How do I make my MXR Flanger quieter, and without a loss of signal

[rosssurf]

Are there any mods that can be done to my vintage MXR Flanger that can make it:

1) quieter ( less hiss, both when engaged and not engaged)
2) so there is a little boost when engaged,...or at least not a slight drop in signal that I seem to hear now.

I searched for MXR Flanger mods both here and on the net and did not come up with much.
any help is appreciated.

Thanks in advance
Ross - Baltimore Md

[Processaurus]

Not knowing much about the circuit, but the general approach to getting audio through a noisy process is to feed it as loud of a signal as you can before it clips, then attenuate the output down to normal.  So an opamp boost on the input, and a passive volume control on the output should improve the situation, as long as you have headroom currently.

[cobra94]

Hi Ross,

What you are looking for regarding the volume mod can be found in this thread.  I have not tried this myself yet but the solution seems correct.

http://www.diystompboxes.com/smfforum/index.php?topic=46029.0

As for the hiss, I just use a noise gate at the end of my chain and turn it on when using the "noisy" pedals.

Good luck and let us know how it turns out.

Peter

[Mark Hammer]

Unfortunately, everything that makes the MXR 117 Flanger so edgy is exactly what makes it so noisy.  I looked at a couple of other flangers, and a number of them have more lowpass filtering than the 117 has.  At the same time, some classic designs like the early Electric Mistress flangers are not that significantly different in terms of deliberate noise suppression, in comparison to the 117.

So, is there something you can do?  Well, one strategy is simply to make certain rolloffs a little more effective.  For instance, in the mixer stage just ahead of the BBD, we see C23 in the feedback loop.  In tandem with R16, that provides a 1-pole rolloff around 16khz; great for keeping oscillations out but not especially effective for hiss.  Consider increasing that cap to 150pf or simply tacking on a small 47pf cap to the copper trace side on the pads for C23.  That will drop the rolloff to start around 10khz.  So, still plenty of treble bite but a bit less hiss.

C12 (a .01uf cap to ground) also shunts treble to ground, starting around 16khz again.  Increasing it to .015uf again reduces that to around 10khz or so, although this and the previous cap change will have more of an effect at higher resonance settings, when the recirculated signal includes clock noise.

Finally, putting a 150pf cap in parallel with R39 on the output stage will also address some of the accumulated hiss, again providing a rolloff around 10khz or so.

Now, all of these suggested additions will likely remove some of the edge of the 117's tone, but perhaps not so much that you find it objectionable.  The nice part is that all these changes can be accomplished by simply tacking on a 2nd parallel cap to the pads where the existing caps are soldered.  So there is little risk of damage, and the mods are all easily reversible.

If you try these out, let us know if they are effective.

[head_spaz]

I would approach this differently.
1.) I would begin by exchanging the LM340T voltage regulator with a quiter one. For a drop in replacement use the LM2940T-15 (Harder to find these days... try LM2940CT-15-ND which is a +15volt TO-220 pkg from Digikey... $1.83). National doesn't even list these anymore, but these are very quiet regulators. Stock up before they're all gone.
This is also a low dropout regulator, which will be needed if that's a 12v transformer. The schematic doesn't say.
You could also sub in an LM317, but it's not a drop in replacement as such, it will require additional components to set the output to 15v.

2.) I'd also add small film caps right on the input, and on the output legs of the regulator to ground. (0.1 - 1.0uF) This will improve stability.

3.) Then I would replace C20 with a larger cap, at least 22uF, or perhaps 47 to 220uF, depending on whether or not you have the real estate.
C20 is the only place I would ever use a tantalum cap, and it would be perfect in this location because they offer a lower ESR than even electros. But they don't last forever. An electrolytic would be my second choice in this situation.

4.) And equally important, if not more so, I would install BYPASS caps across ALL of the ICs, using small 0.01 to 0.1 uF ceramic disk types, (or film) from the Vcc pins of the ICs (the positive lead) going to Vbias, and from the Vee (negative lead) to Vbias.
(This modification alone would probably reduce the hiss to an acceptable level, but doing all these mods will reduce the noise even further.)

5.) If you have tantalum caps in the circuit, I would replace them all with good film caps or electros. (with the exception of C20)
Tantalums go dry after a while which makes them leaky... a common source of noise.

6.) You might also look at replacing any high value carbon resistors with 1% metal films, especially those you find in the signal path.

7.) And if you really wanna get fancy...  remove the resistors of the voltage divider following the Vreg (R8, R12) and replace them with texas instruments' virtual ground... part number TLE2425 in a TO92 package. This is one trick pony. It's an ACTIVE power splitter (bootstrap) and provides a very low impedence output compared to the 33K resistors. We really should be using these more often in our builds and upgrades. These are the hot ticket.

Simple stuff really.
All of these suggestions will improve the design, without really modifying the workings of the flanger itself. These improvements should reduce the hiss dramatically... and would probably improve the dynamics a bit. Good stuff Maynard!

Enjoy!

Oh... I forgot... you can replace R34 in the input opamp's feedback loop with a 20K to 50K trimpot so you can adjust the "engaged" level to your liking.
This should increase the presence a bit, while reducing the noise just a tad, but be careful not to overdo it or you'll overdrive and distort the bucket brigade and lose all your headroom. I would try this first... and if it turns out that the diodes in U2's feedback-loop limits your ability to boost, then I would add a 20K trimpot in series with R39 and adjust it from there.

[Rob Strand]

I spent some time removing noise from my VOX unit, which is the pretty much the same circuit.

One thing which doesn't help the noise is the fact the signal is attenuated before the BBD then amplified afterwards.  If you drive the input without attenuation then re-scale for the same overall gain you should get more than 3dB improvement in the signal to noise of the delayed path:

R35 150k -> 100k
R31 120k -> 82k
R41 82k  -> 120k

To increase the overall gain a tad you need to decrease R37 and R41 by the same factor.   Assuming my modded value for R41,

R37 150k -> 120k
R41 120k -> 100k

I replaced the opamps with TL072 but you can get away with just replacing U4, dropped R26 to 1k.
I also redesigned the filters.

[Rob Strand]

I just realized you want to drop the gain.  You want to increase R37 and R41 by one value.  IMHO you might be better off increasing R41 alone to 120k.

[morgan]

I'm in the middle of restoring one of these. I thought I'd add my findings for posterity.

So far Rob's mods did the most to rid the circuit of hiss for me. I'll add that I did not mess with R37 and I found that raising R41 to 120k caused a slight drop in volume when engaged. I settled on 100k and the effect was at unity gain when engaged/bypassed.

I had other issues prior to starting the restore. The unit made an annoying burst of white noise every 180* of the cycle when the Manual and Width controls were set at just about or max position. If you turned up the speed this would sound like a thudding noise through the speakers.

I traced the issue down to just after pins 11 & 12 of the SAD 1024A. I thought the chip might have issues and was a bit upset knowing the prices they fetch these days. I noticed that R49 20k trimpot on the schematic marked "dist" and thought I'd have a go at it figuring it must adjust something in the chip that controls output distortion? And also figuring that after almost 40 years that pot value might have drifted a bit.

After carefully desoldering a protective drop of solder used to prevent it's movement I was able to adjust it, and just a hair over from the original setting the problem went away.

Getting back the the inherent hiss in the circuit I found that after I made the mods mentioned by Rob the Regen would make a very loud shrieking noise at the near max setting CW (due to the increase in gain I suppose?) also some of the white noise issue mentioned earlier came back. I found that if I adjusted the R49 just a hair more that white noise went away and if I adjusted the R48 20k trimpot to the point where the Regen shriek went away but the gain between effect engaged and off were at unity.

I replaced all the tantalum and the electrolytic caps with new. I increased the 10uf C20 to 22uf in the power supply section as suggested by Head Spaz in this thread. That's about as big a cap as you might be able to fit in there fwiw. I have a LM2940CT-15 on order and will report my findings when I drop it in. I suspect it will clear things up a little bit more but overall I'm very happy with Rob's suggested changes with a tweak here and there.

I didn't feel it changed the character of the pedal at all and definitely improved it to make it usable in a live or studio setting.

[morgan]

I swapped out the stock regulator for the LM2940CT-15. I didn't really notice much of a difference. Done over again I would likely only implement Rob's mods with my 100k resistor change up. Hope that helps people in the future that are looking to quiet this beast down.