Parallel effect loop

[Bob-I]

I'm looking for a schematic for a tube buffered parallel effects loop. I've build serial loops and they're ok, but I'd like to compare to a parallel loop and see if there's any improvment.

Does anyone have a schematic that I can try?

Thx.

[Sven]

Greetings,

Although the concept is rather clear to me from studio gear, I would like a bit of clarification, if anyone cares to contribute some time, about that concept in valve guitar amplifiers. Certainly, one could do search of this Web page and perhaps find some material ... but, can we have it clarified here and now?

My understanding is that signal would flow from preamplifier, than 100% into an effect out through FX send (in my case the effect would be analog delay guitar pedal), where on the pedal one would make wet/dry balance, than re-enter the amplifier circuit (FX return) and go into next preamp stage and/or to the PI and the output stage. Is that serial effects loop? If it is, I do not like it.

I would like the singnal from pre-amp to be split, so ¨branch¨ ONE would continue ¨dry¨ into next stages of the amplifier until the final output stage of the amplifier, while the branch TWO would go out (FX send) into an analog delay pedal (which would be set for 100% wet signal) and than the returned signal would be ¨mixed-in¨ into the ¨summing¨ point where signals from FX return and branch ONE would be added. In that case one would have one pot that would determine how much (perhaps in percentage) of Wet signal (from branch TWO, that is, FX return) would be mixed with the original dry signal from branch ONE. For me, that would be ¨parallel effects loop¨. That would be a better loop, definitely.

Am I right, or is this all wrong? If I am wrong, how could I achieve the second, preferred signal flow anyway?!

And, by the way, are Dumble FX loops, as shown on Dumble schmatics with FX loops, actually scenatio one (for me, serial) or two (for me, parallel, as explained above)?

How could we get scenario two introduced into a Dumble schematics?

Thanks.

Sven R.

Thanks.

Sven

[Bob-I]

Certainly, one could do search of this Web page and perhaps find some material

I've searched quite a few sites with no luck.

My understanding is that signal would flow from preamplifier, than 100% into an effect out through FX send (in my case the effect would be analog delay guitar pedal), where on the pedal one would make wet/dry balance, than re-enter the amplifier circuit (FX return) and go into next preamp stage and/or to the PI and the output stage. Is that serial effects loop? If it is, I do not like it.

Yes, and I agree. I built a Dumbleator and I feel like it's interfering with the dry tone.

I would like the singnal from pre-amp to be split, so ¨branch¨ ONE would continue ¨dry¨ into next stages of the amplifier until the final output stage of the amplifier, while the branch TWO would go out (FX send) into an analog delay pedal (which would be set for 100% wet signal) and than the returned signal would be ¨mixed-in¨ into the ¨summing¨ point where signals from FX return and branch ONE would be added. In that case one would have one pot that would determine how much (perhaps in percentage) of Wet signal (from branch TWO, that is, FX return) would be mixed with the original dry signal from branch ONE. For me, that would be ¨parallel effects loop¨. That would be a better loop, definitely.

Yes, that's what I want to do.

[Sven]

Greetings,

OK, fine, now I understand. Well, the parallel effects loop seems to be the way to go if one does not wish to loose some of the qualities of the valve amp sound ... and even more so in designs where there is an attempt to get a special high quality sound, such as in Dumble models.

Having the entirire signal go through just about any pedal, that is, leaving it to the ¨mercy¨ of the pedal electronic network, is a loosing proposition in many cases. For that reason alone, one might as well use the pedal up front, and than run the signal from the pedal into an amplifier input, and forget about the whole FX loop deal, pretty much ....

Particularly with pedals that add delay or reverb, one is always compromising the final sound if one runs delayed and/or reverberated signals through the amp. In live concert and in studio situation, one would be much better off if one would run the dry amp sound and have the mixing engineer add delay/reverb effects. However, that is not always possible nor practical, even on the big stages and sometimes even in studios ...

Some guitar players who really care about tone very much, would run DI-out from one amp into another amp that would handle the delay-reverb effects (for example, that is what Brian May did). Such scenarios call for two amps, of course. So, in a one-amp scenario we are returning back to a parallel effects loop, as described in my first posting.

Can we look around for a solution of implementation of such a parallel FX loop, in a Dumble circuit.

One more question. Since pre-amp gives signal enough boost to drive any effects pedal (if guitar can do it, than pre amp boost can actually be too much), why is it neccessary to use valve amplification in FX loop of present (serial?) scheme, or is it for impendance matching and some other issues beyond the actual strength of a signal?!

But, having a parallel loop with ability to ¨mix-in¨ a bit of wet signal from delay and/or reverb pedal, is not really a cure-it-all solution, since even a 10% of mixed-in signal can spoil the dry signal in some ways due to phasing or whatever problems end up surfacing in the summed, final singal that goes to the output stage. In Dumbles, also, such an FX loop would have to be introduced after OD stage, since it would be somewhat weird to delay a clean signal and have the dry signal be overdriven ...

Once again, it would be nice to make a parallel loop for Dumble and see how it works and how it sounds overall ...

All the best,

Sven

[Bob-I]

..... in a Dumble circuit.

This one won't be in a Dumble ckt. I'm working on an amp that will be used primarally for solid body acoustic type guitars. The goal will be a small light clean tube amp that can be used for these type of instruments.

One more question. Since pre-amp gives signal enough boost to drive any effects pedal (if guitar can do it, than pre amp boost can actually be too much), why is it neccessary to use valve amplification in FX loop of present (serial?) scheme, or is it for impendance matching and some other issues beyond the actual strength of a signal?!

Yes, there will need to be some adjustment for drive level. The tube is for buffering the signal and lowering impedance for minimum signal loss.

......in some ways due to phasing or whatever problems end up surfacing in the summed, final singal that goes to the output stage.

Good point. I'm not sure how to handle the phasing issue but I'd hope that effects units would maintain the phase relationship.

[Sven]

Greetings,

In a parallel loop, as described above, any way one looks at it, there is going to be difference in phase(s) between two branches of signal after one is kept ¨as is¨ and the other branch passed through some other electronic circuit. There is no way to ¨fight it¨. In the realm of time-manipulation effect (delay, reverb, phasing, flanging - analog or digital) that is the very nature of the effect anyway, so one should not worry about it. Actually, that is good for time-manipulation effect that they arrive 2, 3, ... 20, 100 ms after the dry signal has passed, so there is no need to be concerned about it (since one would need to be in a much shorter time delay than several msec. in order to get into the zone of concern).

I would really appreciate a design of a parallel loop (also for Dumble) that would not ¨touch¨ the dry signal, but merely send a ¨branch¨ of it out through FX send and return it back into a summing point, whereas one would have a pot that would be labelled in percentages of wet signal (delay, reverb etc.) that is introduced.

Sven

[Tonegeek]

......in some ways due to phasing or whatever problems end up surfacing in the summed, final singal that goes to the output stage.

The ideal parallel loop would have phase switch. Problem is, you need another half of tube to maintain the correct phase for a parallel loop. Many series/parallel loops flip the phase (I think the Dumbleator does too). Even if the loop doesn't flip it , the FX may flip it. In a seies loop it is no problem, but with a parallel loop your'e screwed. I have a Tone King Comet which maintains the correct phase, but because my GT-3 flips the phase, it is unuseable in the parallel loop. I am very interested myself in getting a working parallel loop going for the reasons others have stated here and elsewhere. It seems to me a good compromise would be to build a small phase inverter using silicon and have it in series with your FX pedal if you have a phase issue. I may try this. BTW, there is a schematic on the Dumble forums (search on "Fatman") that I believe is series and parallel capable. I bet it could be modded to work on the Dumbleator too. Still it flips the phase...

[David Root]

There is another thread further back here on parallel effects loops you might want to check out. It is mainly about the so-called "Best All-Tube Effects Loop" shown and discussed in Kevin O'Connor's TUT 1 book. This is a complete design with the buffering cathode follower in, dry mix/pass thru option. It uses a complete 12AT7, so, yes the recovery triode does invert the outgoing signal.

Are there many effects that invert their output signal? I would have thought this would have been addressed by effects mfrs. by now, but then I know nothing about effects circuits themselves.

[Bob-I]

Actually, that is good for time-manipulation effect that they arrive 2, 3, ... 20, 100 ms after the dry signal has passed, so there is no need to be concerned about it (since one would need to be in a much shorter time delay than several msec. in order to get into the zone of concern).

Good point, however I'd be suprised if SOME dry signal doesn't come through cause phase relationship issues.

I would really appreciate a design of a parallel loop (also for Dumble) that would not ¨touch¨ the dry signal

I'd be interested in how you're going to split a signal without touching it. By the nature of splitting you'll have to touch it.

[Sven]

Greetings,

Split signal? The simplest one: ¨Y-chord¨!

Splitting and summing signals is what mixing boards in studios and live concerts are about. Implementation of the way a sound mixing board works in a valve amplifierv FX loop, that is what this is about (¨AUX¨ send in mixing console applied to valve guitar amplifier).

One branch of the signal continues into the next stage, while the other branch goes out (FX send, or DI out in some amps, for studio recording, for example). The branch that goes out, enters into an effect (stomp box or outboard rack mounted delay, reverb flanger, chorus) and comes back as ONLY a WET signal (100% wet set at the stomp box or rack mounted outboard effects unit), and than enters a pot that would be marked, for example, 0-50%, to mix in some of that signal with the dry signal. In time-domain manipulation effects, I really see no problem with phases, since one expects delays anyway ... some of them up to 200 msec, some of them several msec (flangers and such).

The difference would be that with the parallel loop the mixing of dry and wet happens at the amplifier, while in series loop it happens at the pedal or outboard rack unit. It is much better if the mixing happens at the amplifier.

I think that such parallel loop should be placed before the PI-power amp output stage, that is, after all the preamplification stages. That would form the best dry signal, than have it treated by the effect -- and ¨deliver¨ it to the power amplifier all summed up.

The whole principle would work as the ¨AUX¨ send on a sound mixing console, really. That is the explanation of the entire secnario, indeed.

All the best,

Sven

[Bob-I]

Greetings,

Split signal? The simplest one: ¨Y-chord¨!

Splitting and summing signals is what mixing boards in studios and live concerts are about. Implementation of the way a sound mixing board works in a valve amplifierv FX loop, that is what this is about (¨AUX¨ send in mixing console applied to valve guitar amplifier).

Sorry, I've never done that without some signal loss, not to mention mixing them back together.

One branch of the signal continues into the next stage

In this case the "next stage" is the PI where the signals need to be mixed.

The whole principle would work as the ¨AUX¨ send on a sound mixing console, really.

Except that a sound mixer would have buffers, mixers etc in the loop. My goal is to do just that but with tubes in a guitar amp.

That is the explanation of the entire secnario, indeed.

And I disagree, which of course is my right




And here's what I did as a test. My Dumbleator got modded with this schematic. IMHO it's working well, but isn't much more effective than a series loop. There's one mistake, on the series parallel/switch.

[drz400]

You need negative feedback to create a virtual ground on the return or you will get bleed and phasing issues. The best way IMO is to use a tweed mixer like the input of the Low power 5E8A Fender or come off each of the plates with coupling cap and resistors to mix, that is shown in the RCA tube receiving manual, but that costs a whole tube for the return. The best way is to use a cathode follower tube send or an Opamp send, an OpAMP mixer for the return with NO gain and a tube for the gain, if Opamps give you the heebie jeebies then do this (from Aikens site), it works like a charm and you can adjust the gain by raising Rf (1M still works fine) but raise it too much and you will loose the virtual ground. In the above schematic you are really better off dropping the signal before the cathode follower especially with the limited headroom if that is a 12AX7 and that 10K load resistor, so you can also have a constant output impedance and then you wont have a loss from splitting either, The pot raises the output impedance depending where it is set, if you must use it at least make it 100K. That 10K cathode load resistor on your send is OK if it is an AT7, AU7 etc but if it is an AX7 you should raise that to 100K otherwise depending on your B+ you could be exceding the Max V across the tube as well as limiting your headroom for no reason
[img:495:356]http://www.aikenamps.com/fb_amp8.gif[/img]

[Bob-I]

You need negative feedback to create a virtual ground on the return or you will get bleed and phasing issues.

Ahh.... I see... I think

The best way IMO is to use a tweed mixer like the input of the Low power 5E8A Fender or come off each of the plates with coupling cap and resistors to mix, that is shown in the RCA tube receiving manual, but that costs a whole tube for the return. The best way is to use a cathode follower tube send or an Opamp send, an OpAMP mixer for the return with NO gain and a tube for the gain, if Opamps give you the heebie jeebies then do this (from Aikens site), it works like a charm and you can adjust the gain by raising Rf (1M still works fine) but raise it too much and you will loose the virtual ground.

I'm trying to keep this to one 12AT7 so I'll play with this for awhile. And no op amps for me, they don't smell right.

That Rf seems really SMALL to me. Typically I use 22M to 44M on a LNFB. Of course this is for 0 gain so maybe it's right. I'll try it.

In the above schematic you are really better off dropping the signal before the cathode follower especially with the limited headroom if that is a 12AX7 and that 10K load resistor, so you can also have a constant output impedance and then you wont have a loss from splitting either, The pot raises the output impedance depending where it is set, if you must use it at least make it 100K. That 10K cathode load resistor on your send is OK if it is an AT7, AU7 etc but if it is an AX7 you should raise that to 100K otherwise depending on your B+ you could be exceding the Max V across the tube as well as limiting your headroom for no reason

So split BEFORE the CF? Hmmmm... I agree that 100K would be the right impedance, the 250K was in there and I was in a hurry. I'll change that when I have more time. No it's a 12AT7 so I'll stick with 10k. All measurements are in range.

[img:495:356]http://www.aikenamps.com/fb_amp8.gif[/img]

Thanks, this is very helpful.

[drz400]

Not SPLIT before cathode follower but the Level Pot before the cathode follower. Like a 1Meg then feed the cathode follower, IF you will have a level on the cathode follower output then make it small like 50K or 100K at most. But you shouldnt need that, then with the cathode follwer output will be less than 1K and that is no problem and no loss feeding a total 10K load so you can probably create 4 splits with no problem. Make sure the output cap is mopre like 1uf or higher to prevent bass loss. 10K load puts the -3db roll off at 15hz -9db at 7.5hz etc.

Using a pot at the send will increase the signal loss dramatically.
Using the pot at the input of the cathode follower would be better

Yes this Negative feedback for a plate follower is very different than tone shaping, we are creating a virtual ground like an opamp mix so that the dry doesnt feedback through the effect since the one mix resistor is not buffered that is a return path and will cause phase notch tones.

The point at where the two mix resistors connect becomes a virtual ground so the signal can not loop back to the input, Sort of works like an opamp mixer except the opamp is better. But I can undestand the desire to stay Tube even though your effects are not tube

There is a good article at aikenamps.com about this plate follower, he calls it a "feedback amp"

[Bob-I]

Good Dr.... I'm not sure what's going on here. I tried the LNFB and the loop loses all output. Without the loop and with the proper loading on the CF I'm getting just about unity. As soon as I add the LNFB all output goes down dramatically.

I'll have to experiment more, perhaps different values for the load and LMFB resistors, but IMHO the simple gain stage is about neutral as to coloring and gain, and the effect drives nicely from the CF.