Orange Squeezer switch delay - buffer question

[lars-musik]

I just built a double pedal with a Klon-ish overdrive and an Orange Squeezer. Now the Orange Squeezer has a slight delay when switched on. I found some entries in the forum that this is probably due to C18 in my schematic needing some time to charge and that this may be avoided by using a buffered bypassing scheme here, so that the circuit always sees the signal.

As I had the KLON schematic at hand I tried to merge its buffer with the OS.

Could you kindly have a look if that makes sense and possibly remedy the delay when switched on?

Thanks!

PS: I know I could use the first opamp of the NE5532 to do the buffering but that one is already in use in the other circuit on the same board. So this TL071 is intentional.

[Kipper4]

Hi Lars
Why not use a bjt buffer? less part.



Anyone know a good reason this won't work?

[antonis]

Anyone know a good reason this won't work?

Only in case of large signal and R_E//R_LOAD not too much greater than r_e (Q1's emitter intrinsic resistor)
(in such a case, it should appear a signal distortion in the form of exaggereted positive peaks and flattened negative peaks, due to r_e increasement/decreasement following Emitter current variations - "unstable" (R_E//R_LOAD) / (r_e + R_E//R_LOAD) gain..)

But Jack's suggestions are always verified utilitarian...!!!

[GGBB]

If it is C18 (I would have thought C20 actually), you may want to think about preventing that from discharging when in bypass, rather than altering the bypass system. Just guessing - but connecting the anode of C18 to Vin via a ~4.7-10k resistor might work - assuming you can do that with your switch (4PDT)? C20 could be prevented from discharging by lifting the ground connection on R24 (maybe - but it might not be that simple). Check the voltages at the cap anodes when in bypass mode.

[Mark Hammer]

If one is using half of a dual op-amp for the compressor, why not use the unused half for the input buffer?

[lars-musik]

That OS is part of a larger circuit - where one opamp was spare. So that NE5532 is already a re-purposed one.

PS: I know I could use the first opamp of the NE5532 to do the buffering but that one is already in use in the other circuit on the same board. So this TL071 is intentional.

[lars-musik]

Now that I think of it - and now that some of my favourite competent helpers already chimed in - maybe somebody might give me helping eye with the whole mess I am planning. This should be a double pedal as an all-in-one solution for my next gigs. I don't think I need more pedal than that for our sixti-ish music – if I get it to work properly.

It is Soulsonic's Stage Overdrive housed together with an Orange Squeezer. I already built one prototype but as mentioned the Squeezer is quite noisy and – even more annoying - I realized yesterday that if I de-activate the Overdrive the Squeezer sound drops away for half a second and then recovers.
Here is the schematic of what I already  built:

PDF: https://www.dropbox.com/s/b6qxr0eesgpgz5w/StageSqueezer_alt.pdf





Now I realise that if I already have got a symmetrical PS I could use it for the Orange Squeezer as well. But as my electronic knowledge is still quite limited I am not sure where I should connect which parts to -9V and where to ground. Or if it's a good idea to have an Orange Squeezer on +/- 9V at all.

PDF: https://www.dropbox.com/s/zmqxsxkblgiccfr/StageSqueezer04_Buffered.pdf?dl=0

[induction]

the Squeezer is quite noisy

The OS was designed for battery operation and injects power supply noise directly into the op-amp non-inverting input, where it gets amplified along with the signal. My build was unuseable with an adapter until I implemented low-noise biasing.

[lars-musik]

until I implemented low-noise biasing.

That sounds like the way to go. Would you elaborate on what you've done?

Thanks!

Gesendet von meinem Lenovo K33a48 mit Tapatalk

[induction]

Sure.

The trick is to put a big cap to ground at the intersection of the biasing resistors (R20 and R21 on the first schematic) so that power supply noise is filtered out. The problem is that the input signal passes through that node as it passes from C16 to the non-inverting op-amp input, so that cap would also filter the highs out of your signal. We want the signal to see the same impedance to ground at that node that it currently sees, while the noise has a straight path to ground.

Here's what I did:

1. Keep R21 and R22 connected to V+, ground and each other as they currently are, but break their connection to C16 and the non-inverting input. The latter two should still be connected to each other.

2. Now put a cap to ground (100u should be ok) from the intersection of R21 and R22 to ground.

3. The parallel resistance of R20 and R21 is about 213k. Put a 220k resistor between the non-inverting input and the intersection of R20 and R21.

Now the path to ground at the biasing node is much lower impedance than the op-amp input, so the noise will be greatly reduced. Meanwhile the signal sees a path of 220k to ground at the non-inverting input just like it did before, so the highs will be unaffected by the new filter cap.

Worked like a charm on mine.

[lars-musik]

1. Keep R21 and R22 connected to V+, ground and each other as they currently are, but break their connection to C16 and the non-inverting input. The latter two should still be connected to each other.

Hi induction,

did you mean R21 and R20? Otherwise I am lost.

Is that correct:



Thanks for your help!

Lars

[induction]

Sorry. Yes, I meant R20 and R21.

[GGBB]

Deleted.

[R O Tiree]

If you have the compressor after the OD in the chain, there will almost always be a momentary drop in signal strength when you disengage the OD, leaving the compressor running.  The reason is that the OD is pushing out a comparatively large signal, so the gain of the compressor is very low (it's doing its job).  It takes time for the compressor to sense that the input level has dropped and then ramp the gain up...

A stock Orange Squeezer has what I have, in the past, called a "turnover voltage" of about +/-0.8V when V_In = V_Out.  Below that input voltage, gain is more than 1 (rising to about 16 with very small signals) and above that input voltage, gain is less than 1 (for example, with an input of +/-2V, output is about +/-1.25V).

The key components governing this recovery time are C20 and R24 in your schematic - it takes a finite time for C20 to release its stored charge through that 100k resistor.  Sadly, however, that recovery time was carefully chosen by Dan Armstrong years ago to give a pleasant sound and reducing either or both to speed up the recovery time will lead to less than stellar results - "pumping", in this case, and it just sounds weird.  The recovery has to be slow enough to stop pumping and fast enough to prevent "breathing".

The simple solution is to just swap the order around, so that the compressor is before the OD (which is what most people do, for just the reason described). 

[R O Tiree]

BTW, if you're interested in how the Orange Squeezer actually works, I wrote this answer about 5 years ago.  You'll need to download the Tonepad schematic (shown in the original post on that thread) so that the component identifiers match my answer.

[lars-musik]

Thanks for the explanation, Mike. For your five years old wall of text, too. I hadn't seen that one before!

BTW, if you're interested in how the Orange Squeezer actually works, I wrote this answer about 5 years ago.   

Now that I got you on the line, one (possibly) stupid question regarding

C20 and R24 in your schematic - it takes a finite time for C20 to release its stored charge through that 100k resistor.

Now that I got a buffered bypass for the Orange Squeezer, could I use the spare switch from the 3PDT to just short out the 100K to ground, when the squeeezer is of? Shorten the finite time?

Lars

[GGBB]

I've always found this quite helpful - from 13 years ago: http://www.diystompboxes.com/smfforum/index.php?topic=33510.msg233149#msg233149.

[R O Tiree]

Now that I got a buffered bypass for the Orange Squeezer, could I use the spare switch from the 3PDT to just short out the 100K to ground, when the squeeezer is of? Shorten the finite time?

As I understand it, from your post above, you want to be able to switch the OD off while leaving the comp running, OD before the compressor in your pedal chain?  If you short out the 100k resistor to ground, the circuit snaps instantly into blatty, horrible distortion (+1V to -4.5V) via a huge "bang" through your amp... not good.  Even if you could have that portion of the switch only "momentary", you'd still get the circuit slamming repeatedly to the negative rail for as long as you kept that resistor shorted, and your amp and speaker won't thank you for it. Neither will your audience.

The resistor is there to "gently" relax the circuit into higher gain as the input signal reduces.  A higher value makes that take longer, ie "breathing", and a significantly lower value makes it "pump", which sounds very unpleasant.  Making it making it small or even zero (short switch to GND) makes the circuit misbehave horribly.

It looks as though this is one of those occasions where the conventional wisdom of "Compressor before OD/Distortion" is one you can't get away from.