yet another ABDX with dry signal but no delay, and a whine at pins 3/4 on BBD

[medeiros68]

Howdy folks,
seems like several people had this problem before, but I am running of ideas as to how to further debug my Aquaboy deluxe.

I am powering the ABDX with a LM317 where I could adjust the voltage precisely to what I wanted.

Here are the IC voltages:
IC1 LF353N
Pin 1 -> 6V
Pin 2 -> 6V
Pin 3 -> 5.97V
Pin 4 -> 0V
Pin 5 -> 5.97V
Pin 6 -> 6V
Pin 7 -> 6V
Pin 8 -> 12V

IC2 SA571N
Pin 1 -> 0.9V   Pin 9 -> 1.8V
Pin 2 -> 1.8V   Pin 10 -> 4.8V
Pin 3 -> 1.8V   Pin 11 -> 4.8V
Pin 4 -> 0V      Pin 12 -> 1.8V
Pin 5 -> 1.8V   Pin 13 -> 12.0V
Pin 6 -> 1.8V   Pin 14 -> 1.8V
Pin 7 -> 3.0V   Pin 15 -> 1.8V
Pin 8 -> 1.8V   Pin 16 -> 0.7V

IC3 MN3008
Pin 1 -> 11.8V
Pin 2 -> 5.8V (clock from MN3101)
Pin 3 -> 10.8V
Pin 4 -> 10.8V
Pin 5 -> 0V
Pin 6 -> 5.8V (Clock from MN3101)
Pin 7 -> 3.5V
Pin 8 -> 10.8V

IC5 MN3101
Pin 1 -> 11.8V
Pin 2 -> 5.9V
Pin 3 -> 0V
Pin 4 -> 5.9V
Pin 5 -> 5.2V
Pin 6 -> 6.4V
Pin 7 -> 6V
Pin 8 -> 0.8V

IC6 TL062
1 - square wave
2 - triangle wave
3 - square wave + offset
4 - 0
5,6,7 - triangle wave
8 - 12V

LFO working fine with the Vactrol I built with an LED and LDR.

I audio probed the signal all the way from the input through the 571 (pin 2, a bit distorted there), into pin 7 of the first BBD (MN 3008). And then pins 3/4 of the BBD, no signal but whine. I put an oscilloscope probe there and I got the clock signal, amplitude around a couple of volts.

I tested several BBDs, suspecting they could be the problem, but all 14 I had gave me the same outcome, and although many were from China (Janeh2100 and Funkward eBay stores), a few were from a real store I found (the last 2 NOS they had), and some pulls from a defunct DMM.

I omitted D2 and D3, as suggested.

See the pics below.

Any help would be greatly appreciated!
thanks!
gmr

[Morgan]

Similar to issues some of the other guys had, the bias voltage on your 571 on pin 7 looks too low. It should be around 1/2 the supply voltage (about 6 volts). What do you have in R13 & R14? Looks like jig head put 47k's in there with his build at 15 volts. Maybe socket R13/R14 & experiment until that compander bias voltage gets up towards 6 volts.

[medeiros68]

Great!! thanks so much.
Did work (partly).
I socketed R13/R14, and ended up 47K/33K, and the voltage on pin 7/SA571 now reads roughly 6.57, or close to half the supply voltage of 12V. Now I have a window on the bias pot to get the signal through. I get repeats, and can have them varying as intended.

The only remaining critical issue is that the clock signal goes with the repeats, and I do not know how to get rid of it. It is quite a bit fainter, but I can definitely hear it.

Any tips on how to clean it up?...

I am using MN3008s,

thanks so much once again!
gmr

[Morgan]

You should thank scruffie! I've just been reading his posts on biasing the output of the compander.

For the clock noise - can you dial it out with the clock trimmer? From the instructions:

Set the Delay pot to Max. Now finely adjust the CLOCK trimmer until just before you hear any high pitch whine in the audio path. This is the maximum delay setting. 

Also, for 15 volt operation, did you see this?

When building the 15v/MN3005 version of the ABDX, use 8k2 resistors for R50 and R51. This will help reduce the added brightness of the vactrol's internal LED due to the higher voltage supply

I think the modulation and mess with your clock a little bit of not set properly.

[Scruffie]

Nicely done Morgan

Beginning to confirm my theory the SA brand has different biasing requirements...

The whine may well be from your wiring and also the board not being in an enclosure/shielded.

[medeiros68]

Thanks Morgan,
I checked the issues with the Vactrol. The good thing about having an LED to monitor it is that we can easily check modulation.

Clock noise -  may go away once I box it (I hope so Scruffie! you may very well be right - with my audio probe I could hear the clock without even touching the pins, because of the capacitive coupling that should decrease once I clear up all the wiring). But the main issue  is that I am pushing the delay to a 5-6kHz (long delay times), and that seems to leak into the repeats. Once I increase the frequency (or decrease the delay time), then indeed it goes away. Oh well, I guess I would need to double the number of BBDs ...

thanks a bunch guys! The SA issue is very subtle, glad to hear that it can be solved by nudging the bias on pin 7,
gmr

[thorpy6]

This is great news that another has been solved. I do believe that the issue should be raised in the build docs though.

[Scruffie]

Thanks Morgan,
I checked the issues with the Vactrol. The good thing about having an LED to monitor it is that we can easily check modulation.

Clock noise -  may go away once I box it (I hope so Scruffie! you may very well be right - with my audio probe I could hear the clock without even touching the pins, because of the capacitive coupling that should decrease once I clear up all the wiring). But the main issue  is that I am pushing the delay to a 5-6kHz (long delay times), and that seems to leak into the repeats. Once I increase the frequency (or decrease the delay time), then indeed it goes away. Oh well, I guess I would need to double the number of BBDs ...

thanks a bunch guys! The SA issue is very subtle, glad to hear that it can be solved by nudging the bias on pin 7,
gmr

A nice enclosure acting as ground can take out quite a bit of whine but if you're just pushing the chips past their limits yeah you're going to get whine starting to bleed in.

You could increase the filter cut offs, but you'll get a darker/muddier sound, so yeah if you want more delay you need more chips.

[medeiros68]

Thanks Scruffie.
This is what I was thinking would happen.
I was able to significantly reduce the whine in the repeats by carefully adjusting the cancel pot. But there is still a constant whine, so will see if that goes away once I box it. If not, I will just have to dial up my clock and reduce the delay time, as you pointed out.

Now, there is a bit of distortion in the repeats, but I was not able to get rid of it only by adjusting the bias and the input resistors (R55 - which was supposed to be 1k for a MN3008, but I put 10k). I noticed that the signal into pin 7 of the first BBD was a bit distorted, so should I keep on nudging R13 and R14 until I get a clean signal there? Or is there another way to lower the compander gain? If the latter, how can I do it?

Thanks Scruffie for the help. This is a fun project!
gmr

[Scruffie]

R13 & 14 should be the same value, see if doing so helps.

[GrindCustoms]

I had D3 installed in mine causing a whine, once removed i got it working perfectly.

[medeiros68]

Thanks guys.
I still get a bit of whine, which I think may go away once it is boxed/decrease the max delay time.
Distortion is mostly gone. But there's always a bit, which I guess it is just me being a purist and comparing the sound with a digital delay. Chances are we cannot get rid of distortion completely.

The configuration which I found to work well was:
V=12V
R13/14=22k
R27/55=10k (I guess 1k is equaly as good).

I also found that there seems to be some interaction between all pots; meaning, if I bias BBD2 with too much gain, I get T1 messed up and need to recalibrate it. Cancel has a minor dependence on the BIAS pots.

Next step is to get the modulation working properly. I do not like the way it sounds, and it could be my vactrol only works at 9V.

I ended up testing all my MN3008, and indeed,some Asian eBayers will sell you fakes (or I just could not bias them with the numerous combinations of compander resistors, supply voltage and R13/14 I had tried). In 8 I bought, Ihad 4 working and 4 not working.

I hope these comments will help people in the future! The documentation lacks some info that is perhaps dispersed between multiple threads. Maybe this thread can help a bit in that direction.

Thanks everyone for your help!
gmr

[medeiros68]

Hi guys-
I finished my ABDX, and although I really like the delay, I feel that the sound is a bit dark (not only the repeats, but the direct signal).

I have not modded any of the capacitors in the signal path; anyone has any idea how I can make it a bit brighter?
thanks!
gmr

[Scruffie]

Analog delays just naturally are dark by design, it's part of why they're liked, the DM-2 it's based off is actually one of the brighter ones, if you mess with the filters to make it brighter you risk letting whine and aliasing through.

You can try messing with the pre-emphasis and de-emphasis if you want, C3 & R5, C6 & R9, anything further down the signal path I wouldn't play with unless you know what you're doing.

[medeiros68]

Thanks Scruffie for the fast reply,
but I am afraid I was not referring to the intrinsic darkness of the repeats, but rather the direct signal.
I see  C3 & R5, C6 & R9, their 3dB roll off frequency is about 14kHz; so I think they are not the culprit.
I don't think C5/R10, C4/R7 matter either (>200kHz). Not sure how to improve it.
thanks for the tip!
best
gmr