AVR Freaks

Yes, I requested their turnkey service, in which I send them a BOM, and they would purchase the parts themselves. In my case, the BOM included the full MPN (MCP1703AT-3302E/CB) and a digikey part number for cross-checking (MCP1703AT-3302E/CBCT-ND).

Yes, that's what I would expect from a pcb house that I'd send my business to.

I'll review the boards and assembly service later (when this issue is resolved).

In the meantime, I got a reply from them. They've claimed that their supplier sent them the wrong part. I was given no further explanation of how that happened, or who the supplier was. They've also offered to send replacement parts, presumably for me to do the rework myself. So far, their response has been quite underwhelming.

- S

Another update. PCBWay has refunded the full price of the assembly and components. Kudos to them for taken responsibility in this way.

- S

ps: As requested by Ross, I'll post a review of the boards in this thread later today when I have a bit more spare time.

Of course :-)

Here's a quick review of their service. I had several different board orders done in a single batch. Some of the boards were PCB only, some were PCB plus assembly.

The quality of the bare boards was very good for all the orders. Alignment of layers was precise. I think my gerbers had a soldermask swell size of 0.1mm. I don't know if pcbway did any adjustment there, but even on the fine pitch (0.5mm) SMD pads, I couldn't find any soldermask overlapping the pad. And on those 0.5mm SMD pads, there was adequate soldermask between pads. For one of the orders (a 4 layer board), the bare boards came individually wrapped in bubble wrap. For another, the bare boards were put in a ziplock bag. For the others, the boards were stacked together in shrink wrap bubble wrap. There's some minor scratching on the surface of the boards, most noticeably on the boards placed in the ziplock bag, but I don't see any damaged soldermask or flaws that might affect solderability.

There was one board that had an obvious manufacturing problem - there was a gouge in the top copper underneath the soldermask. I'm guessing this happened during photoetching. Electrically, it's not an issue, as it's part of a ground pour, and there's no continuity break.

One of the orders was for HASL lead-free, but it looks like I got a free upgrade for that order to ENIG.

Overall, I'd say the bare boards are very good.

I also had assembly done on several of the board orders. These were turnkey assembly orders in which I supplied BOMs and PCBWay purchased the components on my behalf. These were all prototypes - the number of boards assembled per order was 5 or less. I'm guessing that the process used by PCBWay was to stencil print solder paste, hand place most of the components, bake in a reflow oven, then hand solder any remaining components. The components that went through the reflow oven were generally well soldered - you can see in the pic in the original post that all of the visible components had even amounts of solder filleting, and the components were well centred on their pads. I can see on one of the boards around 2 or 3 components that show signs of hand soldering - there are a blobs of uncleaned flux residue around those components. I'd guess that those components either were missed in the reflow pass, or somehow fell off before reflow. I can also see some uncleaned flux between the pins of some of the fine pitch SMD components. There are three or four areas where the solder paste smeared a little bit. I don't think these affect functionality of my boards, but I haven't thoroughly tested them yet.

Assembly-wise, I'm happy with them, particularly considering the hand placement. The most difficult board had components on both sides, and included 0402 passives.

There was an issue with packaging of some of these assembled boards. All of the assembled boards were placed in separate static shielding bags. However, the larger boards (100mm x 100mm) were wrapped in bubble wrap before placing them in the static bags, which kinda defeats the purpose of the anti static bags.

Time between payment and shipping was 20 calendar days. There was also a period of about a week before that during which the BOM was revised several times due to unavailable components. Shipping was 5 calendar days (DHL express).

Overall, the only major issue for me so far was the wrong component discussed in the original post, which they rectified by refunding the assembly and components cost for the affected board.

Communications was fine. Turnaround was either the same day, or in the morning on the next business day.

There are a couple of things which I need to do for future orders. Firstly, as suggested by others I'll give them a list of approved suppliers (digikey, mouser etc), and only allow components from other suppliers with my explicit approval. Secondly, I'll check board photos to see if the right components have been loaded (although that's not always possible - component markings aren't always legible). And finally, I'll include instructions not to put bubble wrap inside static bags of assembled boards.
 

Pics:

Flux residue would be a terrible name for an alt rock band:

TSSOP8 - 0.65mm pitch - footprint (please excuse my poor macro photography). Alignment of pads to soldermask and drilling can be seen here:

eeprom mounted in that TSSOP 8 footprint. The white specs around the "U6" silkscreen are bits of solder paste residue:

- S

Almost all of the white spots in that photo, including the one you've highlighted, are dust specks, and quite likely caused at my end. There's also a small scratch on the topmost pad of C12.

- S