Caution: JLC's .5oz inner layer may have caused issues for me.

[Rat_Patrol]

I'm going to leave my original post, but I'm going to edit with this:
Upon thinking about responses in this thread, I'm running under the assumption that JLC's 0.5oz inner layer is highly likely to have been causing my issues. While I am not positive, other than manufacturer of the PCBs, this is the only difference between my JLC orders and orders from other vendors. PCBs from JLC had a failure rate that I did not experience with other PCB vendors that use thicker inner copper.

I would also like to issue a formal apology to JLC. While I'm not sure that the copper thickness difference was causing my issues, I'll give them the benefit of the doubt as I've not had issues with their PCBs when I order small 2 sided boards. Hey JLC: thicker inner copper layer option would be nice!

Consider this a caution about making sure your design will fully function with 0.5oz inner copper if you use JLC for your PCBs.



Original post:

Over 3 orders, I'm having up to a 50% failure rate of their PCBs.

I have a 150x200mm 4 layer board that I was running some short production runs of. Think less than 100 boards.

For some reason, which I thought was on my end for over 3 weeks, the micro would keep rebooting randomly. I could not figure it out. I replaced and re-replaced every component on that board by hand over multiple of these faulty boards to no avail. I checked everything I could think of. I've spent more time staring at the scope than I have in a long time. Last thing I figured it could be was the PCB. In these 3 orders (2 small, 1 larger order of the same gerber files), I would just set aside the non-working PCBs for later diagnosis and re-work. Well, the last 3 weeks I have spent going over these boards, trying to figure it out as they were stacking up.

The boards that worked just worked fine. I thoroughly stress test the boards before shipping, and not one has come back for warranty. So when they work, they seem to work fine.

I got some fresh PCBs in from PCBWay (same gerbers), a company which I had used for years (more expensive than JLC), and they are so far working perfectly with zero issues at all; not one failed board. I'm not sure what is intermittently wrong with those JLC boards, but I'm done wasting time and money with JLC, at least for complicated boards.

[mikeselectricstuff]

Unless you can pin to down to an actual measurable PCB fault, I think it may be premature to blame the PCB, as it could be something like a marginal component tolerance or over-sensitivity to something which could come back to bite you later.
Random reboots don't sound like much a PCB issue to me.

[Mangozac]

I'm surprised by this because I use JLC for production and I know another local company who uses them for complex 4 layer boards. Neither of us have ever had an issue with production boards.

Do you order with random test or 100% test? I choose this based on the board complexity.

[lutkeveld]

I had that too with JLC, it was related to a via annular ring that was below their tolerance.

Nevertheless, they should have caught it in either the ordering process or during quality check.

Are you sure your design complies with their tolerances?

[Rat_Patrol]

I had that too with JLC, it was related to a via annular ring that was below their tolerance.

Nevertheless, they should have caught it in either the ordering process or during quality check.

Are you sure your design complies with their tolerances?

As far as I know my design was good for their tolerances.

ETA: Though I would suspect something like the problem you pinned down.

Frankly, I'm not interested trying to "fix" the issue with JLC anymore if PCBWay can give make me boards that work.

[NivagSwerdna]

How are you generating your Gerbers?  Which EDA tool? Which version?

[Rat_Patrol]

How are you generating your Gerbers?  Which EDA tool? Which version?

Interestingly, this board is from Sprint Layout.

PCB way has no issues with the gerbers.

[48X24X48X]

I had that too with JLC, it was related to a via annular ring that was below their tolerance.

Nevertheless, they should have caught it in either the ordering process or during quality check.

Are you sure your design complies with their tolerances?

As far as I know my design was good for their tolerances.

ETA: Though I would suspect something like the problem you pinned down.

Frankly, I'm not interested trying to "fix" the issue with JLC anymore if PCBWay can give make me boards that work.

They won't fix things when there's issues. Period.
I recently have a batch of 100 pieces of production board. On the board there are JST 2.5mm through hole connectors that I have used many times in other design and were produced by them. The pins can go through the pad holes all these times as I followed JST recommended size regardless whether they are prototype or production boards made by JLC. But, to my horror after mounting all boards with the SMD parts, I found out these through hole is having a hard time getting the pins in. They argued that these are "production boards", you must adjust the hole size. I said all my previous batches of the board work out just fine with the same footprint for both prototype and production boards, why it is any different this time around? They just don't want to take any responsibility or even giving any sort of apologies. I had a $12 module mounted on every piece of the board and had to sand down the pins in order to complete the work. My bad for not checking before doing the SMD assembly. But, like the few review I have read on PCBshopper.com, "when things go right, everything is rosy with them but when things go bad, they won't even bother to take their responsibility". That sums up JLCPCB.

[thm_w]

They won't fix things when there's issues. Period.
I recently have a batch of 100 pieces of production board. On the board there are JST 2.5mm through hole connectors that I have used many times in other design and were produced by them. The pins can go through the pad holes all these times as I followed JST recommended size regardless whether they are prototype or production boards made by JLC. But, to my horror after mounting all boards with the SMD parts, I found out these through hole is having a hard time getting the pins in. They argued that these are "production boards", you must adjust the hole size. I said all my previous batches of the board work out just fine with the same footprint for both prototype and production boards, why it is any different this time around? They just don't want to take any responsibility or even giving any sort of apologies. I had a $12 module mounted on every piece of the board and had to sand down the pins in order to complete the work. My bad for not checking before doing the SMD assembly. But, like the few review I have read on PCBshopper.com, "when things go right, everything is rosy with them but when things go bad, they won't even bother to take their responsibility". That sums up JLCPCB.

What hole size, for reference?
Maybe they adjusted it down or up to meet the drills they happen to use. But I'm surprised if its that far off.

[48X24X48X]

They won't fix things when there's issues. Period.
I recently have a batch of 100 pieces of production board. On the board there are JST 2.5mm through hole connectors that I have used many times in other design and were produced by them. The pins can go through the pad holes all these times as I followed JST recommended size regardless whether they are prototype or production boards made by JLC. But, to my horror after mounting all boards with the SMD parts, I found out these through hole is having a hard time getting the pins in. They argued that these are "production boards", you must adjust the hole size. I said all my previous batches of the board work out just fine with the same footprint for both prototype and production boards, why it is any different this time around? They just don't want to take any responsibility or even giving any sort of apologies. I had a $12 module mounted on every piece of the board and had to sand down the pins in order to complete the work. My bad for not checking before doing the SMD assembly. But, like the few review I have read on PCBshopper.com, "when things go right, everything is rosy with them but when things go bad, they won't even bother to take their responsibility". That sums up JLCPCB.

What hole size, for reference?
Maybe they adjusted it down or up to meet the drills they happen to use. But I'm surprised if its that far off.

It is 0.9 mm as recommended by JST on the XH series connector (for 3 pins and above). It just doesn't line up with all the bullshit they gave because I had production runs with them using the same exact footprint.

[wraper]

It is 0.9 mm as recommended by JST on the XH series connector (for 3 pins and above). It just doesn't line up with all the bullshit they gave because I had production runs with them using the same exact footprint.

But did you measure actual diameter?
EDIT: datasheet shows 0.9mm +0.1/- 0. So tolerance does not allow to go below 0.9mm which certainly can and most likely will happen due to tolerances. You should expect there will be 0.9mm drill and plating thickness, so by default you should expect something like 0.86mm for 1oz copper, and that is already out of spec. So regardless of how badly they made PCB, it's your fault as well.

[48X24X48X]

But they need to have consistency as well. It worked well for many batches of boards before that. All made by them. That needs explanation but they kept quiet and not referring at all to those orders when queried despite me giving them all the order numbers. This is for me, the "art of Tai Chi" by irresponsible manufacturer. I have had boards redo by PCBCart, JSDPCB whenever there's a mistake.

[wraper]

But they need to have consistency as well. It worked well for many batches of boards before that. All made by them. That needs explanation but they kept quiet and not referring at all to those orders when queried despite me giving them all the order numbers. This is for me, the "art of Tai Chi" by irresponsible manufacturer. I have had boards redo by PCBCart, JSDPCB whenever there's a mistake.

There is a such thing as tolerances. If you made a marginal design, don't blame the manufacturer when you received a batch unsuitable for you but within provided manufacturer tolerances. Even if previous batches were fine. I guess your complaint was like: "it was fine before, so it must be your fault", so they told you to sod off. While proper complaint should be: "hole diameter is 0.8mm, which it's out of spec of your published +/-0.xx mm tolerance".

[Mangozac]

EDIT: datasheet shows 0.9mm +0.1/- 0. So tolerance does not allow to go below 0.9mm which certainly can and most likely will happen due to tolerances. You should expect there will be 0.9mm drill and plating thickness, so by default you should expect something like 0.86mm for 1oz copper, and that is already out of spec.

Plated through hole size is the finished hole, so after plating. Still, there is no tolerance allowed for so does have the potential to be an issue.

[wraper]

Hole size Tolerance (Plated)   +0.13mm/-0.08mm   e.g. for the 1.00mm Plated hole, the finished hole size between 0.92mm to 1.13mm is acceptable.

https://jlcpcb.com/capabilities/Capabilities

[48X24X48X]

But their explanation was: "for production batch we don't do adjustment/compensation, for prototypes, we will do adjustment". So, they do play around with the holes diameter. My guess is as a large panel consist of probably other designs, they probably want to save some drills sizes and rounded mine. I will be more careful on the holes sizes after this with all PCB manufacturer.

[wraper]

EDIT: datasheet shows 0.9mm +0.1/- 0. So tolerance does not allow to go below 0.9mm which certainly can and most likely will happen due to tolerances. You should expect there will be 0.9mm drill and plating thickness, so by default you should expect something like 0.86mm for 1oz copper, and that is already out of spec.

Plated through hole size is the finished hole, so after plating. Still, there is no tolerance allowed for so does have the potential to be an issue.

But you should consider they likely have drills with something like 0.05 mm or even 0.1mm step. So expecting they will use 0.93mm drill for 0.9mm plated hole is quite optimistic.

[Mangozac]

But you should consider they likely have drills with something like 0.05 mm or even 0.1mm step. So expecting they will use 0.93mm drill for 0.9mm plated hole is quite optimistic.

Yes that's right - as you said he hasn't allowed for any hole size tolerance. I was just pointing out that you don't have to make further allowances for the plating, as this is included in their drill size selection tolerance.

[phil from seattle]

It is 0.9 mm as recommended by JST on the XH series connector (for 3 pins and above). It just doesn't line up with all the bullshit they gave because I had production runs with them using the same exact footprint.

But did you measure actual diameter?
EDIT: datasheet shows 0.9mm +0.1/- 0. So tolerance does not allow to go below 0.9mm which certainly can and most likely will happen due to tolerances. You should expect there will be 0.9mm drill and plating thickness, so by default you should expect something like 0.86mm for 1oz copper, and that is already out of spec. So regardless of how badly they made PCB, it's your fault as well.

My experinces with JLCPBC showed that they have a pretty heavy hand with through hole plating. I generally oversize my through holes by 10-15% just to be sure.  Other manufacturers are a little more sparing on TH plating. I'm not going to defend them but they are relatively predictable for me.  I can live with their shortcomings.

[jmelson]

Over 3 orders, I'm having up to a 50% failure rate of their PCBs.

First thing, before stuffing the board, put it under a microscope or strong magnifier, and look for the probe marks on the pads.  The flying probe tester leaves a TINY dimple in the center of each pad when it tests the board.  If you don't see probe marks, the board was NOT TESTED!  Yes they can have some horribly bored girl looking for obvious shorts and opens, but a bored human will certainly miss a few.  Flying probe testers are not perfect, but they reduce failures to the PPM level.

I have had excellent luck with PCBway, well it isn't actually luck, they do the testing.

Jon

[Mangozac]

First thing, before stuffing the board, put it under a microscope or strong magnifier, and look for the probe marks on the pads.  The flying probe tester leaves a TINY dimple in the center of each pad when it tests the board.  If you don't see probe marks, the board was NOT TESTED!

As I said above though, you have to pay extra for it (as an option) if you want 100% test. Otherwise they will just do random test of a few panels in the batch.

[JLCPCB Official]

Over 3 orders, I'm having up to a 50% failure rate of their PCBs.

I have a 150x200mm 4 layer board that I was running some short production runs of. Think less than 100 boards.

For some reason, which I thought was on my end for over 3 weeks, the micro would keep rebooting randomly. I could not figure it out. I replaced and re-replaced every component on that board by hand over multiple of these faulty boards to no avail. I checked everything I could think of. I've spent more time staring at the scope than I have in a long time. Last thing I figured it could be was the PCB. In these 3 orders (2 small, 1 larger order of the same gerber files), I would just set aside the non-working PCBs for later diagnosis and re-work. Well, the last 3 weeks I have spent going over these boards, trying to figure it out as they were stacking up.

The boards that worked just worked fine. I thoroughly stress test the boards before shipping, and not one has come back for warranty. So when they work, they seem to work fine.

I got some fresh PCBs in from PCBWay (same gerbers), a company which I had used for years (more expensive than JLC), and they are so far working perfectly with zero issues at all; not one failed board. I'm not sure what is intermittently wrong with those JLC boards, but I'm done wasting time and money with JLC, at least for complicated boards.

Dear sir, could you provide your order number to me. There must be some misunderstanding. JLCPCB is the biggest PCB prototype company. We are confident of quality control.

The fact:
-By 2020, we have 6 factories, all the production line with advanced equipment;
-File Review Before Production
-Multiple testing before shipping
-Strict management of factory staff and  office staff
-High-quality raw material
....
in those ways, JLCPCB can make the quality complaint rate to be 0.25%(In the 20,000 orders, there is less than 50 orders with quality complaints). We want to know the specific reason for your order problem, we promise if we make a mistake, you will be definitely compensated. Similarly, we hope to get a fair and open evaluation.

Please provide me your order number, or leave your contact, we will find out the real situation as soon as possible. This is what a responsible PCB company needs to do.

Thanks for all the attention.

[Mecanix]

This is what a responsible PCB company needs to do.

Thanks for all the attention.

Agree. Tell'm, JLC!
Disclaimer: I've ordered hundreds of pcbs from your fab house, 2L & 4L in al colors and shapes and none (literally none) came in bad. Always within tol.

[Ice-Tea]

Seems to me you're blaming (and shaming) JLC without any actual evidence.

In addition, you seem to discard the possibility there might be a problem in your gerbers or design based on the assumption JLC is to blame. That's a dangerous approach. I personally would not be shipping products unless I knew *exactly* what the issue was. If JLC didn't flag a problem with your design, perhaps you should run it through some other vendors online verification to see if they catch anything. Or double check your design rules and DRC.

Just moving on is asking for trouble IMHO.

[mikeselectricstuff]

Seems to me you're blaming (and shaming) JLC without any actual evidence.
 I personally would not be shipping products unless I knew *exactly* what the issue was.

This, x1000