MN3007 ADA Flanger Clone Questions

[Ed22]

Hello,

i've tried with other caps for C29. I only have values higher than 39 pF. So I've tried 101pF and 150 pF. but it was even worse. I got really high frequencies and again had to turn up T4 trimpot to get rid of them. I don't know if it would be better with less than 39pF. One thing is sure, it sounds the best with 39pF. Trying to change R68 (linked to C29) would have an effect?

But i've noticed something that may be part of the issue : I just can't set the calibration at 69kHz for the minimum frequency. I can only have 50kHz or so. I used a scope for this. I've followed Moosapotamus instructions in his video. But once I set the high frequency at 2,8 MHz, I turn T5 trimpot to raise the lower frequency and I don't have enough room to raise the frequency. When I adjut T5, there is very very little change. Since the sweeping noise occurs mainly at the bottom of the wave of the LFO, I wonder if this is the issue : the lower frequency of the calibration is too low. Maybe it's the reason why I have a delay-like effect when I set manual knob at minimum. How could i fix this? Changing T5 and get one with higher resistance?

thanks,

Ed.

[Rob Strand]

OK as far as the whine is concerned it's probably a simple case of the clock frequency going too low.

I think step one is to fix the 50kHz vs 69kHz problem.   The simple solution would be to change C29 from 39pF to 27pF.
If you don't have 27pF you can use 39pF in series with 100pF as a temporary solution.

As to why you need to change the cap, it could be just tolerances in the CMOS gates.  On the other hand it could
be a wrong resistor value somewhere.   There's no check voltages on the circuit that I can see that confirms other parts of the circuit.

As far as problems go elsewhere in the circuit.  If there is a DC offset error in the Manual circuit or the LFO then that's going to introduce a frequency shift in the VCO.  Also the calibration procedure starts with the Manual circuit so if something is wrong there the whole procedure gets out of whack from the start.

From what I can work out the output of IC3d pin 14 should go from 2.36V to 12.30V when the manual pot is adjusted from one extreme to the other.  It would be worth checking that.    BTW, I worked out those voltages by hand so there's a chance I made a mistake - I *think* it's OK and the range looks believable.

So if the manual control looks more or less OK I'd probably reduce C29 as mentioned above.

[Ed22]

Hello Rob,

thanks for your help.
I'll give it a try and let you know. I'll also check component values in these areas.
I've checked the power supply section and found values closed to what they should be (as Moosapotamus noticed in his video). I have 14,53V at positive rail.

Ed.

[Ed22]

Hello,

OK i've checked ALL components (took me some time... ), caps, resistors, etc. All are correct. I've put out the PCB from the enclosure. I've changed R65 (between T5 clockrange trimpot and range pot) from 150K to 47K and I'm now able to calibrate the circuit in the good frequency range  . BUT, I still have this weird sound...I noticed that i can hear it more when manual and range pot are at low settings. I've tried with different values for C29 capacitor (clock) : 28pF, 33 pF... the best results are with the original value (39pF). Anyway i've spent HOURS with it and still have this sound. Maybe it's the way I wired the circuit, i' don't know. If anybody has an idea...you're welcome. 

[jmasciswannabe]

I've got two ADAs on flintlock boards that I can't get distortion out of the top of the range of the sweep. Like you said, such a shame as everything else sounds amazing.

...anyway, Are you using the voltage doubler? I always had an oscillation noise in the sweep until I did away with the voltage doubler and ran it at 18v. Pop that sucker out and run a couple of 9v batteries in series to the regulator and you can eliminate that out of the equation if you havent already.

Good luck!

[Rob Strand]

Maybe it's the way I wired the circuit, i' don't know. If anybody has an idea...you're welcome.

I think there two issues:
- The ability to set-up the unit to the factory specs.   By that I mean setting it up according to the instructions
   and getting 69kHz min clock frequency.  If that can be done and the unit is set-up like that, do you get the sound?
   and do other people get the sound?
- The second is setting the unit up for the sound *you like*.   Maybe that's not the same as the factory spec.
  Maybe you like the sound of longer delays.   That's where you will get stuck since it's very likely
  those longer delays are the cause of the noise.

It is normal to get more noise with longer delays.  Normally you would need to compromise and drop the filter frequencies.

The doubler jmasciswannabe mentioned is worth looking at.   You can remove it and temporarily use a external DC supply.

Just to be clear, when you don't play anything you might hear a slight noise but when you place the guitar you get a siren like whining noise?

[Ed22]

Hello,

jmasciswannabe : yes, i've installed the voltage doubler. Thanks for the idea, i'll try without it. And yes, this unit sounds fantastic, it's very frustrating. 

Rob : by replacing R65 from 150K to 47K, I'm now able to calibrate the circuit from 69KHz to 2,6MHz. But the whining sound is still here. I've tried with another MN3007 and this sound is still here.
I understand what you say by "setting the unit up for the sound *you like*". Sound really depens on calibration in this circuit. I can get rid of this sound by calibrating the circuit way higher (about 8 MHz i think). I've noticed that : considering a cycle of the LFO, this whinning sound appears twice while the wave does its cycle. It's clearly something that has to do with CD4047. I'll try with another CD4047, who knows... Or maybe it's an interference in this part of the circuit. But I've checked solder joints and haven't seen anything wrong.
When I play rythm guitar it's OK, but if I play a nice arpeggio and let the sound ring...then we can hear that whining sound. In a recording situation, it will be heard. Vintage-style circuits do produce hum, i know. But this whining sound is not classic analog hum.

thanks for your help!

[Ed22]

OK...i'm confused...
I've put the circuit on the floor, on a sheet of paper, on the carpet. Plugged it in the power supply of my pedal board instead of the BOSS AC/DC apdapter i used to power it with. And the whining sound dissapeared! 
I then put the circuit on the table with my pedalboard and the whining sound was there again!  until I touched the circuit and lift it up quite a bit.
It's clearly a ground issue. So jmasciswannabe, you're probably right : it comes from the way it is powered. It will probably work with 2 9V batteries. Before getting rid of the voltage doubler (using 2 9V batteries is so  ...), i'll make some other tests (plug the circuit in another wallmart plug, or test with other AC/DC adapter). It surely comes from the quality of the DC power supply that has to deliver the best filtered voltage possible for this circuit.
I'll tell you if I find the solution.

thanks!

[Ed22]

have tried with different adapters and got the same results.
I'm starting to think of a different way to power it : with a 18V wallmart adapter (instead of 2 9V batteries) and pop the voltage doubler out. I've seen the charge pump of madebeans pedal and doubling voltage is not as easy as it seems : it does have limits.
http://music.codydeschenes.com/wp-content/uploads/2013/02/Charge-Pump-Madbean-RoadRage-TRANSFER.pdf

Any suggestion of what type of adapter (max voltage?)? Will a universal (providing enough intensity) adapter be OK?

[Ed22]

 
finally managed to make this clock sound disappear!

First, i've powered the circuit with 2 9V batteries. I could hear again that clock noise...
The last thing I could do was, still with the 2 batteries, swapping the 39pF cap with 28 pF (39p and 100p in series, as Rob suggested). Then recalibration (i've done this maybe 50 times )Miracle! the clock sound disappeared!

To resume, there were 3 wrong things : the clock capacitor, the resistor before range pot (in order to properly calibrate the circuit) and the power section : i still don't know what is really wrong here: the voltage doubler itself or the quality of the tension delivered to the circuit. this design seems to be really sensitive to this.

This is so cool to be able to listen to this thing with no noise! it's an outstanding flanger.

Ok the drawback is the batteries... I'll try with a 18v adapter (i have to buy one). It would be great if it works.

thanks Rob, jmasciswannabe and Stephen

Ed.

[Rob Strand]

To resume, there were 3 wrong things : the clock capacitor, the resistor before range pot (in order to properly calibrate the circuit) and the power section : i still don't know what is really wrong here: the voltage doubler itself or the quality of the tension delivered to the circuit. this design seems to be really sensitive to this.

This is so cool to be able to listen to this thing with no noise! it's an outstanding flanger.

Excellent detective work.   Thanks for posting the conclusions.  (Sorry I haven't been here for a few days.)

FYI:  Some doubler chips on the web seem to be fakes they are charge pumps with low clock frequencies.
         Recent post confirmed it.
         Maybe if you measured the switching frequency of the charge-pump chip you would know your chips have
         a low switch frequency and *if*  the chip had a higher switch frequency the circuit would work without
        noise as drawn.

[Ed22]

Hello Rob!

i'm happy when I can find the solution of issues in forums  , so if someone has the same kind of issues, he would have some ideas to start troubleshooting with.

about the voltage doubler : i measured frequency at pin 7 of LT1054 : 21,25 kHz. Is that a "low" frequency? I've read 1054 has a nominal frequency of 25 kHz.

I'll make some internet search because I really need to understand how voltage doubler works. I must admit that i don't understand what's going on here. More specifically : how the "switching thing" works?

thanks,

[Rob Strand]

about the voltage doubler : i measured frequency at pin 7 of LT1054 : 21,25 kHz. Is that a "low" frequency? I've read 1054 has a nominal frequency of 25 kHz.

That looks OK to me certainly within the board specs of the datasheet. 
The low frequency units are around 5kHz and 10kHz, which is within the audio band and can be heard as a whine o whistle.
The reason why the LT1054 is better for audio is that  25kHz is above audio.

*However*  with an analog delay the analog delay chips switch the waveform  internally and they rely to some extent on the power supple being clean.  So what I think is happening in your case is that process is demodulating the power supply ripple down into the audio band.   It's basically the same processes used to demodulate AM radio signals into the audio band. 

I'll make some internet search because I really need to understand how voltage doubler works. I must admit that i don't understand what's going on here. More specifically : how the "switching thing" works?

Normally the chips produce a negative voltage but with the added diodes you also get doubling.
From this circuit,
http://moosapotamus.net/files/ADA-MN3007-rev5-Jan2010-Documentation-rev20120225.pdf

Steps:

1) pin 2 is switched to ground, cap C39 gets charged to +9V (less diode and switch drop) through diode D8.

2) pin 2 is switched to +9V, that makes the cap (C39) voltage add to the +9V rail.  The voltage on the positive
    side of  cap C39 is now at 9V + 9V (less drops) which is a bit less than 18V.  At the same time
   D8 becomes reversed bias and not long has an effect, and D11 becomes forward biased and the
   18V (less diode and switch drops) at the positive terminal of C39 gets dumped into C40 via D11.
   So C40 gets charge to just under 18V.  The diode drops and switch drops at each of the steps
   drops the voltage a bit from the full doubled voltage.

While in step 1, the charge in C40 holds the output at 18V.  The load will cause discharge C40 and that causes
ripple on the output voltage.  This is exactly the same way you get ripple on a rectified and filtered DC supply
which uses a transformer + diodes + filters caps.

[Ed22]

Hello Rob,

thanks a lot for all these informations. Really appreciate!
While searching things about charge pumps and LT1054, i came accross this thread :

https://www.diystompboxes.com/smfforum/index.php?topic=101135.40

Charlie (Moosapotamus) said he had issues with the charge pump and had noise... i should have searched first for "LT1054" in this site before "ADA Flanger clone", I would have had the solution of my problem. 
Anyway, trying, making tests, is always better because we have to think by ourselves and I now better understand what I've built  . The issue definitely comes from this charge pump. But i don't forget the clock cap. About the resistance before range pot...as you said, it's a matter of what we like to hear. And while tweaking again this calibration, i said to myself : "i don't think this thing has to go so low in the frequency range". 69kHz is a bit too low. It makes a delay-like effect in the first half of the range pot, too much dramatic when enhance pot full up (even if i didn't set the circuit to self-oscillate, but just right before). So I've raised again the resistor in order to get something close to 95KHz, keeping 2,6 MHz as the maximum. That's much better and musical to my ears.

thanks Rob!

Ed.

[Rob Strand]

While searching things about charge pumps and LT1054, i came accross this thread :

https://www.diystompboxes.com/smfforum/index.php?topic=101135.40

Good find.  Charlie's comments convinced me it's some sort of demodulation.

A simple solution is to add an RC filter to the DC output of the charge pump,
Good starting values would be 2.2 ohm resistor and 100uF cap.
Large caps and larger resistors help filtering but larger resistors will drop the voltage.
The 2.2 ohm is a good value but if you were desperate to squeeze a bit more out you
could raise it to 4R7.

See around figure 9,

https://www.maximintegrated.com/en/design/technical-documents/tutorials/2/2027.html

The existing C40 of 100uF is a good large value to do some tests.  However once you get the filter right
you could play games economizing trying to reduce C40 later on;  too small will make things worse again.

And while tweaking again this calibration, i said to myself : "i don't think this thing has to go so low in the frequency range"

If you don't need the clock to go so low there's definitely benefits edging-up the clock.    I was always playing with that on my first flanger  .

[Ed22]

Hi Rob!

that's a really NICE link! thanks!
Just to be sure : i just have to insert a 2.2 Ohms resistor between C40 and D11(last diode of the charge pump, so output supply)? Because in fig9, they're talking about adding a RC filter to the input supply. So input or output supply?
And why Vout (pin 5) of LT1054 is linked to ground, regarding Charlie's schematic?

Ed

[Rob Strand]

that's a really NICE link! thanks!
Just to be sure : i just have to insert a 2.2 Ohms resistor between C40 and D11(last diode of the charge pump, so output supply)? Because in fig9, they're talking about adding a RC filter to the input supply. So input or output supply?

Ahh.  Something went wrong with my head the other day .  Maybe the coffee had entered my mouth but the caffeine hadn't got to my brain yet.

There's two possible causes for the noise: Input ripple and Output ripple.    The words in my post are about output ripple but the link I posted was for input ripple.  Very confusing.    At first I just wanted a picture.

Anyway, let me get it right this time!

For input ripple:   The Charlie's schematic has a very small input cap C38 (2uF).    Maybe beefing that upto 100uF might help and it's definitely worth trying.   Beyond that you would need to go an input filter, like the previous link I gave.

For output ripple:   So what might be happening with this circuit is the 18V rail is too close to the minimum voltage for IC7 and ripple is getting through to the output.  The addition of D7 doesn't help.   Perhaps *replace* D7 with a resistor say 1 ohm.   The added 1 ohm and the existing large cap C31 should provide filtering of output ripple.  That's what my previous post was about but changing D7 to a resistor is a better solution.   If possible, measure the voltage drop across the added 1 ohm resistor.  You want the voltage drop to be a small voltage, less than 0.3V.

One or both of these might help.   If not we might have to increase C39 with the hope of increasing the output voltage a bit.

And why Vout (pin 5) of LT1054 is linked to ground, regarding Charlie's schematic?

I'm not 100% sure why they do that it looks redundant.  However the connection comes
from the datasheet, figure 20,
https://www.ti.com/lit/ds/symlink/lt1054.pdf

---------------------
Some more points about the regulator IC7.   
When the input of the charge-pump doubler is 9V the output will be less than 18V, say 17V or less.    The regulator IC7 needs about 17V to regulate.   These numbers aren't exact as the exact voltage where IC7 stops regulating depends on the load current and also the tolerances of the IC7.  The main point is the charge-pump output is very marginal in keeping IC regulating the voltage.   If the input power supply was say 10V then that could help the cause.   The point here being that because the regulation is marginal you can get varying results depending on your exact power supply voltage or the state of your battery.

Using larger caps around the charge pump and ditching D7 all help.  You want to keep the voltage drop across the added resistor to a minimum.

[Ed22]

Hi Rob,

OK thanks!
I'll try this since it doesn't involve lots of work. I have some paint job to do in my house (way less interesting than debugging a flanger  ) so i won't be able to do this for some time. For the moment i can use the flanger with 18V (it works like a charm like this ).
I'll let you know when i try this.

again, thanks a lot Rob,

Ed.

[Paul Marossy]

I was looking for some info I got from somewhere about a depth control that you could add to this circuit, but couldn't find it in this loooong thread, so I must've found it somewhere else I guess. I have it scribbled onto a schematic but I am not sure what I have is correct... doesn't seem to make sense.

Anyway, I thought I'd post a link to a video I did a few years about my build where I used wood to make an enclosure similar to the one A/DA used but a little smaller and with a slightly different form factor. I still like using this for the really nice chorus like sounds I get with it.

[PRR]

Happy 13,000!