effects loop idea

[daz]

Hmmmm....thats odd. I assure you it sounds a lot better, pretty much like it did pre SLO loop. The level is hot since like i said it's very loud very quick. But i thought that was the recovery gain not input, because the level meter needed to have the input knob up a bit to read where it should. Before it was all the way down i think. But in any case i'll try whatever you suggest because i certainly know the way i configured it isn't going to be right. I think it's just the location that makes it work now because like i said, before i relocated it the tone stack didn't work the same. Now it does, and i think that has to do with why it didn't sound good.

EDIT: I was mistaken, the input level was all the way down. Could have sworn it had it higher. It was late so....

Anyways, still too early to play it but just as a test i put a 470k from the treble wiper and that sounded better. (***maybe***...just drop the tone knobs to match) It sounded bright w/o it which is another thing i didn't notice last night. But still, the harmonic richness is back and thats the main thing. However you think i should drop the level and impedance, i think that will do the trick and make it perfect. In a bit i'll be able to turn it up a bit,....too low now to really be sure whats what.

EDIT2: i checked AC on the treble wiper and it went to at most 5v strummed hard, averaged around 3. But it was where you said (around 25vac) at one side of the pot. Don't know if thats a proper way to determine it, but if so maybe thats why it works?

[daz]

Wow, this is strange....i put a send and return point from the treble wiper in my drawing but somehow posted a earlier draft and now can't find the correct one ! I probably shouldn't have had those 4 beers last nite. In any case of course the insert point is between the wiper and that pot.

[daz]

More....if i turn the treble up a lot (more than usual) it goes into what sounds like blocking and gets super thin. I didn't look at the DSP in meter but i assume it was causing blocking at the unit. Tried a 100k off the wiper with a 10k load after than as you suggested and that fixed that issue for the most part. Tone really is very good but i can tell that it definatly should be configured correctly so i'm anxious to see what u come u with.

[Chuck H]

A lot of guys do fine padding off the treble wiper. It's not the "proper" way to do it, but it works. You just can't run long cables or it'll make you grumpy

I just spent about an hour drawing up some stuff. Not unlike what you're doing now. Plus another with a proper pad, but let's skip that for now. As long as the 100/10k pad isn't affecting the tone stack negatively. Oh, and don't use a coupling cap to the send. You don't need it and it's actually better without it. So now it's as simple as dialing in the recovered drive level. And that's easy. Here's one drawing I did. It doesn't have the split plate, but that's how I'd go for division if necessary. And I've used a 220k/22k instead of your 100k/10k (yours is half the impedance, so, better if it sounds good). Anyway, you're almost done.

[daz]

I appreciate this Chuck, but i feel bad you're spending all this time. You can just describe it to me and thats good enough. Anyways, i will implement this right now and post back. Oh, and i already canned the cap at the send. Also, no issues with cables because i just use 2 1' cables and the DSP sits on the amp. I never use anything in the loop that need to be on the floor.

That said, i have 2 Q's...1-How can i determine the proper level hitting the master? I can't turn it up all the way here, not even close, and it gets loud vet=ry quick on the dial. I wanna do a split load once it's otherwise all done but i'm not sure how to determine the signal is within a reasonable size. Can i turn the master up say 1/4 and check the ACV strumming a chord and expect an approximate voltage at that point? I don't have a huge load to test it and it i tested it at full volume i;'d be looking for a new place to live. Second, and reason i can;t use a non isolated speaker jack? Mine is but it's a cheap plastic one without the proper lugs and i have some new switchcraft at the ready. The mounting spot iis very close to the central ground point.

[daz]

Ok, the difference is that it's brighter, but i like it the other way in that regard. Otherwise it's pretty similar. But eitherway the treble still seem wonky when you turn it way op. No like before when ther were no resistors between the send and treble wiper. But same thing just less of it, and both your diagram and how i had it before do that. It;s not an issue in that i never turn the treble up all the way or even close, but i think it indicates a issue that should be dealt with. Maybe a bigger series R? Or what about lowering the level into the tone stack? Anyways, can you post that other drawing you mentioned? I'd just like to see what exactly it is. I'm guess you're using both triodes?

EDIT: would sending the treble wiper to a triode isolate the treble pot? Then configure it for very low gain then use the second triode for recovery? Just a thought, tho i suppose if it were feasible you'd have already drawn it. I just know that the treble pot has got to be isolated somehow.

I found that the nastiness when i crank the treble was apparently oscillation. I put a 100pf cap across the PI and the treble knob now works fine.

[Chuck H]

Well... The treble pot wasn't isolated before. It went to a master volume control, which is... Wait for it... A voltage divider- BING BING BING BING BING! So not much different from what you're doing. The tonal change you perceive could be a few of different things or some combination of them including, but not limited to:

the slight difference of the load after the treble pot (this should actually be pretty small)

additional and/or different harmonic distortions introduced by an additional gain stage

the difference that 180* phase difference makes to your ears and guitar in a fixed playing position

the difference that 180* phase shift has on the layout and proximity effects in your amp

Sending the treble pot to a buffer stage MAY help by correcting the phase IF that's part of the cause. It also has the benefit of reducing the impedance of the send circuit. Additional distortions from additional processing may be somewhat fixable by employing local NFB to the recovery stage. This could simultaneously be used for any necessary attenuation in that stage. I think we should try this first. I have a good idea in mind. To try it I need you to do something.

Turn down your master, crank the preamp and put a signal into the amp. A signal generator is best for this but I think you'll be plucking and peeking It'll have to do. Now measure the output at the treble pot wiper. Right at the wiper, not at the new voltage divider load.
Post your findings and I'll draw it up. It should be pretty damn close to unity too.

If the oscillation you encountered didn't exist before the loop addition it is likely a consequence of the additional lead dress/layout OR the aforementioned phase difference. Hopefully when the layout is finalized and we've corrected for phase you can remove the cap and be back to your original design. At the least you will probably be able to reduce it's value to 47pf, making it pretty much inaudible.

Don't sweat the effort on my part. I have a little fun with this stuff.

[daz]

Ok, i'll see what i can figure out as a signal generator. But note that i did this with the guitar before and mentioned in a above post that the treble wiper showed a high of 5VAC and a average of 3VAC when strumming the strings hard with the guitar on 10. I'll try it again now and see if it still hits a 5VAC high. Post back in a minute

Ok, this time it was double. Maybe i had something set different, i dunno. But with the gain on 10 and slamming the strings constantly it hovered around 10 VAC. If i can figure out a signal gen i'll try that, but any thoughts about what kind of signal? I can generate a 1k sine wav in my audio app. or any frequency you think will be good. Looking for a cable now....

Oh, and yes, the oscillation may well be dress because i have wires all over the place just tacked on. It's pathetic, comical. But if not at least now i know how to combat it. The cap DOES dull the top end a hair tho so hopefully i CAN do w/o.

[Chuck H]

It doesn't surprise me you got a different reading. This is very dependent number. You DO need to be accurate with your setting remaining the same for all test conditions and the same level of signal going in. My ESTIMATE is that with a signal generator you'll have between 15Vac and 20Vac there with the preamp going balls out. But I've never tested the signal level there in a Marshall type circuit before.

[daz]

Ok, got a read with a signal i created in an audio app. Made 2 of them, a 1k and a 440hZ. I saved them as mp3 and they were recorded to -3dB, but i have no idea how all this should be done as far as getting max signal etc. Anyways, plugging a mp3 player into the amp the 1k sine wav with the gain on 10, it read 7 VAC. With the 440hZ signal it doubled to 14VAC

[Chuck H]

Well my 20Vac estimate is without an appreciable load. With your master volume a 'halfway' setting has something like a 100k to 250k load depending on the taper PLUS a series resistance for voltage division. I only mention it because it's these conditions that you designed the amp with and has at least SOMETHING to do with how the tone controls "feel". So we want to preserve that. I think That works because a 220k series resistance to a "virtual earth" is part of the design I'm thinking of and will "look" like a load to the tone stack. Perfect.

Ideally you want about 100mVac at the input. But it's unimportant now. Clearly by your test with the different frequencies (smart thing to have observed too) it's clear that there is some loading of the signal. I think I'm pretty safe using a 15Vac marker for design purposes. I'll draw something up and re post.

[Chuck H]

This is a more conscientious project. It uses local negative feedback for attenuation and to achieve low impedance and transparency (low harmonic distortion contribution from this circuit). That also allows us to use an inverting phase for the buffer and return the phase to your stock, pre loop condition. Take your time planning it out because the grid junctions on these circuits can be cumbersome with four connections to the same place. I'm not 100% on this being unity gain. Very easily adjusted though.

[daz]

Yeah, this one will be logistically tough but i'll find a way. I'll have to start this tomorrow but i can't wait to hear it. Just out of curiosity, why is the second feedback resistor so much bigger? Is the smaller one how you achieve low impedance? Thanks Chuck ! (for the gazillionth time !)

[Chuck H]

This is the same method of feedback used for something called a "virtual earth mixer". They call it that because the impedance at the input is so low that it's virtually 0V to the signal and there's almost no interaction between different input sources. This with equal sized feedback and input resistors and the gain is unity. You could join ten channels and they wouldn't affect one another. As a bonus, the circuit also has a low output impedance.

When you stagger the value of the feedback and input resistors you can either reduce impedance and gain or raise impedance and gain. On the send triode I'm going for low impedance, a gain reduction and as much transparency as possible. On the return triode I'm going for some gain and as much transparency as I can by using the feedback resistor for any attenuation.

My hope is that this is the simplest loop possible that will return your amp to it's original phase and be tonally close to invisible.

[daz]

Well, what a fiasco....a contortionist's nightmare at best, and an addition to an already sloppy and cluttered amp that put's it squarely in the running for worlds most jaw dropping slop fest of electronic parts. But so what....Chuck, you magnificent bastard ! If anything it sounds better then it did at any point. Certainly as good as it's best at any given point. It took a couple hours being so cluttered but worth every painstaking minute. The last 2 incarnations took about 20 minutes to wire up and i could have lived with the last one. But this is the RIGHT one. The send level is such that it requires the knob to be almost all the way up on one setting of the +4/-20dB switch, and right about in the middle on the other. So i'd say it's about perfect. The only problem is logistically it was a near impossible task with no room for anything, so I'd fear gigging with it w/o a backup. It's so cluttered i could not have even done it without a 7X eye loupe and jammimg my face into the amp. Luckily i rarely do play out anymore so thats not a serious issue. I suppose i could get some terminals to mount around the socket to make it more reliable but I'm leaving it for now. I will have to play with it a while before i can be 100% sure there are no issues, especially at volume. But I cranked it a few seconds and so far so good, and i can't imagine any problems. Oscillation is gone too by the way, so no need for that PI cap any more. Again, a big thanks to you Chuck. You really outdid yourself on this one and I truly appreciate it....A big THANKS!