't take it too hard on yourself... it's a learning experience, and we've all been there. (For example, I'm currently quite self-screwed on account of choosing to go with classic Arduino (i.e. 5 V main rail) in my project, since all other sensors I was retrofitting were already 5V... but the project (and code) hit the ATMega328p's limits, and now I need to re-do tons of wiring, add level-shifters, re-do a case... Luckily, it's just a private project (and I probably applied plenty of feature creep that I find hard to quit). Alas - to the redrawing desk! :)
Have you considered TPU? I assume it'll interact poorly with the jumper housing but would make the spring effect far easier and more compact, the less stable jumper housings could be mitigated by leaning further into the bent jumper ends, bending them into channels such that the metal jumper heads provide the stability that the PETG did previously while still working as a single print and being more compact whilst providing the same spring effect. This would provide more give in the alignment pin too so you'd likely need to have that at an angle to act kinda like a pair of tweezers with that and the screw under compression to hold it on.
You came up with a bench dog! Woodworkers have used that self-locking mechanism for centuries. Look up bench dogs, and you'll see all the different designs. Awesome that you came across it accidentally.
In the old Nokia N8 mobile phone, some of the spring-loaded pins were built in. This solution is safer, more robust and, what is important, very small. I find it very interesting to be part of your tinkering. The way and not the destination is important.
Out of curiosity, have you considered taking your pin out pattern and making another custom PCB with the same pattern inverted, and connected to a usb/serial output? You could still align using the ground and drill bit that way, and you might be able to shave the other PCB closer to the other pads potentially eliminating the issue of interference with those closer components.
I got an idea to decrease the no-components area, I think that the main issue is the width and depth of the pin headers holders. The solution would be to give a longer and thinner tip to leave space. You could simply bend completely backwards, in a U shape, each pin header tip once inserted, as such the connector wouldn't get out as the folded tip would rest against the printed holder while leaving a thin 2-3mm long tip that would leave much clearance for components. Then there would be the risk of shorts with the tips' exposed sides, that could be covered by a short piece of thin heatshrink tube
Первая версия мне понравилась больше (как и многим в комментариях). Что если сделать расположение контактов не по кругу а в линию? На линию ниже, по краям, расположить 2 отверстия диаметром 1 мм для пинов который будут выравнивать и фиксировать весь разъем. Вместо сверла можно попробовать использовать PLS штырьки и весь разъем распаять на маленькой плате, без 3d печати. Просто не на всех печатных платах бывают отверстия 2 мм для крепления, иногда отверстий и вовсе нет, а футпринт там внушительный получается вместе с полигонами земляными.
Хм, мне нравится. Мб я слишком упоролся своей идеей с М2 дыркой. К тому-же на практике оказалось что крепить коннектор в углу не всегда удобно. Обычные PLS думаю слишком хлипко будет. Но я попробую наверное несколько вариантов такой идеи сделать с прочными штырями. Еще тогда уж на плату можно и JST разъем добавить, чтобы провода не паять. Спасибо за идею!
Why not make the pad a 2.54 pitch through hole. Then you can insert your Dupont wire at an angle, just like how you make your self locking mechanism, but lock in on every pin instead of a guided pin. And to securely hold them at an angle, just use rubber band. And I believe that people normally do this, as I have seen a lot of pcb with debug 2.54mm pitch connector not soldered. Vcc Rx Tx Gnd
Very neat! I feel like you should combine the first idea and the current iteration. I like the plastic springs but they're too bulky for this imo. How about creating your own sprung pins with the same wires you used in the second version but with small steel springs braced against either their original plastic end that you removed (maybe cut down in length) and a custom printed housing. Seems like it could be pretty close to the original in size
Yeah, that could be an option too. But most people in comments tell that having pogo-pins is not a big deal. So now I think that I overcomplicated everything and just need to slightly optimize the pogo-pin design :)
Can you use a nut on the screw to secure it on the backside? It looks like you would not need to debug/upload software too often, so using the nut should not be too much of an inconvenience? Maybe can even 3D print a nut with thin walls so it doesn't bump into alignment pin.
I have an idia for you to test! Make holes for pins and use plated VIAs. Use screw with a smaller diameter, and no aligner pin. Try to spring-load pins in horizontal direction. That's it! Now, your plug should be both aligned and hold tight on the screw. Remember, use screw pin with smaller diameter to have a good friction. Smooth pin or pin with bigger diameter woun't work!
How about using tiny nedimium magnet. It could hold the pins in place. You ca than move the connection to any part of board snd shoul also allow to better pin layot.
Do you know of any clips for QFN60, for example? It's hard to think of SOP8 chips which are actually worth attaching a debugger to. But if you do use one, then sure - sounds like the best idea, as long as you have enough clearance around the IC.
Very interesting video. What about those cool magnetic USB connectors? You could solder one end to the PCB, and just allow the magnetic USB to connect that way. You might need to bolster the strength of the magnets tho.
That's a fun idea, but sometimes there is a need for the low-profile board. For example these small boards are placed inside a servomotor, really close to moving parts (because there is a sensor on the board) so nothing on the board should be higher than 1-1.5mm. That's why ideally I don't want to solder anything to the board if it is possible. Cheers!
@@Positive_Altitude Hmm, what I meant was more like a surface mount soldering of the pads, with physical placement of the magnets to match the opposite USB connector. You would have to find a way to embed the magnets into the board. But whenever you want to do the diagnostic, your USB connector would then "stick" to the surface of the board where the magnets were located and the soldered pads for the PCB would already be in place.
Funny idea: A clamp with the connectors on the inside of the clamp touching the pads on the pcb. Alignment could be done with small notches in the pcb.
@@Positive_Altitude Ur right. When it's mounted you wouldn't have access. But with the design in the video it's the same isn't it? I mean you are using one of the mounting holes for alignment. I thought it was meant for updating the boards before they are mounted.
Problem I can see is the pads being so close to that grounded mounting hole opens your design up to shorts if the mounting fastener is the wrong size or shifts slightly or if a washer is used.
Yeah, that's simple and effective, but I thought that the board edge is not always accessible if the board is mounted. So I decided to make a different design.
I'm guessing the drill bit has a rougher surface than the rod because it gives more surface area for the drill chuck to bite onto and keep it from slipping
I would say that it's very fine surface. I also tried to make the steel rod surface more rough by sanding, but it didn't make a difference. Somehow the drill bit is just much more "grippy", you can clearly feel it with your hands when inserting the pin. There should be some cool physics :)
in the first design and the last design, why not use a wingnut with a nylon spacer if needed on the ground bolt? and then an entirely new idea is a clothespin style clamp that has the pins orientated on one side of the clothespin (when i say clothespin im talking about the 2-piece spring loaded pinch your finger style not the friction based one-piece dolly peg)
if the circiut is not powered you could have a "U" shaped PETG part that slides over the PCB corner, and you could secure that by reusing the screw hole. if the "U" shape is no good you could try a clamp design like a clothes peg.
Thank you! I was thinking about a similar design, but I want to be able to use it when the board is mounted. And it's not always possible with a design like this that requires some extra space around the board edge.
@@Positive_Altitudea lot of negative comments under this video, i dont get why people are like this, but i really like this idea and design process. Keep doing what you do!
As a professional hardware developer, i've been using the tag-connect for probably the last 10 years, and what i really dislikes it is the issue that i need the extra space for the clamping pins, so i need to use their back clamping accesories that always fails. i loved how you used the torque generated by the pins to lock the device in place, simple yet brillant!
tag-connect looks pretty awesome, is it's footprint supported by kicad? gonna check it out now
Thank you! Yeah, that work was inspired by tag-connect and it's a pleasure to read this from a person, who uses a tag-connect :)
Oh man, trying to develop using those back-clamping accessories is a horrendous experience!
As a person on TH-cam, I'm replying to this comment
@@MrValhallaKingThank you for your effort!
All those intermediate designs weren't useless if they led you to the final design =) Very cool project, gonna try to print it for myself!
For sure!
dude you already achived the perfect tiny connector that nobody is possible. i definitely like the first one more
For the metal rod, you can buy piano wire which is super hard. You cut it with a file or Dremel as it will damage most wire cutters.
Cool! That's good to know, I will try it. Thank you!
@@Positive_Altitude Or a shaft from a rivet ? Great channel. You may really like mine too!
A nail.
(No, not the finger kind, the hammering kind.)
i like your honesty and precision and critical thinking...great work Dima.
Been doing similar design and hacking since the late 1970's and early 1980's, mostly with the Commodore 64, then microcontrollers with PICs and others. So, let's go back through the design. A original, B the large 3D version, and C the revised 3D print.
First, study B, then C, then A. Recognize that similar to C being more simple but a better redesign of B, actually A is the simpler, better design overall. Go back to A, and make it more easily manufacturable than either B or C.
Go to 4:44 or so in the video, to have the picture for reference. As it is, the wires are separated, stripped and soldered into the pin header cups. That's extremely tedious and time consuming, we're going to redesign it all and make it a breeze to assemble.
First, make the cylinder 3 times taller. This will make it much less tedious to handle. It will still be mostly the bottom 1/3rd as the working tool head, the upper 2/3rds will be mostly about making it easier to assemble and attach to the tool.
So start with a solid 3 high cylinder. Have your center hole, and from the top still recess a 2h hole for the screw head, so you can still use the same screw for the center, just with a hole for the scredriver to get down to it. No need to use a 3x longer screw.
Now for the drill. Don't make a hole. Make a U shaped slot into the cylinder, for the drill bit to press into. Narrow the upper part of the U, so the drill snaps down into the bottom of the U in the correct place. Assembled in moments by pressing it in, instead of trying to force it through a hole.
The cable can be made FAR more easily. Instead of separating the wires, use .1 header cable or similar, and just cut it straight across without separating it. Then use a razor blade on both sides, to shave away only the outer part of the insulation on each side. So you will end up with all conductors exposed and in place, with the remaining plastic between them separating the conductors. You can then tin all of the wire ends at once, just sliding your iron and solder down the row of conductors.
On the upper 2/3rds of the cylinder, make an IDC cable recess, so the cable lays against the cylinder and becomes part of the cylinder form. This will hold all of the conductors in place at once for soldering to the pins, and makes it ridiculously faster compared to individual separate wires.
And of course for the pins, you make U channels so they pop right in like the drill bit, with correct spacing around the cylinder side so they match the IDC cable conductor spacing. You pop in your pins, cut then scrape both sides of the cable, lay the cable against the cylinder for positioning, and then solder all of the connections to the pins in one operation.
And remember, there are springs in the pogo pins. Forget the pogo pins, and use small nickel plated springs directly, it should be fine for low duty operations like this. There are a lot of other design ideas for this, but YT hates much longer comments so I'll have to put it in another eventually..
Adjust the cylinder for ridges in the upper part, and you can use zip ties around it to hold the cable tightly for strain relief, or glue etc to the cylinder.
TLDR: You can use relatively standard 1980's style .1 IDC cable making techniques to make constructing the first tool version even easier than the later versions, and probably better overall. Or even use .05 IDC cable if you want to go smaller.
I've worked in Engineering Development for nearly all of my career and I'd say that it's pretty common to develop something that works and then go back and simplify and refine the design for the final product. You have to do the proto first to understand the limiting factors and assess your design concepts. It's an iterative process.
Nice work, great video and your English is excellent. Thank you.
Stop Russia!
Personally, I liked the spring-loaded pin design better.
He liked it too, but wanted to design something highly accessible.
Notching the alignment pin slightly at the board depth would cause it to lock into place once the rotational force was applied by the pogo pins, removing the requirement to use a harder steel drill bit.
Yeah, me too. And most people here say that having pogo-pins is not a big deal. Maybe I should just slightly polish my first design, or maybe even make a small pcb to mount pogo-pins to it.
Small rigid PCB, locking alignment pin and pogo pins would be perfect. There's a lot of engineering in a pogo pin, and it helps to provide a reliable connection.
Same ! Smaller and probably safer. And this kind of contacts are esay to find and cheap as well.
Very nicely made! My compliments. The original one left the most usable area on the board. And I don't really see an issue with it. So I personally would use that. Btw... the use of the mounting holes as part of the connector is brilliant!
Thank you!
Interesting to see. My suggestion is that you use those 5x Through-holes in a line. Adjust the diameter to make the (square cross-section) pins lightly loose. The pins are supported by plastic sleeves that hold them apart, but allows independent and minor flexional play. Reaction force can come from a guiding hole (and could position anywhere), or just use position near a border of the PCB for that, with a flat “hand” holding from the border to the bottom side.
beautiful man. I like it much more than the large professional one, and your design doesn't take any space compared to the professional one!
Cool project. The locking principle is also used in woodworking. There the clamp, called a holdfast, goes into a hole in the workbench and is hit with a wooden mallet to lock it. To release it you hit the back of the holdfast.
For an improved version... I'd consider some spring-loaded pogo pins soldered into a custom pcb as the programmer adapter.
I'd use 2x pcb's, one spaced apart from the other, to make it more stable.
I'd agree here, great idea. I've done similar and it's super reliable.
Yeah, that is great. I think I will go this way for the next version. Thank you.
@@Positive_Altitude You're welcome, I hope it goes well. Looking forward to seeing it 😊
I got my pogo pins from adafruit; they're about 1mm diameter, but you'd likely find similar on digikey=mouser. Look around, some have different heads for different purposes.
I agree. Pogo pins are designed for this job, and also not too expensive or hard to get.
Oh, this is a really cool idea. Honestly this video have popped up a couple of times already, but just skipped thinking you discovered tag connect. But I was completely wrong. Good job!
What I’d do, is take your original design but solder the pogo-pins into a custom PCB, instead of gluing them into a 3D print. Maybe an extra thick PCB, or two stacked PCBs. I don’t think you’d have an issue of solder wetting inside the pogo-pin, but I’ve never used them before. It might also be an option to heat-set the pogo-pins into a 3D print, like you would with a threaded insert.
What I was doing for my UPDI AVRs was to just mount some 0.1” header strips inside a clothes-peg. Two pins on one side, and one pin in the middle of the other side. It clamps onto the PCB, and since it’s three pins it is self-levelling, though not self-aligning. But I found that the peg ends were flexible enough to bend a bit even if the pins were slightly different lengths, so I’m pretty sure it would work fine for 4 pins too. Just not 5, but hey a lot of STM32 uses don’t actually need the reset pin anyhow. And those new CH32V003 MCUs only need three pins to program.
Awesome work, the hobby wires you were talking about are called DuPont terminal connectors. It was really nice to see all the work and attention to detail, and looking forward to more!
this is GENIUS. the only downside i could think of is if a flanged bolt is used to mount the board, it could short the power to ground. but then you just don't use a flanged bolt lol
This is a really good idea. It was nice to see your process as you pursue it.
Hello Dima, luckily I found your channel today and subscribed.
I am impressed by your comprehensive knowledge and skill in so many related fields. You are an important role model for me. Thank you very much for this easy-to-understand explanation in a very pleasant English language. Best wishes from Vienna.
I liked the gray/metal prototype a lot better.
I've used an array of sim card connectors to connect boards by pressing them together.
Wow, that's smart, I like it, thanks!
smart, ;)
I've been dealing with this for 25 years.
Just use a row of 5 holes, SIL spaced at 0.1".
Then the connector you use a row of SIL pins, get the long reach pins about 1" long, and bend a couple of the pins a little bit.
The bent pins hold it in place with the spring tension of the long pins themselves, plenty secure enough for program/verify operation.
You can put the holes anywhere on the PCB, and even stack them as 2x3 etc they dont have to be one line. They also make ideal probe points later for multimeter or scope testing.
You can get the ground from the center screw and move the 3.3v to the alignment pin. The alignment pin could just be an unbent connector. This leaves three pads/spring peaces.
1. I would use the male pins with their housing.
2. Insert them in flexible holders in the body.
3. Prepare matching holes with traces to the required signals in the PCB around the center.
4. Compress the pins so they align with the holes, insert them then release them so they make contact and hold the fixture.
This can be made smaller and can waste no PCB area.
Great design, I can relate the locking trick as welding table clamps which has a rounded plug to attach clamp
It's an old technique predates even that. The method has been used as clamps called "holdfasts" since at least 1st century AD.
@@b5a5m5 that's really cool
Yeah, I didn't know about it :) That's cool.
If you have a number of through holes arranged in a circle (one for each pin in the programming pin) and then make a holder that is springy radially. The radius of the hole placement on the board is slightly larger or smaller than the radius of the placement of the pins in the connector. This gives the locking force when mounting the connector. One could have either slip-ring locking even harder.
really like the flexure approach. excellent use of creativity and 3d printer! looking at the design, my opinion (but i haven't used it) would be to put 1/2 of the pins on the bottom layer. this would reduce the square area required and be more stable. think alligator clip that uses the square corner of the PCB for alignment. Still, really impressed with a working prototype that looks clever and usefull!
The first thing to consider is if you even need that many pins. When I program STM32 boards all you really need is 3 pins, swdio, swclk and gnd. You don’t really need the others, so if you can cut your pin count down then it may make it even easier to design your programmer and take up even less board space.
Another idea would be some sort of clamp or peg that goes onto the corner of the board, it would need the backside clear (although maybe not) or you could put contacts on the back side too and it shouldn’t need any sort of spring mechanism on the pins themselves meaning they don’t have a larger footprint than the pads since the spring would be in the clamp.
this is really cool! the “jamming a metal rod against the side of a hole” mechanism reminds me a lot of the old woodworking tool called a “hold-fast”, which is basically a ?-shaped rod of metal (literally shaped like a question mark, can definitely find on youtube) that you hammer to hold a piece of wood in place while you’re working on it. there’s maybe some inspiration there; i’m not sure.
It's an amazing idea, very very cool. I love designing stuff like this with 3D printing, each iteration brings you closer :-) Thanks for showing this project!
Thank you, I'm glad that you enjoyed it :)
That's a really cool idea and process! It occurs to me the footprint looks to end up about the same, maybe slightly larger, than a standard JST SH connector.
Get rid of the pins, just tin a mm or 2mm of bare wire and model a cup for the wire to sit into on each spring. That would remove the bulk of each pin housing and allow you to bring the angle of the spring inwards a bit. If you’re concerned about longevity you could snip the pins down and solder wire to just the tip.
I know I'm late to the party, but I liked the first simple solution you used to validate the idea.
What if you created just the "head" of the unit and used female connections in it to plug into?
You could get rid of the bulk of the wire head-block that takes up a lot of the space.
That way the wires are separated too.
Make a pcb small mimicking the pads and holes , make a plastic pad to cover the new pcb, with the connector will be your plug, in the ailment could also be plastic, and in the middle make a thread and put a plastic screw from the bottom of the pcb. If the pads are in the back of the pcb will be much better same idea , but the screw will also hold the pcb.
Think of a tiny, fancy, drink umbrella. Now, imagine the paper cover removed. You have a tiny umbrella skeleton. You will be 3D printing this. In the 3D print, print a thin (3mm) bolt for the umbrella's handle. Make pin holders on the ends of the bare clothe rib supports. I think this need not be anything other than a hole in the end of the ribs. The Dupont programming pins will go down, vertically, through these holder holes. Insert the bolt handle vertically through the PCB hole. On one of the holders, insert a non-electrical alignment pin. Align the pins to the pads. Align the alignment pin with its hole. Add a nylon nut to the bolt. Tighten down the nut, by hand (only,) to add pressure on the pins onto the pads. Now, the springy part sort of becomes the jig's body and the rest of the body is sort of eliminated. Hope that makes sense. If it works, please publicly post it (its print files) as open source, give me due credit and send me 10 copies. Hope it works. Hope it helps. Cheers.
Really innovative idea and thanks for sharing. My biggest concern would be how to do in-circuit programming if installed in some housing/enclosure? The point of using a mount hole as part of the connector means that would be “populated” with an actual mount screw on the board, if you are talking about a “production” or already installed unit you are trying to troubleshoot or upgrade.
However, maybe this could actually “kill two birds with one stone”, if you rely upon _something_ acting as a nut or mounting stud underneath the connector. Engaging the center screw/bolt into something on the back of the pcb mount hole could actually help with alignment, correct contact depth & tension, and eliminate the (some) need for reliance upon connector torsion for tension.
This could then be used both out of enclosure (where you 3d print some floating stud (mount post, maybe with heatset threaded insert, etc), or if mounted in an enclosure, you can remove the enclosure’s screw from the mounting location, and then the screw from the connector would use the existing mounting boss/stud in the enclosure for security.
I didn't take a close look at your drill bit, but some drill bits are centerless ground, which leaves microscopic grooves around the shank, try to spin your new rod in a drill and sanding it with a 320 sandpaper and see if it behaves differently.
I am enjoying your channel, keep up the good work!
Doing it in a corner is a clever approach but adds a certain complexity. As @SystemsPlanet says, creating an small array in one edge can be more effective. You can design an upside down "F" shaped clamp and a mirrored contact mini PCB as a start point for a minimum clearance clamp.
I tried to make a design that does not require access to the board edge. My servomotor boards are mounted in a tight space so I wanted to be able to program them without removing them from the device.
Very clever, I have never done any 3D printing design so Im really impressed with the detail you can achieve. I still prefer the spring loaded pins to a bent metal pin, the spring loaded pins are not that hard to get considering the reliability of the connection.
Yeah, modern printers are amazing. There is great progress compared to 5 years ago. It doesn't require lots of experience and hours of tuning any more, even some cheap 300$ printers can produce excellent results right off the box. 3D printing is very simple now.
I think I will get back to the spring loaded pins idea, since everyone likes it more :)
is it the drill bit or the sharp edges of the screw gripping thread the pcb from inside as its angled also?
At first I thought that it was the screw, but the gap between the screw and it's hole edge is much bigger, so the screw could probably do that too, but it requires more tilt.
Lovely idea sir, thank you for sharing!
I like your design well done.
Very cool! I wonder if you could use the binding action of pins in through holes to make the five jumper wires just directly plug into a board. Of course through holes might not be something you want in your design...
I had a project using a PIC MCU (also a 5 wire interface) where I'd left a 2.54mm header footprint, but instead of populating it I ended up just plugging the bare 5 pin header into the female connector of the PICKIT 2 programmer to make a "male pin PICKIT programmer", shoving this into the empty footprint, and torquing the whole thing sideways to cause all the pins to make binding contact with the holes.
You would think this would be a terrible connection. I was expecting that one pin would not connect, or that it would produce intermittent or vibration sensitive connections. But it worked great, I think I got through probably 50 programming cycles with one failure due to contact intermittency.
Of course this was a tiny application with a 5 second programming cycle. It was easy to hold the programmer steady and pushed to the side for that long. But I wonder if you could design a 3D printed housing for 5 jumpers which just positioned them to enter 5 through holes, and then pushed them all sideways? Or, you could design it so that the jumpers were angled inwards like yours already are, without bending the tips of the pins for surface mount pads: then the pins could enter holes that are more closely spaced than the 2.54mm pitch but of larger interior diameter at an angle, and maybe there could be a similar binding action.
Maybe through holes are just too much real estate for this application since they use both sides of the board and all inner layers. Anyway, super neat implementation, really liked the reflection and commentary you did on the design process!
Very cool. I respect you for sharing it.
If I were you I would go with a clips design. Think a small clips that sandwiches the corner of board both front and back. It can be slightly spring-ly or not at all. All you need is some notches aligning with the holes. Maybe one side of the clips can have positive cylinder and the corresponding side can have negative cylinder.
Nice to follow your thought process...
I would love the files for your original design with the pogo pins. It is so simple and perfect
from a technical point of view - quite well done.
from a video publishing point of view - also well for the body - nice mix of presentation speech and intermixed object shots.
i would just recommend to give a bit more sort of a teaser in a visual sort at the beginning.
and maybe sort of a 5seconds logo and/or animation in form of a recognizeable opening scene for your channel to make regular watchers comfy... 🙂
Thank you for the feedback. Cheers :)
you could put a little ledge on primary pin or alignment pin
4:40 going back to this, what if you used a two-piece core to mount the pogo pins instead? So you'd be able to just seat the pogo pins into the main body, clip in the rest of the body over them to lock them in-place, (with them half-exposed on the outside, like the Raspberry Pi Pico's exposed metal contacts) and then you can easily solder the wires onto the pogo pins from the outside. As for latching, I'm not sure if an FDM machine could do it, but what about two plastic-latches on extended stems? (alternatively, if you embedded a small magnet into the design then the "clip" could simply be placing another magnet on the opposite side of the board)
Clever point and informative tutorial!
How about moving the contacts completely off the PCB by making a sacrificial break-away area on the opposite side of the screw hole? You could also make a double-sided break-away protrusion from that corner which fits an 8P SOIC test clip. Those are extremely cheap! I have received them free with programmers that only cost a couple dollars. Just make the “mouse bite” edges castellated with the whole thing being the same width as an SOIC-8.
This is probably the best solution. A commodity connector with compliant pins that has a very well defined footprint and is available from many vendors and will be long into the future.
Mill a slot into the board near the edge for 4 of the contacts and ground it and then have 4 cut through plated holes on the edge of the board that match the other 4 contacts. As long as one is getting boards made at a fab where milling the slot has low cost there is no reason to shy away from it The clip is self supporting, a flat bit of ribbon can be soldered to the 4 contacts and you can add a 5th ground contact so a permanent connection can be made for an expansion interface or during development.
Could be a DIL-8 test clip as well if bigger is better.
Personally i like ones you can buy. works reliably and easy to replace. Have used them many years.
i mean tinkering is fine if you doing it for fun. But if you doing it for work, no company will accept board with DIY thing. Maybe few yers later another engineer needs to work on software patch. They have no time to mess around with "can be made out of scrap" designe. Worker hours are valuble, so if worker cost say 30 USD/hour, you can buy ready made thing for that money.
For the alignment pin you are 'jamming' via tilt, that's a lot like what fixture tables use in welding. To avoid needing to bend the wire, tilt the printed hole to add a preload! Very clever and happy accident though figuring that out.
Good point, thanks!
Another idea would be is print the alignment pin but have a locking clip where the screw is like the expensive clip. Instead of bending the pin CA glue the wires and they will have some play like the spring loaded pins.
Can the courtyard (of no-components) zone can be reduced by putting the pads near the edge of the board?
I would, the screw hole is ground and then 2 contacts on each side of the corner, Need thin tracks to route out from the corner but nothing crazy.
Nice content!
If you really want to use this kind of mechanism, I wouldn't make 5 individual springs, but instead focus on just a single spring on the drill bit side. If it doesn't work with a piece of steel rod, it's probably because the surface is too smooth for the tiny contact area - so maybe roughen it up a bit (or make a dent or hook) or just take another material that has more fiction. (And since we're dealing with 3D printing... Layer lines can have their benefits too if you know how to use them, you know? ;D)
I would take a step back though and think about what the actual goal is: Press pins firmly against the contacts. Why not go with a clothespin-style design where you clamp it on from the edge and a single spring that puts pressure from the bottom and the top holds everything into place? If you're worried that not all pins make perfect contact, print everything in 90A TPU, leave tiny comb-like gaps in the pin arms that act like mini springs (keep in mind that you probably won't need more than 1mm action anyway) or incorporate o-rings / washers made of rubber, and if necessary increase the clamp pressure.
Сверло в качестве "ключа" - топ) Обычно ставлю B6B-PH разъемы (они есть с гнутыми в шахматном порядке пинами, чтобы без пайки держались, их можно не припаивать). Либо можно двухрядные с 1.27 мм шагом, но сами пады на плате сделать аналогично в шахматном порядке, чтобы с натяжением разъем вставлялся. И целесообразно добавлять либо SWO, либо банально UART для удобной отладки прошивки.
Хитро) я не думал про такой вариант. Наверное идеально, если есть место на плате под сквозные отверстия, спасибо!
While I do like the design, I think you've pidgeon-holed yourself with your initial decisions, i.e. the choice of reusing the corner screw hole. IMO, the answer has been there all along - it's "DIY electronics" and PCBs, which means that one can make initial provisions that would make for a simpler and more useful end design.
For example: you disregarded the edge-aligned/based type of a fastener on account of that your previous motor project would not be able to accomodate it. But _you_ control & _can_ acommodate _it_ (a different debug connector) in yoir future board revision! :)
For example, for the board edge kind, a simple 2-3 milimeter deep inset region along one edge of the PCB would be enough to accomodate a lower "jaw" that would lay below the PCB. On the PCB - contacts. As part of the debug connector: ordinary pins glued together (assuming original commector casing, or alternate) and tensioned towards the jaw via any kind of various means. And bonus: keep the steel pin, just offset it a bit deeper into the PCB, past the contacts, as a stabilizer and position lock.
Alternative, simpler solution:
Instead of pads, use those cuttable round-holed IC row socket holders. Connector? Laterally spread-out regular connectors with some sort of locking mechanism (tiny metal jaws that lodge under the plastic of the IC row connectors?)
All this avoids the use of the unreasonably expensive pogo pins... although, when used on the probing connector, cost isn't really that much of an issue.
I like your comment. Indeed I locked myself too much to some decisions. The initial idea of reusing the mounting hole looks great, but actually it creates some constraints to the board design: there should be a hole, the connector should be located where the mounting point is. I wanted to make the connector very small and flexible, so it could be used in any design and now I see that these decisions actually act against it. There were few eye opening comments like yours and some other. I will try to make the following versions quite different and hopefully much better.
@@Positive_Altitude Just don't take it too hard on yourself... it's a learning experience, and we've all been there. (For example, I'm currently quite self-screwed on account of choosing to go with classic Arduino (i.e. 5 V main rail) in my project, since all other sensors I was retrofitting were already 5V... but the project (and code) hit the ATMega328p's limits, and now I need to re-do tons of wiring, add level-shifters, re-do a case... Luckily, it's just a private project (and I probably applied plenty of feature creep that I find hard to quit). Alas - to the redrawing desk! :)
Awesome solution!
Thanks!
Where are the 2 minute paper fans here... W"hat a great time to be alive" 🙋🏾♂️🙋🏾♂️ 8:35
If you like the radial pin pad lay out.
Why not design and print a button style connector.
Interesting video
Thank you. What is the "button style connector?"
Have you considered TPU? I assume it'll interact poorly with the jumper housing but would make the spring effect far easier and more compact, the less stable jumper housings could be mitigated by leaning further into the bent jumper ends, bending them into channels such that the metal jumper heads provide the stability that the PETG did previously while still working as a single print and being more compact whilst providing the same spring effect. This would provide more give in the alignment pin too so you'd likely need to have that at an angle to act kinda like a pair of tweezers with that and the screw under compression to hold it on.
You came up with a bench dog! Woodworkers have used that self-locking mechanism for centuries. Look up bench dogs, and you'll see all the different designs. Awesome that you came across it accidentally.
In the old Nokia N8 mobile phone, some of the spring-loaded pins were built in. This solution is safer, more robust and, what is important, very small. I find it very interesting to be part of your tinkering. The way and not the destination is important.
Out of curiosity, have you considered taking your pin out pattern and making another custom PCB with the same pattern inverted, and connected to a usb/serial output? You could still align using the ground and drill bit that way, and you might be able to shave the other PCB closer to the other pads potentially eliminating the issue of interference with those closer components.
I got an idea to decrease the no-components area, I think that the main issue is the width and depth of the pin headers holders. The solution would be to give a longer and thinner tip to leave space. You could simply bend completely backwards, in a U shape, each pin header tip once inserted, as such the connector wouldn't get out as the folded tip would rest against the printed holder while leaving a thin 2-3mm long tip that would leave much clearance for components. Then there would be the risk of shorts with the tips' exposed sides, that could be covered by a short piece of thin heatshrink tube
Nice! But other people here convinced me that I should probably just use pogo-pins design :)
Very well , Dima !
Than you! Cheers :)
i definitely prefer to use off the shelf pogo pins than to bend probes by hand
Yeah, I see. Everyone likes it more :) I will get back to pogo-pin design
Первая версия мне понравилась больше (как и многим в комментариях). Что если сделать расположение контактов не по кругу а в линию? На линию ниже, по краям, расположить 2 отверстия диаметром 1 мм для пинов который будут выравнивать и фиксировать весь разъем. Вместо сверла можно попробовать использовать PLS штырьки и весь разъем распаять на маленькой плате, без 3d печати. Просто не на всех печатных платах бывают отверстия 2 мм для крепления, иногда отверстий и вовсе нет, а футпринт там внушительный получается вместе с полигонами земляными.
Хм, мне нравится. Мб я слишком упоролся своей идеей с М2 дыркой. К тому-же на практике оказалось что крепить коннектор в углу не всегда удобно. Обычные PLS думаю слишком хлипко будет. Но я попробую наверное несколько вариантов такой идеи сделать с прочными штырями. Еще тогда уж на плату можно и JST разъем добавить, чтобы провода не паять. Спасибо за идею!
cool current result. just 3 types parts to meachnically insert into the main 3D printed body.
Why not make the pad a 2.54 pitch through hole. Then you can insert your Dupont wire at an angle, just like how you make your self locking mechanism, but lock in on every pin instead of a guided pin. And to securely hold them at an angle, just use rubber band.
And I believe that people normally do this, as I have seen a lot of pcb with debug 2.54mm pitch connector not soldered. Vcc Rx Tx Gnd
Very neat! I feel like you should combine the first idea and the current iteration. I like the plastic springs but they're too bulky for this imo. How about creating your own sprung pins with the same wires you used in the second version but with small steel springs braced against either their original plastic end that you removed (maybe cut down in length) and a custom printed housing. Seems like it could be pretty close to the original in size
Yeah, that could be an option too. But most people in comments tell that having pogo-pins is not a big deal. So now I think that I overcomplicated everything and just need to slightly optimize the pogo-pin design :)
Can you use a nut on the screw to secure it on the backside? It looks like you would not need to debug/upload software too often, so using the nut should not be too much of an inconvenience? Maybe can even 3D print a nut with thin walls so it doesn't bump into alignment pin.
I have an idia for you to test!
Make holes for pins and use plated VIAs.
Use screw with a smaller diameter, and no aligner pin.
Try to spring-load pins in horizontal direction.
That's it! Now, your plug should be both aligned and hold tight on the screw.
Remember, use screw pin with smaller diameter to have a good friction. Smooth pin or pin with bigger diameter woun't work!
Oh, nice! That's an interesting one, I will think about it, thank you!
How about using tiny nedimium magnet. It could hold the pins in place. You ca than move the connection to any part of board snd shoul also allow to better pin layot.
Nice job Dima. My name is also Dima 😀
there are smd springloaded pins. you can just reflow solder them
I kind of liked the first version more - if those pogo pins last and make good connection all the time - I would prefer the small version
Talent / work / luck the triumvirate of successful venture.
...Or don't wate ANY board space and buy a SOP8 (or what every you need) Programming clip that just attaches directly to the IC.
Do you know of any clips for QFN60, for example? It's hard to think of SOP8 chips which are actually worth attaching a debugger to. But if you do use one, then sure - sounds like the best idea, as long as you have enough clearance around the IC.
thank you very much!
Very interesting video. What about those cool magnetic USB connectors? You could solder one end to the PCB, and just allow the magnetic USB to connect that way. You might need to bolster the strength of the magnets tho.
That's a fun idea, but sometimes there is a need for the low-profile board. For example these small boards are placed inside a servomotor, really close to moving parts (because there is a sensor on the board) so nothing on the board should be higher than 1-1.5mm. That's why ideally I don't want to solder anything to the board if it is possible. Cheers!
@@Positive_Altitude Hmm, what I meant was more like a surface mount soldering of the pads, with physical placement of the magnets to match the opposite USB connector. You would have to find a way to embed the magnets into the board. But whenever you want to do the diagnostic, your USB connector would then "stick" to the surface of the board where the magnets were located and the soldered pads for the PCB would already be in place.
Why not use pogo probe clip with 1.27mm pitch, cost way less, much cleaner and take less space in pcb.
I'm not impress by the final product, but by the journey was awesome! What a time to be alive!
Though:
can you get pcb SMA with pogo pins?
that way you could make the pattern in any shape with same setup.
great video. love the story elememnt!
Do you mean to make a PCB with pogo pins for the connector part? Yeah, I will try to do that, it should be cool 👍
Funny idea: A clamp with the connectors on the inside of the clamp touching the pads on the pcb. Alignment could be done with small notches in the pcb.
I was considering this option too, but in some boards I cannot get access to the board edge when it is mounted.
@@Positive_Altitude Ur right. When it's mounted you wouldn't have access. But with the design in the video it's the same isn't it? I mean you are using one of the mounting holes for alignment.
I thought it was meant for updating the boards before they are mounted.
Problem I can see is the pads being so close to that grounded mounting hole opens your design up to shorts if the mounting fastener is the wrong size or shifts slightly or if a washer is used.
Molodetc, Dima!
What about a cloth clip design? And 4 stright pin at the end and a hole?
Perhaps you could make it 2 parts with a v-torsionspring and essentially create a clamp to connect to pads on the front and backside of the pcb
Yeah, that's simple and effective, but I thought that the board edge is not always accessible if the board is mounted. So I decided to make a different design.
Thanks Dima!
:)
I'm guessing the drill bit has a rougher surface than the rod because it gives more surface area for the drill chuck to bite onto and keep it from slipping
I would say that it's very fine surface. I also tried to make the steel rod surface more rough by sanding, but it didn't make a difference. Somehow the drill bit is just much more "grippy", you can clearly feel it with your hands when inserting the pin. There should be some cool physics :)
in the first design and the last design, why not use a wingnut with a nylon spacer if needed on the ground bolt? and then an entirely new idea is a clothespin style clamp that has the pins orientated on one side of the clothespin (when i say clothespin im talking about the 2-piece spring loaded pinch your finger style not the friction based one-piece dolly peg)
use threaded rod...it will work for sure...because of higher friction
if the circiut is not powered you could have a "U" shaped PETG part that slides over the PCB corner, and you could secure that by reusing the screw hole. if the "U" shape is no good you could try a clamp design like a clothes peg.
Thank you! I was thinking about a similar design, but I want to be able to use it when the board is mounted. And it's not always possible with a design like this that requires some extra space around the board edge.
you could use a paperclip like design to clip on both side of the pcb and in a corner edge
I was considering this option too, but in some boards I cannot get access to the board edge when it is mounted.
Look for High Speed Steel (HSS) for the 1mm rod.
4:15 dude just figured out how a variety of shop/fixture-table clamps work.
:) yep, now I know about it
@@Positive_Altitudea lot of negative comments under this video, i dont get why people are like this, but i really like this idea and design process. Keep doing what you do!