Bed Of Nails And Accuracy In PCB Manufacturing

A few days ago, we mentioned the new ARM-powered Teensy 3.0 project on Kickstarter. The creator, [Paul Stoffregen], decided to share the trials of building a test fixture along with a shocking comparison of the accuracy of different PCB manufacturers in an update to his Kickstarter.

Because [Paul]’s Teensy 3.0 has more IO pins than should be possible on such a small board, the test fixture to verify if a board is defective or not is fairly complex. To test each board, a Teensy is placed on dozens of spring-loaded contacts arranged like a bed of nails. From there, another Teensy (this time a Teensy 2.0) performs a few tests by cycling through all the pins with several patterns.

Because the spring-loaded contacts require rather precise drill holes in the PCB of his test fixture, [Paul] thought it would be neat to compare the accuracy of several board houses. In the title pic for this post (click to embiggen), [Paul] demonstrates the capabilities of OSH Park, Seeed Studio, and iTead Studio. The lesson here is probably going with a US company if quality drill work is a necessary requirement of your next project.

43 thoughts on “Bed Of Nails And Accuracy In PCB Manufacturing

  1. I’m living in The Netherlands, and the US-based services aren’t attractive to me due to shipment cost. I’ve sent a few designs to Itead, and the quality of the PCB’s is OK, not outstanding. 0402/0.5mm pitch QFN is on the edge of their possibilities regarding my quality standards. If anyone knows a price-comparable solution (10pcb’s 50x50mm for < EUR15) for Europe, please comment below!

      1. Hmmm. Doesn’t sound interesting at all unless you are in a real hurry. And to the prices you must add 21% VAT.

        Time= 5 days Default
        Number of layers=Double Side PCB
        PCB Size=50(mm)x 50(mm)
        Finish=Immersion Silver
        Quantity=1 PCB's 24.75 €
        Quantity=10 PCB's 67.50 €
        Price for delivery: 6€

        Still 9 times more expensive than Seeed or iTead. Or I’m missing something?

      2. @lionel, yes, the holes are plated, they do multilayer and hdi, so holes not plated is a no no.

        @S I’m always in a hurry with prototypes. I can’t wait weeks for testing a prototype. But it depends on the purpose of the board.

    1. Hi,
      I also live in the Netherlands and I order from OSH PARK.
      even when shipping to europe was 15USD is was cheap..
      The quality is great and now it is even easier to order.
      The only thing you need to look out is if you have more than one PCB and want to share the shipping costs.

  2. I dont understand why only board from itead has stopmask on the pins(which look more like vias on this board), also the holes are different diameter from those on other boards, I have used their service few times and i didnt have any problems like this. if all those three boards were the same design, why there are such differences in hole diameter and stopmask?

    1. Yeah, I had the same question about the soldermask over the holes. My files clearly had those areas exposed. iTead responded “the error was made by the factory when they manufactured since the hole size is a little small”.

      It’s not clear from Paul’s blog post, but the holes are supposed to be 13 mil.

      Seeed averaged 17 mil holes, and an average offset of 4 mils. All 10 boards passed with no shorts or breaks. Soldermask-to-copper registration was about 7 mils. Seed requires 8 mil drill-to-copper clearance, and rejected my first board which had only 2.5 mil drill-to-copper clearance.

      iTead averaged 12 mil holes but had an average offset of 7 mils, and coated all of them with silkscreen. 7 out of 10 of my boards had shorts, even though I’d paid for “100% E-Testing”. To their credit, they offered a full refund. Seed requires 6 mil drill-to-copper clearance, and rejected my first board which ad only 2.5 mil clearance.

      My service (OSHPark) averaged 12 mil holes and had an average offset of under 1 mil. Soldermask was correct, and had a soldermask to copper registration of 2 mils. The fab is allowed up to a 3 mil drill-to-copper clearance. In practice, I’ve never seen it more than 2 mils off.

      1. I updated the kickstarter page with this extra info.

        Sorry about Hack-a-Day’s summary, which implies I made those test boards. You’re the one who did all the hard work and you definitely deserve the credit for making exceptionally high quality boards available at a great price.

        Without your service, I would have been building this test fixture the old way, spending all day drilling holes through a 1/2 inch thick piece of plastic… and making many attempts before getting one close enough to actually use. I did that for Teensy 2.0 and it took all day.

        Thanks! :-)

      2. Sorry, Laen, I’m having a little trouble parsing the data here. You said “The fab is allowed up to a 3 mil drill-to-copper clearance.” By that do you mean specified drill location to the edge of the capture pad, in other words the annular ring? If so, that’s more than a factor of two better than most low-cost board houses, which is great. If that’s the case, why is the annular ring design rule 7 mils, on par with most other board houses? Etch tolerance? The other way I could interpret your statement is that the drill, in order to prevent a breakout, must fall no closer than 3 mils to the edge of the cover pad. For a 7 mil annular, that would imply a drill to copper registration of 4 mils or better, which is still very good for the price. Finally, it could mean that the maximum allowed registration error between drill and copper is 3 mils. All of those would imply a very good layer registration…any data on that? For the test structures, what was the hole aspect ratio, and were the measurements made at the entrance or exit pad? Thanks, and thanks for providing such a valuable service to the community!

  3. That pic is something OSH Park gave to one of their clients to show they are the best! Not only that but the Seed board isn’t obviously isn’t even drilled, so is hardly relevant to the accuracy of their drilling!
    I know we are not journalists but some reporting standards must be maintained.

      1. @ibuildrockets

        I don’t think it’s just the large holes. First, the background behind the boards appears to be tan, almost copper colored. It’s also the fact that the camera seems to be held directly over the seed board which gives it a direct view through the holes. The other two are at more of an angle to the camera giving the holes some shadow for contrast. This probably also makes the Seed board’s holes look bigger although I’m not sure they aren’t actually bigger as well.

      2. Hi, Paul here… the author of this article and creator of Teensy 3.0.

        Those images were provided by Laen. I also saw those 3 boards in person with my own eyes. Laen lives here in Portland and comes to the local Dorkbot meetup occasionally.

        They are a grid of very tiny holes. His manufacturer is indeed much better.

    1. I’ve used Laen’s OSHPark service for over a dozen 2-layer boards. The quality is consistently outstanding, and I’ve put several boards together at the limit of the fab’s design rules. 2-3 week turn-around with free domestic shipping is also nice. I believe he offers the best PCB prototype service around for US customers, not only in terms of quality, but also price.

  4. As an engineer that makes PCBs for a living I can say with certainty OSH Park makes quality boards. I have used several fab houses over the years and Laen does not skimp on the board specs ( yes gold plating finish is important!). You can try the PCB prototype “special” at other houses but the quality on the finish does not match up. Laen must have a good relationship with whoever sources his boards.

    1. I’ll second raynor’s comment on the gold finish-

      The quality of white-tin plating from overseas manufacturers is not-so-good. If you go with gold plating though, you can always get a super even finish on the pads.

      It also happens to be a useful tool for easy spotting of unsoldered pins/pads.

  5. > The lesson here is probably going with a US company

    The only lesson I see here is that, assuming the images are actually representative of each company’s production standards, OSH Park is better than Seeed and iTead.

    That’s three individual manufacturers. This evidence on its own isn’t enough to justify any claim about US companies in general versus the rest of the world in general.

    1. I’m sorry. I’ve got to say that although I like and agree with the thought that American manufacturers are better, this article doesn’t support that.
      I think it would be a useful thing if the Hack-a-day readers had a way to upload board pictures from different manufacturers for everyone to compare along with other info such as turnaround cost, etc. I’d like to buy local, but it is all currently crap-shoot as to what qyuality you can get.

    2. Yeah, totally. All over the world there are high quality fabs and low quality fabs. This isn’t a comparison between American fabs and fabs in other countries. It’s a comparison between three low-cost batching fab providers.

  6. You are all of course assuming that Seeed & iTead use the same PCB agents every time and that the process is the same.
    That is not how Seeed & ITead work, no matter what tehy may tell you.

  7. Dont understand why the itead picture shows such a crap board. Did someone make a mistake exporting and renaming the necessary gerber files for itead?
    I can only say i ordered once at itead and the quality/price ratio is unbeatable. The boards were excellent!

    This is typical american “if its foreign its bad” jabbering. They are into this sort of product comparison

    1. But did you use lots tiny vias, with 13 mil drill and only 7 mil annular ring? Or did you try making a bed-of-nails test fixture, with lots of larger drill holes using tiny annular rings requiring highly accurate drilling, like the one in my kickstarter update?

      Those lower quality services work pretty well for less demanding designs using lower specs.

      As for USA vs foreign, there might be really high quality PCBs outside the USA and domestic vendors with poor results. But in terms of low-cost prototypes, these 3 services are the main ones hobbyists are using lately. And coincidentally, the 2 Asian vendors don’t have nearly the tight tolerances of the USA-based manufacturer. That doesn’t means all international vendors are bad. It also doesn’t mean you shouldn’t use them at all.

      But if you’re making a PCB that requires high quality and especially accurate drilling, like the ones I made for this bed-of-nails test fixture, you should absolutely use Laen’s OSH Park service over the others.

      1. I used itead for some boards with qfn packages and lots of 0402 components( but using 24mil hole diameter for vias) and I had no problems with soldering any of those parts. I Am not saying that OSH Park service is bad(it looks like it is the best of those three), but for Europeans it is much more expensive then Itead and Seeed, and processing&shipping time is roughly the same.

    2. “i ordered once”

      There’s your problem, sample size = 1.

      You’re assuming that Chinese vendors are consistent, and you’re wrong, with Chinese people better be prepared to receive something different every time.

      Take it from a South American who buys a lot from China and very little from USA.

  8. Setting aside the cheap-batch-PCB fab comparison (which I believe is relevant, for these three specific services only of course) the actual article at Kickstarter is a really nice illustrated writeup on bed-of-nails testing. Well done Paul! (and even though I love reading them, don’t spend *too* much time on the articles, because I’m looking forward to that Teensy 3 board in the mail!)

  9. this is all misleading, as the specifications from the board house are an indication of the accuracy. If they want large annular rings, it’s because they are not good at drilling in the middle and want some extra space to make a mistake. however, the offset is typically consistent (i.e. it’s a registration problem) such that the relative location of all the holes will be correct. Also, there are some good offshore fab houses, you just need to pay more for PCB mfg. My last batch had 7.8mil holes with an annular ring width of 5mils (total diameter ~17.5mils). The boards look great, but it cost me $350 for qty 25.

  10. I’ve used all 3 services (osh and itead only once but at least a half dozen orders from seeed) as well as an PCBFABEXPRESS. I’ve never made a board with such small pads as the bed of nails (except for at pcbfabexpress), but i’ve had good results from all 3 board houses. I did notice that on seeed boards some vias would be coated by the solder mask more then other but it wasn’t enough of a difference to be a deal breaker.

    If I’m doing a volume project, i’ll go with either seeed or itead but if i’m doing some prototypes I’ll use OSH. i’d use them more if i could get just a regular green board :( I like supporting an american board house but unless you are making entire panels most american board houses charge an arm and a leg to make 1,5 or 10 boards. For the price of 1 board from an american board house i can get 30 boards from seeed or itead, but i can get them much faster from american board houses and i trust the quality much better. Usually my main motivations for picking one house over the other are cost and time. If i need the board next week I got with PCBFAB, if i can wait a month then i go with seeed or itead

  11. >The lesson here is probably going with a US company if quality drill work is a necessary requirement of your next project.

    I find this comment particularly disturbing. How many test boards were sourced from fab houses outside of the US? This blanket comment smacks of the “If it ain’t American, it ain’t shit” campaigns of the past. Don’t get me wrong, I’m all for regeneration of the US economy and encouraging the American public to buy US made goods, but remember even though you are a US based blog, you are hosted on the internet so making comments which can be taken as discriminatory or just plain rude isn’t very clever. BTW I’d be interested to know what % is the readership of HAD outside of the US. I expect it’s not insubstantial.

  12. Could I inquire if you could be fine with fee based posts?
    All I’d require is for you to submit content material
    for myself and just a website link or mention of my site.
    I can also compensate you.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.