Ask HN: How to improve PCB prototyping iteration time?
I'm building wearable tech, and the JLCPCB lead times are my biggest bottleneck for iteration. 3-4 weeks is way too long. Has anyone figured out a way to make this quicker? Strongly considering moving to Shenzen to build the prototype. Has anyone done this, and was it quicker/worth it?
96 comments
[ 3.2 ms ] story [ 135 ms ] threadThat seems like a profoundly personal decision that would have more ramifications than you expect.
Breadboarding should be possible, same with soldering together prototype PCBs as long as you're not concerned with wearing the prototype yourself. There's also (albeit expensive) circuit modelling/validation software you could try to use. I don't think many circuit designers have a workflow where they perpetually buy finished prototypes to check if they're working or not, though.
If you just need a lot of unique double-sided boards to play around with in "alpha" and you can wait longer for the "beta" boards from your supplier, you can use several different small scale processes to accomplish that, such as photoresist, milling or printing.
Do you need your prototypes to be on flex PCBs? If you can get the electronics ironed out on FR4 you're going to save a lot of time and money on prototype revisions.
If you're waiting for JLCPCB to assemble all your boards for you you're slowing your process way down. Pay for 2-3 day processing at PCBWay and then assemble in-house. If you can prototype on FR4, PCBWay can literally manufacture the boards overnight and DHL them to you before noon the next day.
It's been a couple decades since I was involved in the H/W business, but there are usually tricks you can use to assemble even very fine pin-pitch devices in your kitchen. Heat gun, microscope, etc. I'm guessing there are 100 YouTube videos on this. Assemblers will tell you it needs ZYX special machine that only they have, but actually you can do it yourself, albeit slowly and with defects here and there.
Back when, there were decent PCB houses in the UK (up in the midlands typically, not near London). Perhaps they're all out of business now but you might go looking for one of them vs back and forth to Asia for prototypes.
Edit: a quick search suggests that there are many UK manufacturers still alive. E.g. [Forward, Newberry, Leicester, Stevenage]circuits.
For the parts with pins you can't reach with a probe, be sure to leave a way to test connectivity of the pins with a DMM. For the super tiny parts with the hidden pads, I like to have a series resistor that I can remove, and then use a DMM to check that you can find the ESD diodes on each pin of the package.
PCBWay can ship bare boards by noon the day after you order them if you pay the $99 to rush the order. Much cheaper than moving to Shenzhen.
Hm... my experience with JLCPCB is that I get 7-10 days iterations from sending my gerbers to getting my boards in SF Bay Area. Still long, but much shorter than your experience.
Is it shipping, PCB manufacturing or PCB assembly that takes so much time?
If you have to move, move to a place near the start of your DHL driver's route as that can make the difference between 9am and 4pm delivery (and that goes for RS too).
Don't move to Shenzhen. You don't have the social connections for that and it'd just give you a bunch of new problems.
[0] https://www.pcbtrain.co.uk/
You don't get many engineer hours for $100.
Of course everything is modulo working capital.
If you can afford the time, effort, and expense needed to move to China, maybe you should look into working with a local board shop that could provide more rapid turnaround.
The best way to improve your development time is by doing fewer iterations. Often easier said than done, but you should be designing for test and rework and thoroughly reviewing your designs before submission. You said you're making flex PCBs -- have you considered doing fast-turnaround prototypes on rigid PCB to make sure your electrical design is good? What kind of problems are you having that are forcing you to do extra prototyping runs?
Yep we've been optimising to reduce reruns, but the time spent doing so + the lead time is super frustrating. Standard engineering woes I suppose...
Hardware has an enormous set of things that cost money to surmount. Packaging, testing, certifications, etc. If chunks of $2K matter, you're never getting over the next hurdles.
Hardware isn't software. It isn't cheap.
I do a LOT of hardware and the difference between $200 and $2000 matters a lot to me.
Look at the BeatBuddy. It's one of the most successful projects to come out of the whole crowdfunded hardware thing.
Look at how much money he raised and then how much more money he needed in order to get it finished and how long it took.
If $2K vs $200 prototypes matter to you when building a hardware product, you're in deep, deep trouble.
This kind of situation is burned in to my psyche. [1]
One thing I learned is how one might bootstrap hardware on the cheap properly. You can’t launch until your product is perfect and you are actually ready to scale and you have to be willing to charge the right amount on your kickstarter.
So you have to know what you are doing, but I believe a cheap bootstrapped hardware product can be built and launched pre-funding and lead to success. Of course it’s difficult but we all know hardware is hard.
There’s a lot of creative people out there who can’t afford $2000 prototype batches and I want to see what they’re building. The point is, don’t panic-launch early. That’s an excellent way to fail in a very painful way. But that’s orthogonal to how much budget you have. Even a well funded project can launch too soon.
[1] http://www.tlalexander.com/business/
You'll be limited to single or double sided, without vias but considering its rapid prototyping, these are the compromises to be had.
You could always try CNCing prototypes but this is fraught with trouble. Fiber lasers offer the simplicity, repeatability and speed that normal laser cuters, but with metal.
[0] https://www.youtube.com/watch?v=PoYcjyghDx4
You should be able to buy quite a few local turns for the price it would cost for you to spend time in Shenzhen.
Unless you're Chinese is very good or you have a trusted local contact, you're not likely to benefit much from heading to Shenzhen in person.
1. Figure out if you can do fewer slow iterations. What's driving the need for a full PCBA run for each iteration? Might be able to split out to rigid assembly and flex for example.
2. Run more experiments in parallel. If you have multiple ideas or variations to test, design them all and fabricate them all. Flex antenna? Make like 30 parametric variations.
If design then becomes the bottleneck, then automate that next.
Then just glomp the flex down to the test PCB with some nice Samtec connectors (you can enough free samples for a small batch of test units even) or an edge connector on the stiffener.
If you’re bootstrapping, find a good quick turn board house and assemble the boards yourself. Get an lga-12 stencil and hot air tools, maybe a benchtop reflow oven.
I just don’t think this really is a better way of doing things.
Also, really take a look at JLC's component prices and compare them. Some of their markup, namely for ICs, is borderline criminal.
I find pricing really scattered everywhere. Often mouser and digikey, which others swear by, have worst pricing (and availability). But I think these places are intentionally high on some items to discourage production runs and keep it a prototyping service. Not sure why they care though.
The equipment isn't expensive and you don't need much labor. I bet a 2-3 week wait costs you a lot more than a part time assembly tech.
None of the engineers that I've talked to have been impressed with the current state of automated design or layout tools.
I just want an easy way to get into PCB design but there is just so much learning curve
If you want to learn how to design PCBs, I think it’s feasible to jump in and just start doing it (assuming you know how to design the circuit you want to put on it).
What Fritzing has going for it is that it is very simple.
Think of it like firing up Notepad when you just need to print a few words on to a sheet of paper, while everyone else is using Word.
I wouldn't recommend it for anything beyond maybe one maybe two DIPs and some surrounding passives and a few headers/connectors.
https://www.youtube.com/watch?v=BVhWh3AsXQs&list=PLy2022BX6E...
(That was a newer playlist than the videos I watched, and I think even that playlist is for an older version of KiCAD, but I found that Chris covered the material in a way that resonated with me, without an excessive amount of time-wasting along the way.)
I had taken a digital electronics course in college (I was Mech E, but we could take EE/CS courses if there was space) which is how I got the basics, but that was back when TTL/CMOS and PALs ruled, which isn't exactly the case anymore. Programming circuits in FPGAs is much newer (and by all accounts, easier/nicer).
BUT:
If you're just starting out/don't yet know which part of a soldering iron gets hot, I'd recommend starting with some Arduinos (clones), then probably do some ESP32 using the Arduino IDE (which is awful but easy to get started with), and that will give you more of a grounding (no pun intended) to set you on a good course for figuring out what you want your first PCB to do.
There are a thousand-plus Arduino makers on YouTube. Most are dreadfully slow-paced once you know the basics, but probably OK paced to start. I'd look at Andreas Spiess, the LED content from Dave's Garage, Adafruit's got a bunch of great content out there, Great Scott is good, Dronebot workshop is the Bob Ross of electronics in some ways, or 50 other channels. For more basic EE topics (including things like understanding decoupling caps, etc), Dave Jones at EEVBlog is great (as is the EEVBlog forum).
There really isn't these days. There's KiCad and LibrePCB, there are fabs that give you boards for pennies or even assemble the whole thing on a hobbyist's budget, there are cheap off-the-shelf dev boards that you can use as a base for your first designs, there are tools that verify your designs and can autoroute simple things, there are component libraries... I have just went through my first PCBs recently and it's incredible how approachable it became. Merely a decade ago we had to print, etch and drill a PCB ourselves for a group project as anything else would be way too costly (I worked on software and only watched though :P)
I usually get boards right first time now. But it took me 2 years to get there.
Having prototyped in both places I’ll make some arguments for. I’ll preface it’s only worth it if you’re beyond the limits of what American facilities can do and it's a step function in workflow and getting setup. There is an acceptable cultural tunnel vision in our field that developed in the 1980s for "how to do things" and hasn't changed much beyond "4 layer pcb on FR4 with surface mount components" - going against that grain requires an interdisciplinary mindset - as in "make a functioning circuit on a piece of toast" level of creativity[1].
If you’re building wearable tech there’s a strong chance you’ll need to make flex pcbs sooner or later. Those are comically cheap in China and stupid expensive stateside.
Especially when you start pushing the boundaries - there’s so much low hanging fruit for experimenting with your PCBs when you’re in the factory making them. Most US manufacturers will only let you use one color for a solder mask. In Shenzhen we pioneered using full RGB to print any graphic on your pcb back in 2017. Even on top of the chips themselves. It’s now pretty easy to source. This is literally just by being on the factory floor and saying “hey can you do step 5 before you do step 4? - we want to take the boards to this other factory across town first” And they say “sure”. Likewise if you want to mount your parts sideways or upside down to save space. Or say, take a literal sea shell through a copper PVD machine and mill away some traces and mount some chips. They do not care and will gladly take your money and make it happen.
One time we couldn’t find a single vendor in the US or Europe who would embed chips in the middle of pcb layers[2]. This was a weekend project for one of our Chinese vendors - who also had never done this but it sounded like fun so she said “no problem”.
Can get turnaround on prototype boards with assembly for free once you have a cm or just $50 if you don’t. One American vendor comically insisted one couldn’t mix flex and rigid boards for one of our designs for less than $10k. In China it cost me $80. Likewise we mounted chips to non-traditional media like credit cards with no sweat.
Any chip we wanted was available through HQB or TaoBao when Digikey was still backed up on Covid.
Test fixtures (the laser cut jigs that you program and test pcbs) are $100. Stateside they were half as useful and $2000.
You’re one blue Buick minivan ride from Guanzhou where all the garments in the world are made. Being at the intersection of these two cities is a strategic advantage.
Cost of living is cheap.
It will make you a better engineer by exposing you to the dirty business of manufacturing first hand. I'll go so bombastically hyperbolic and say not moving to Shenzhen as an EE is like not moving to Nashville as an aspiring country musician.
In China the saying goes “anything is possible, nothing is easy.” I’d mostly agree with this but also point out that price is completely-orthogonal-to-possible in China, absolutely not in the US, and straight up forbidden in Germany.
[1] https://talk.dallasmakerspace.org/t/breadboard-electronics/1...
[2] https://twitter.com/pmg/status/1248148053795540992/photo/1
p.s. If you enjoyed this comment, you may also enjoy my "So you want to start a factory?" reply on a post from a few days ago [3]
[3] https://news.ycombinator.com/item?id=40001222
I miss Naomi's tweets!
For some data, in 2019 about 3 million Americans went to China and the number detained for political reasons (and not working for the US government or journalism) was probably in the single digits?
In the US I've driven about 600,000 miles and been in one traffic accident. In China I've ridden about 30,000 miles and been in 6 traffic accidents.
Plenty of other things people have gone on at length about things that are just different risks in both places.
I figured I was over indexing but it’s nice to hear it directly from someone who knows. Thanks!
OSH Park does flex:
https://docs.oshpark.com/services/flex/
I feel like something might be slightly wrong here. Are you iterating because you haven't finalized the core innovative tech? Can you develop that first, on breadboards or with lab bench equipment?
Are you iterating because of minor aesthetic tweaks? Can you just make one PCB with LEDs in all the places you might want them, turn off the unneeded ones in software, then leave them out of the final BOM?
Are there just straight up mistakes, that could be fixed with slower iteration?
Maybe it's just that I've never quite made it to the level of career success where there's a budget and time for this kind of iteration, but I still think something could improve.
To answer your specific question, Shenzhen is a good place to work on hardware, but it is not like you just fly in, and people start building for you. Your own lead time (establishing yourself there) will likely be in weeks if not months. This is from my personal experience in Shenzhen and other hardware founders'.
If you want to build the prototype faster, identifying the bottleneck is a great start. Which takes most time? For up to medium complexity, PCB fab (the board itself) is the most time-consuming part. Anything more complex, it could be the SMT assembly. Each part of the entire process can be optimized: PCB fab can be done in-house quick (chemical etching, CNC machining, etc). Assembly can be optimized too (solder reflow oven, solder paste stencils, etc) on a small scale. I would estimate a batch of 5 boards of 50 SMT components each can be produced under 8 active engineer-hours. That's your entire prototype in a work day.
Do you have professional/academic training in electrical engineering? EE is no less complex than CS, and perhaps studying up the entire process will help you identify what's possible and what's the bottleneck.
Far less! That's about how many boards of similar complexity I could hand-stuff and reflow in an hour!
OTOH, most of the stuff can be and should be prototyped on standard rigid boards.
For standard 2L rigid boards:
If you work in 1-week cycles, you can usually send a PCB out on Thursday or Friday and have it in your hand on Monday or Tuesday to assemble and test. About $40-$100 per round.
If that's not fast enough, a domestic fab shop that can do 1 day turn on bare boards for not terrible pricing - about $150-$200 per round (unless you can will-call at the dock). 2 or 3 day with soldermask/silkscreen will probably be around $300-$500 per round.
Prices are rough estimates off the top of my memory from the shops I've used.