Messages

they're in parallel to the rectifier caps

R78 isn't, it's a pull-up to VCC, while the rectifier cap is going to Gnd.

I have tested the circuit on my breadboard with different chips and I get similar behaviour. I actually noticed very faint signal pass when I moved R78 to go to gnd instead of vcc.

edit: I'll try to get some build and layout pictures uploaded today

What confuses me are R77, R78 and C44. I don't get their purpose when almost every comparable delay schematic (examples) does not have them. Especially R78 as a pull-up to VCC.

I mean the rest of the circuit shouldn't matter as long as the compander works as intended, right? So I could sub any tried and tested compander architecture used in other delays for the one used in this circuit, yeah?

Just to note as well, you wont hear any signal on the expander inputs because it's inverting.

Huh? I'm not sure I understand. When I audioprobe my circuit I clearly hear the input signal (which is the delayed signal from the BBDs) on pins 2 and 3. Why would I not hear what's "upstream" in the circuit?

Why they thought that was a good idea on a standardized part though...

So I just woke up and had a thought, the expander is not working, the compressor is, so I should verify that pins 5 & 6 and 5 & 4 have the correct resistance, as they are flipped in this build. However, both sides are the same within a 0.1k margin. On both the 20k/30k "version", as well as the 15k/25k version.

Why they thought that was a good idea on a standardized part though...

The weird thing is apparently I have two types of yellow thailand SA571, two with the cookbook values and the other three with the lower values, they look exactly the same except for the part marking. Maybe those were faulty batch?

So what would you propose? Play with R79? I think I'll breadboard the circuit and see if it even works on its own the way it's designed.

Edit: alright I think I fried a chip because I forgot the voltage regulator, ugh. At first i got good voltages though so maybe onto something. I'll try again tomorrow.

holy moly, I just checked my other chips, most of them measure at 15k and 25k, and another yellowish font thailand sa571 and the only one NE570N (white font S type) I have (I didn't know I did) measured smack dap at 20k and 30k

Hi scruffie,

voltages are the same, upon just rechecking it, pin1 is at around 0.15v.

I have attached the circuit snippets of the compander sections. I would rather not post the whole thing as I'm currently part of tracing it and it's not verified yet, but will do once it works.

the compander measures 19.4k between pins 11 & 12 and 29.5k between pins 12 & 4 (which I assume you meant). This is in line with Figure 9 of the cookbook (R3 and R4).

I will try to build just the compressor/expander circuit on my breadboard, theoretically the sound should not change going in and out, right?

I am currently building an analog delay (not a project here) and I was pointed to this forum apparently as there are multiple BBD delay experts and y'all seem to be building analog delays everyday. I have read multiple troubleshooting threads concerning the ne570/sa571/v571 companders on here, and the compander cookbook, but it sounds like some of the sa571 behave/misbehave a little differently than expected.

I'm using one of those thailand sa571 (i actually have several of them to compare and they seem to all behave similarly) and it works in my aquapuss clone, but in the delay i'm building I audioprobed the signal through the whole delay line (compresser works, all 8 bbds work, filtering, etc) up to the expander input and NOTHING, not even white noise, on the outputs (in this case pins 6 and 7). Here are the voltages (it's mirrored from the aquaboy, imho they look fine, maybe pin1 is a little low):

compander pins:
1: 0.2v
2: 1.8v
3: 1.8v
4: 0v
5: 1.8v
6: 7.6v
7: 7.6v
8: 1.8v
9: 1.8v
10: 6.7v
11: 1.8v
12: 1.8v
13: 15.2v
14: 1.8v
15: 1.8v
16: 1.2v

So now I'm wondering if pin 5 should be biased differently (there is a 8k2 on there now, i socketed it and tried a 15k but still no sound), but the voltages at the output 6&7 look ok at half supply.

What I also find interesting is that while the dm-2 apparently has pin 12 (output bias) unconnected, madbeans aquaboy dlx has a 20k there.