Testing a Coolaudio V3205

[Ksander]

I was given a couple of 3205 BBDs, which were ordered on AliExpress. To test whether they work:
- I programmed an Attiny85 to generate complementary square waves of 10-100kHz, to use as clock signals, and
- I use a voltage divider (10k potmeter as variable resistor + 100k fixed resistor) to get a 14/15th V_{GG}.
- 5v V_{DD} for both the Attiny and 3205 comes from a 78L05

As audio input, I feed the V3205 a guitar signal, buffered using a tl072. The clock signals look clean on an oscilloscope and the buffer output fed to an amplifier directly also sounds fine. The 3205 however produces no output other than some noise.

Is the test setup flawed, or can I conclude that the 3205 doesn't work?

Given that they are from AliExpress, i don't trust them, but I'd like to make sure.

[ElectricDruid]

You need the correct bias voltage on the BBD input, and you need resistors on the BBD outputs (pull downs/ups, can't remember which).

It definitely won't work without the correct input bias. The output, I'm not sure, but the "minimal circuit" for a BBD and the "datasheet circuit" are basically the same thing - there's *nothing* there you can really leave out.

[Ksander]

You need the correct bias voltage on the BBD input, and you need resistors on the BBD outputs (pull downs/ups, can't remember which).

It definitely won't work without the correct input bias. The output, I'm not sure, but the "minimal circuit" for a BBD and the "datasheet circuit" are basically the same thing - there's *nothing* there you can really leave out.

Do you maybe know what the bias voltage should be? I didn't have a bias network, so added that (1M/1M) between V_{DD} and GND, but that didn't help. I also didn't have a pulldown (100K, what I gather from the few schematics I can find), but that didn't make a difference either.

[ElectricDruid]

Well, the "official" MN3205 datasheet shows a bias divider with 100K at the top and a 100K trimmer at the bottom, with a note to "adjust for minimum distortion". That implies a bias level at least somewhat less than 1/2 Vdd. It also shows 100K pull*ups* to Vdd. You sometimes see schematics with the two output pins simply hardwired together, in which case the pull-up value is usually halved to 51K or 47K (the two 100Ks would be in parallel). To reduce clock noise, I think a 10K trimmer in place of the 5K6 mixer resistors is a good idea, but you're not there yet. Let's get a signal through the thing first, if we can.

[Ksander]

Yes, getting a aignal theough was the first goal.

Anyway, with these changes, there now is a result. I've got two of these ICs. One of them doesn't do anything, but the other produces a 100Hz wave. It is pretty clean, and independent of the clock signal. But that's not what it should do, so... fakes?

[ElectricDruid]

Certainly sounds pretty weird.

If it's a fake, how does the chip *look*? The 8-pins-on-a-14-pin-package is pretty unusual. What else could you find in such a package that you could relabel? Clipping three legs off each side would be pretty obvious, I'd have thought.

[Ksander]

They certainly look the real deal, and I admit that i'm not certain about the surrounding circuitry, but then again, if it would have something to do with that, I wouldn't expect an exactly 100Hz wave output, and I would expect it to vary when changing the clock signal (which works well, verified with an oscilloscope)?

[Ksander]

Scope image

[PRR]

, I wouldn't expect an exactly 100Hz wave output,

Are you in 60Hz land or 50Hz land? Ripple on 50Hz AC-to-DC power lines is often 100Hz. i.e. it is feedin power line nor audio signal. This may be a bad chip or bad wiring.

[Ksander]

Ah

, I wouldn't expect an exactly 100Hz wave output,

Are you in 60Hz land or 50Hz land? Ripple on 50Hz AC-to-DC power lines is often 100Hz. i.e. it is feedin power line nor audio signal. This may be a bad chip or bad wiring.

Ah! I didn't think of that. I am in the Land of 50Hz, so that chip could be bad. I'll put that one aside and experiment further with the other chip.

[amptramp]

That package looks like the kind used for reed relays - DIP packages with pins clustered at each end and other side pins missing.