How Many Bad FV-1's In a Batch?

[vigilante397]

They're not that expensive, but I would love to not have to do that. https://www.newegg.com/p/0X6-04KT-0MVH2?item=9SIAM2VB8J3683

And actually I said I have that crystal working fine in two other designs. I have THIS EXACT DESIGN working, and I literally have a boxed one on my pedalboard right now that is working. I don't think the problem is with my board, unless this particular board has issues, which would be disappointing as they were supposed to have full electrical testing.

[octfrank]

What do X1 and X2 look like when it is not operating properly? Frequency?

[vigilante397]

My oscilloscope is in my office at work which I'm currently locked out of (non-essential worker), but if I find a reason to go in sometime soon I'll probe the board and let you know.

[free electron]

To be 100% sure the clock is not the source of the problem i'd try to:
- remove the xtal + caps
- generate a 32kHz 3.3V square wave externally
- feed it into the pin X1/clock input (10).
There are many ways to create the clock signal: 555 timer (low voltage version), two cmos inverters + crystal (basically replicate a typical circuit which is built in the FV1) or even a 3.3V version of arduino with one pwm channel set to 32kHz 50% duty.

Also, it is recommended to add a small 100n ceramic caps close to the input and output of the voltage regulator to ensure it's stability. Long shot, but maybe the combination of factory tolerances of the chips makes the vreg oscillate and pollute the supply voltage creating the distortion at low levels.

Have you checked the DC bias voltages on audio in and outs? Compared them with working unit?

[vigilante397]

I wanted to update this because it has been semi-resolved. I still don't know what happened with that board as I still have a fully-functioning one boxed up and on my pedal board, but I removed the FV-1 that was on this particular board and put it on another board and it worked. So I was wrong, the issue was not faulty FV-1's, it was something wrong with the board, now I just need to find out if it was in the design or just a manufacturing error on that one board.

[bartimaeus]

as others have said, C17 isn't a necessary part, and i haven't seen it on any designs besides this one. with those clock crystals, you really just need to try a bunch of slightly different values for C15 until you find one that works with the crystal from your  supplier.

10nF for C11 is VERY low. the datasheet value is 1uF and some folks use values up to 22uF. it sets the reference for the converters so it could be why you're getting bad sound quality. apparently the fv1 has some internal filtering there, so that could be why it sometimes works.

you're also missing the 100nF cap on the DVD pin.

i'm curious, why did you stray so far from the datasheet? i'm a little surprised that some of the boards worked to be honest, maybe the chip is more resilient than i thought.

i also notice that you have copper pour for the ground plane in a bunch of places that are both very close to ic pads and also not actually connected to the ground plane. i know that's fine with better manufacturers, but it does put the board at greater risk of shorts during the manufacturing process.

[vigilante397]

as others have said, C17 isn't a necessary part, and i haven't seen it on any designs besides this one. with those clock crystals, you really just need to try a bunch of slightly different values for C15 until you find one that works with the crystal from your  supplier.

10nF for C11 is VERY low. the datasheet value is 1uF and some folks use values up to 22uF. it sets the reference for the converters so it could be why you're getting bad sound quality. apparently the fv1 has some internal filtering there, so that could be why it sometimes works.

I've played with different caps, as mentioned above. I have 10pF on the working boards, but for some reason it didn't want to work on this board. Also 10nF for C11 is a typo, it's actually 10uF. They have the same footprint in my library so sometimes I get careless during schematic captures.

I have 4 different FV-1 based layouts, I have never had any issues except with this one board from this one design.

[bartimaeus]

ah, gotcha! that's pretty weird. interesting that 100nf cap appears to be unnecessary!

[vigilante397]

ah, gotcha! that's pretty weird. interesting that 100nf cap appears to be unnecessary!

It's still not a bad idea, I just tend to be a little lax on my filtering and decoupling caps when using linear regulators. If I were using a switcher or LDO I would absolutely have it in there.

[vigilante397]

Alright, this has been completely resolved now. I had some free time and this board has been silently mocking me for months now. Turns out one of my op-amps was bad. The design was fine, the board was fine, the clock was fine, the FV-1 was fine, a stupid op-amp was bad Swapped the op-amp and it's good to go.