Way Huge Aquapuss delay clone issue

[Ed22]

Hi,

I've built the way huge aquapuss clone project (see freestompboxes forum). This is the mini MN3005 version (to fitin 1590B enclosure). It works but has an issue: i play a lespaul and the delay repeats tend to distort, specially with the bridge humbucker. I'ver tried to set the bias pot but i can't manage to reach "the sweet spot". Maybe i could decrease the gain of the delay...but i have no special knowledges in electronics. Maybe there is one or two values to change in the input stage. i don't know if the issue occurs because of the input or because of the BBD stage...
Here is the link : the schematic is available at the end of this page :

http://freestompboxes.org/viewtopic.php?f=7&t=4216&start=140

the link: http://www.digitaldestruct.com/images/AquaPuss/AquaPussSchematic.jpg

I've spent lots of time on this project, i can't give up so close to the end!
Thanks for any kind of help,

cheers,

Ed.

[petemoore]

  I have trouble turning down the volume on my Epi LP, it just doesn't turn down until it's at <2 mud.
  If attenuating the input reduces or eliminates this distortion [cool, there it is]...but then you have a problem, low output, which of course it not too difficult to overcome.
  I imagine those chips may not be happy with voltage higher than what they're working at now, is why the suggestion to reduce the signal voltage from the 'outside' so to feel more headroom as far as they're concerned.
  It's hard to tell what voltage the circuit is supplied.
  Attenuate before the circuit, gain recovery after is the general gist.
  I've had luck with compressors ['n Les Paul humbuckers] keeping signal hot, but without the big peaks so signal would become overcook by the next unit, but the other idea allows you to actually have a wet and hot output.

[Ed22]

Hi,

So you suggest to reduce power supply? i've tried some other voltages: at 7,5V, it does distort too. on the other hand, i haven't tried it with a compressor in front of it. (but i guess this delay shoud work well without a compressor (my ad900 works fine in any setup.) it seems that high frequencies distort (this is why it's more noticeable with the bridge pickup).
i suppose it comes from the duo MN3005/NE570N. sound may be different with another MN3005 for exemple. I understand that if i change the gain of the input, i should increase the output section. But the thing is that i don't know which parts to change to try this...

thanks,

Ed.

[Andre]

What happens if you increase the supply voltage to 12V?
I believe not all MN3005's like 9V.

I know Boss used MN3005's selected for low distortion at 9V in their DM-2 delay

[petemoore]

So you suggest to reduce power supply?
  Nope, reducing supply reduces headroom/increases distortion, if distortion is caused by hot signal/lack of headroom, could be, we can't be sure it is.
  Increasing supply is something to hesitate doing, put it off at least until sure it'll make desirable alteration and not fry the chips-or.
    i've tried some other voltages: at 7,5V, it does distort too. on the other hand, i haven't tried it with a compressor in front of it.
  Dern LP just don't turn down ''right'' here, I should try some different pots, yea, sometimes I have an easier time turning down the boosted guitar output, less mud down at low, compressor can lift-and-limit the signal dynamics, transporting more overall 'juice' without making the 'pipe leak'.
  (but i guess this delay shoud work well without a compressor (my ad900 works fine in any setup.) it seems that high frequencies distort (this is why it's more noticeable with the bridge pickup).
  Compression alters dynamics, sometimes I use it to 'shoe-horn' signal more smoothly to a box's input, something to try if you have it.

[petemoore]

i suppose it comes from the duo MN3005/NE570N. sound may be different with another MN3005 for exemple. I understand that if i change the gain of the input, i should increase the output section.
  9v supply, not sure what's going on with those chips.
  But the thing is that i don't know which parts to change to try this...
  Hey, neither do I...I was just trying to prevent over-voltage or damage to your board.
  I kind of suspected...that's why I was making external, and off-board mod suggestions.
  The volume control to input / gain recovery mod can be tried out using say a booster, just set the guitar volume ''low'' [for my LP, that is the mud around the <1 setting...I need to diddle with that pot's taper]..attenuate the signal using a regular volume control [I like the way a boost or buffers output VC works, better than what my guitar does]...anyway reduce signifigantly enough the signal input volume to the echo, then try that out...if no distortion, you're set. Then stick a booster on it to bring the volume back up at echo output.
  I'd need a booster to unitize bypass/echo volumes, depending on how much attenuation was needed at the input.

[petemoore]

  Now...after all that was said...and after having read some HC reviews and other...
  It could be that the circuit requires debugging.
 

[Ed22]

Hi,

i'm afraid you're right. 12V changed nothing...
circuit debugging... 

thanks,

Ed

[petemoore]

circuit debugging... 
  Seems like it this time, as you may have disco'd, sometimes it's hard to tell.
  Audio probe seems a logical way to find where this disto resides, perhaps the neighborhood of it's introduction is a good place to then use the DMM to verify everything it possibly can.
  The debugging page is a great read and follow for...debugging.
  "Distant point'' measurements can verify more than one thing at a time, find the farthest test point [schematic] to measure a resistor and test continuity of all connections in between...ie...say..from the actual transistor lead, then through as many connects as possible and that resistor.

[sean k]

First thing I'd try is pulling R7 and R10, the 10k resistors, out of the circuit and see what happens. Everything on that circuit is set at unity and the only boost and cut, at the beginning and the end, is going on with those 2 resistors in conjunction with the 6.8n caps in series with them. Lifting even one end of each resitor wouldn't be a chore and you may see a difference. I'd lift R10 first and see what happens as it looks to be creating a high freq boost at the beginning.

Other than that I might be inclined to bypass the compression and expansion for the feedback routing and just have the feedback pot at C27 and make it 20K and through a 20k resistor to the beginning of R18. That would allow you to bypass the comp/expand for each repeat but still have it on the in and out given the first signal has actually gone through the comp stage then you'd be just repeating it through the filters and the 3005 and each signal exiting would be expanded... but only on exiting as it were.

Actually as I look at where the feedback signal goes back into the 3005 it basically doesn't go through compression agin but on exiting it goes to expansion so that may actually be your problem. each repeat is being expanded without being compressed again so the idea would be to either have the feedback signal enter the comp so it both comps and expands or just move the feedback point to  after C27. So lift the connection of VR4 to VR5 and reconnect it to just after C27. So that'd be two simple little adjustsments that may make a world of difference.

[Ed22]

Hi Sean,

thanks for all these technical informations. I'll try them and see what happen.

just one thing i haven't understood:

Other than that I might be inclined to bypass the compression and expansion for the feedback routing and just have the feedback pot at C27 and make it 20K and through a 20k resistor to the beginning of R18.

"hot lug" of the feedback pot on r18 through a 20K? and what about C27, where do i solder it on the pot?
sorry,

cheers,

Ed.

[sean k]

Hey Bro, I was kinda working it out as I went along so all the info is in the last paragraph. Lets say that where VR4 is joined to VR5 is at the top of the pot, it's kinda orientated that way in the schematic anyways, so it's just a matter of lifting the connection at the top of the feedback pot, VR4, and rejoining it to the negative side of C27.

What it means is that each repeat is not expanded in the SA571 before it goes back through the 3005. And don't worry about changing the value of the pot. VR5 stays as it is and the connections of C27 to the SA571 stay the same. You just lift the end of VR4 and piggyback it to the C27 negative where it enters the SA571.

The signal then, at the output, is expanded as it is set at VR5, but the signal is split off for repeats is taken before the expansion and just goes through the tone networks before and after the 3005 and the 3005 itself.

Notice that the feedback signal at R24 goes straight to the output of the compression side of the SA571 at pin 5 so it actually bypasses the comp input at pin 6. But at the output of the 3005 and it's tone shaping networks, C27, the signal actually goes into the expansion section and the repeat pot sets the repeat that goes back again so each time the repeats go round they are expanded which can't be a good thing.

If you do what I suggest and don't get enough signal going around then lift the end of R24 where it goes to pin 5 and move it to pin 7 thereby completely bypassing the SA571.

I'm suggesting these things, not because I'm familiar with the actual working of the circuit, but simply because I see on the schematic what may be causing your problems but I may be wrong.

I thought my own delay pedal, with PT2399's, had problems because it was getting dirty and dirtier on repeats, but when I finally used it with a single coil guitar, I'd been using it with preamped piezos on all sorts of instruments, it tidied up really nicely so it wasn't the FX at all but the simple fact I was always throwing far too much signal at it. That said I've tried to give you the info you need to do some playing about with your circuit. Good Luck Hombre!

[Ed22]

Ok, thanks very much for the comments!

I know delays are sometimes unpredictable. I also know that my built was OK (someone helped me to fix it), but i still have this issue. therefore my LP has no hot pickups. i'm sure it would work fine with single coils. But manufacturers don't build two sorts of delay each time they release one, that's why i don't understand! Anyway, i'll try your ideas and see what happen.

cheers,

Ed.

[rustypinto]

Hi Ed,

To get the results you want, you should understand the devices you are working with more. The compander used in that delay (SA571) and the BBD chip (MN3005) both have some very specific needs at different power supply voltages. Yes more voltage is more headroom, but you have to make sure everything still gets properly biased.

The op-amps in many circuits use a 1/2*(supply voltage) resistor biasing, so you can get away with using different power supplies for different head room in general. But not so with the MN3005 and SA571.

Reading the datasheet/ app note of the compander reveals a lot (BTW: ON is now the owner of SA571):
http://www.onsemi.com/pub_link/Collateral/AND8160-D.PDF
http://www.onsemi.com/pub_link/Collateral/SA571-D.PDF

Firstly, your input signal goes buffer > filter > compressor (1/2 of the SA571). From the Aquapuss schematic, R23 and R25 set the DC output voltage of the compressor to 3.0V if you use 10k/10k. The idea is of course to set this right in the middle of the power supply like the op-amp circuits mentioned above do (e.g. 4.5V if using 9V). Using the equations from the datasheet show that a better value for R23/R25 is 24k/24k at 9V. Both the MN3005 and SA571 can handle 12V, so i recommend using that (what would R23 and R25 be in that case?).

Now for the MN3005. Its mandatory that you calibrate this device using an oscilloscope and test signal. If you don't, you could only be passing half of your guitar signal through (hence the distortion). VR1 from the Aquapuss schematic sets the bias of IC3. Again, the idea is to set the bias so that your signal doesn't clip. With BBD's, this point is often not at the middle of the power supply. The best way to find this point is to input a test signal (like a triangle or sine wave at -20dB/75mV) while looking at the output with an oscilloscope. You adjust VR1 to find the point where the top or bottom of the test signal stops clipping.

Finally, the output of IC3 goes into another filter, and into the expander (the other 1/2 of SA571). The output should be set between the power supply again, and using the stock Aquapuss schematic shows that you're getting 3.0V again. If you want to get 4.5V in the case of a 9V supply, you have to add a resistor from pin 12 to ground. Adding this resistor decreases an internal resistor to the SA571 which sets the output bias. Doing another calculation using the equation in the datasheet yields 24k is a good value to get 4.5V.

Delays don't have to be unpredictable. I think if you apply some of these methods you can have an awesome sounding unit!

Let me know if you have any questions. Best of luck!

[Mark Hammer]

Marcus has some good tips.  My experience is that a scope is not absolutely required for setting the bias, however a well-behaved trimpot IS.

To my mind, since there is a compander in the circuit, any audible distortion is likely from the bias-setting rather than from an input overload.  Do make sure that your input signal is as clean as possible when setting the bias, and if it is possible to listen to the output through headphone to do your calibration, try that.

[Ed22]

Hi Marcus (and Mark),

thanks for your tips. I've had to leave this project for a moment. So i haven't done anything on the unit since that time.
i'll check it and follow your advice.

thanks a lot,

cheers,

Ed.

[finalrest]

HI,ED,I THINK YOUR SCHEMATIC IS WRON,CHECK PIN 3 OF MN3005.IT should be connected with VR3.