Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

March 29, 2024, 02:50:10 AM

Login with username, password and session length

Recent Posts

 

Parallel effect loop schematic

Started by Marek, September 30, 2008, 05:07:11 AM

Previous topic - Next topic

Marek

I'm looking for a good SS parallel loop schematic. I want to plug in a DSP effect just before a power amp input and to be able to mix the effect with the original signal. I could only find Marshall schematic but I've seen bad opinions about this loop (I'm not sure whether they are justified). Any ideas?

Mark

teemuk

#1
The thing is, there really isn't a universal FX loop circuit that would work with each and every design. You have to tailor the circuit to suit each amp amp/application individually so that signal levels in the FX loop's "insertion point" are correct and that the loop won't load the circuit.

You also are not telling us if the power amp has an input sensitivity matching the line level of the loop or not, or if the planned effect loop takeoff point can fluently drive an FX loop. You also mention nothing about other features that the loop should have.

But in practice, there really are not that many ways to put together a parallel FX loop. The only differences are pretty much in whether the loop needs an external buffer / amplifier stage to drive it, some additional gain or attenuation, or whether you want to have controls for signal amplitude pre / post the loop – usually either in the form of "volume control" or line level selector to choose between +4dBu or -10dBV. Those are quite simple features to implement if necessary, so if you have seen one parallel loop schematic, then you have pretty much seen all of them.

The catch is that you can't just replicate the circuit portrayed in some random schematic and expect it to work. You have to take your design, determine the signal levels in the loop takeoff and insertion points to find out if they work properly with the effects you plan to put in to the loop and if the stage preceding the takeoff point is already enough to buffer the loop by itself. If not, you need to modify the designs accordingly. Without seeing the actual circuits and knowing the signal levels it's impossible to say anything concrete on this regard.

I don't know if Marshall's parallel loop has any other issues than the universal one for all parallel loops: They are prone to phasing problems (effects that introduce a lot of phase shift can skew the frequency response or at worst case trigger oscillation) and effects that are based on varying frequency response do not work very well unless the "mixer" control is turned to fully "wet".

Marek

Thanks Teemu,

The basic idea is to take existing amp and plug in the effect between preamp and the power amp. So the signal levels should be typical - between 0.7 a 1V. I intend to simulate preamp and the effect loop in LTSpice before I build it (guess who convinced me to use the tool :)). Your statement about effect based on varying frequency response not working well made me wonder whether the idea is correct. Do you suggest that parallel loop is not a good solution for some effect (like e.g. chorus)? And such effects should be connected in series? The DSP module that I'm going to use will operate with line level signals.

Mark

teemuk

#3
The issue with parallel loops is basically this: Let's consider an utopistic effect that does nothing but inverts the signal. What happens when you run this effect in a parallel loop and mix it with the dry signal? That's, right: The two signals cancel each other out. You get similar effects when the phase of the FX loop signal shifts, when the signal from FX has some latency etc. These summed together with the dry signal can act as comb filtering and therefore skew the frequency response really badly.

The second, but quite lesser, issue is with effects that alter frequency response and overall dynamics, e.g. an EQ pedal or compressor. Say, you dial in a nice bass boost and compression from some effects but what happens when you mix it up with the dry signal: That's right, the magnitude of these effects is quite miniscule since you're basically just crossfading between bass boosted, compressed signal and dry signal with wider dynamic range. If you want these kinds of effects to perform properly then a series loop is likely a better solution.

Not to say a parallel loop could liven up a distortion effect by mixing it up with some dry, clean signal, though...

...But the fact is that the architecture of parallel loop is such that it inherently suffers from problems such as those aforementioned. I'm not saying that parallel loops are no good, though. The truth is, it depends, and in some cases the parallel loop can be superior in comparison to series loop. Ambient effects, such as reverbs, may for example benefit more from parallel loop configuration. Think of spring reverb circuit for example; within amps it's always a parallel loop. So are most chorus and flanger effects, although problem with them is that the parallel loop of dry and wet signal is already introduced within them internally and hence using a parallel effects loop therefore forms two parallel paths for the dry signal. The very same thing can happen with many outboard reverb effect units as well – unless they have an option to output 100% wet signal. If one dry signal path in such effects, or some effect in the loop happens to invert the signal you already got yourself some trouble...

So, benefits and drawbacks...

But, you can "convert" a parallel loop to series loop when you dial it 100% "wet" so basically a parallel loop can offer more variety. Not to mention that the consequences of jack intermittancy in a parallel loop are less sever than in a series loop.

Marek

Teemu,

It seems that, as you say, there are more drawbacks than benefits. The main reason for the idea of using a parallel loop was to avoid sending whole signal through the DSP module. But you are right that in case of most of effects (compressor, equalizer, phaser) it makes no sense to use such a loop (although in some cases the effect already uses a parallel loop internally). It makes sense only for effects like delay, echo, which send only the "wet" signal to the output.

Mark

Jack1962

Quote from: Marek on September 30, 2008, 05:07:11 AM
I'm looking for a good SS parallel loop schematic. I want to plug in a DSP effect just before a power amp input and to be able to mix the effect with the original signal. I could only find Marshall schematic but I've seen bad opinions about this loop (I'm not sure whether they are justified). Any ideas?

Mark

What Teemuk says is perfectly correct however let me understand what your after , do you what a wet dry mix between your effects from a effects lloop? If so do you have a dual channel or a single channel amp(see where I am going with this)?

                                              Rock On

Marek

The amp is in the design phase  :) but the plan is to have one channel only - this is for a bass guitar. As I said, I want to avoid sending whole signal through the DSP module. Signal level will be ~0.7V. And yes, I want to do a wet-dry mix. As Teemu wrote - in some cases it does not make sense (eg. compressor). Maybe I should consider a parallel-serial switch.

Mark

Marek

What do you think about this effect loop: http://www.runoffgroove.com/splitter-blend.html ? I think it meets my requrements (as long as effect provides only "wet" signal" - the second signal path may be just shorted).

Mark

Jack1962

looks like a solid circuit for what your wanting to do , as the best thing I see about it is that it looks simple, the more links to your signal chain the more problems you will have( just something to think about).

                                       Rock On

armstrom

#9
I'm in the same boat at Marek. I have a DSP board I need to build a mixer for. I like the blend pot approach because it allows you to get fully dry or fully wet signal. Most reverb mixer circuits assume you always want full dry signal and then just allow you to control the volume of the wet signal. Some of the effects on my DSP board (like rotary speaker cabinet) only make sense without the dry signal.

So... For the runoffgroove design can you get away with just the two "send" buffers to split the signal? My DSP board has unity gain so Vin should = Vout therefore no recovery gain is needed. Also, it's meant to drive a load greater than 2K Ohms so it should be safe for just about any power amp input without a buffer. So I'm thinking of doing the "send" half of the ROG circuit. The "Green Send" would be my dry signal and run directly to lug 1 of the blend pot. "Red Send" will feed the input of my DSP module with the output of the DSP board running to lug 3 of the blend pot. No recovery buffers (and one less IC!) Do you think this would work? Should I buffer the output of my DSP board anyway? I know I don't need any of the phase inversion stuff (JFET, switch, etc..) since this will be an internal effects "loop" so there's no opportunity for anyone to invert the signal.

armstrom

#10
I've attached a rough schematic of my idea. I removed the 1M pull-down resistors after the buffers since there shouldn't be any situation where the outputs are floating (I'm assuming that's why they're there, to eliminate problems when nothing is plugged into the green or red channels). I also added a foot switch to cancel the effect and ensure that the full dry signal is passed to the output.

I'm not sure if this is what the OP is after. A circuit like this is not as fool-proof as the runoff groove design and should probably only be used for internal effects where you know all the variables (Impedances, if the effect inverts the output, etc..) It does however eliminate a lot of components (hopefully without any issue! that's the feedback I'm hoping for ;) ) and should be a pretty easy build.


Marek

Hi Armstrom,

I also plan to build the circuit for use with a DSP module. And of course I need it in cases when only "wet" signal is provided from the module. I plan to build the circuit from ROG as it is (with one small correction on the output). I want to be sure that I can get only "dry" or only "wet" signal on the output. With your version you cannot be sure of that. Also if I use 2 pairs of input-output, the circuit will be more flexible (for 2 DSP modules??). Please note that one output can be shorted to the input in a jack so I even wouldn't need additional jack-jack cable. Do you want to build it as a stomp-box?

Mark

armstrom

Yes and no. I plan to build a stomp-box version with an additional gain stage in front (to avoid low-clipping the DSP's DAC) and maybe an output buffer. However, the primary purpose is to integrate it within an amp.

I guess you could modify my footswitch to also short the output of the DSP to ground when the effect is turned off. That will ensure that there is only dry signal making its way out. Another option if you really must have only dry signal is to use standard 3PDT true bypass wiring (assuming a stomp box build) Then the reverb circuit is taken completely out of the loop. As for being able to get a full wet signal, that shouldn't be an issue with my setup. Sure, some of the dry signal will bleed through, but very little. You can increase the size of the blend pot to reduce it even more.

-Matt

Marek

Good idea with true bypass wiring. However, I think that you cannot increase the blending pot - at certain value it will start having influence on the frequency characteristic of the next stage (or amp). But this can be easily corrected with a few resistors connected to the ground. I will try to find such a schematic - I've see it lately.
At the moment I plan to build the ROG circuit with small modifications and later I will think what the finial wiring will be. E.g. I can omit the phase switch or add true bypass (as you suggested).

Mark

Jack1962

sure it will work in theory , guitar impedance's are high and most opamps are as well , most amps(guitar & bass)look for a 8-18k input impedance.

                                   Rock On