The Unknown Spring Reverb

[razabri]

Another update:
I pulled out the 2k2 input resistor, 4k7 tank one and the input cap - slotted the holes and I've been trying different values. I don't have much to choose from, but higher value on the input does reduce distortion, different cap makes somewhat difference and the tank resistor between few different values makes huge differences. Right now I have listened to it and tweaking it for too long - I don't know if I'm getting anywhere or going away from the right combo... Maybe I should somehow place some trim pots instead of resistors and roll the right values.
Also, I've been testing it with the 9v battery and I don't know if switching to 12v will make audible difference. Haven't tried the 3080 bypass mod yet - I'm also not sure if this will reflect on the tone.

[Rob Strand]

I did some calculations.

So the 4k7 on the tank output  is *just* OK.  I wouldn't worry about it at this point.

The biggest problem is the 2.2k input resistor.  It needs to be increased to about 1Meg.

Yes! it's *way* off.

The 100nF (0.1uF) cap that goes connects to the 2.2k is OK for now.

Try just changing the resistor!

So a couple of things:

-  Have you tried running with the LM3080 bypassed with wires?

- On your "wire" pic, you have show the 68k lifted or cut on one end, that's OK
  However, I don't understand the  wire that goes from that point to pin 3 of the opamp.
  That wire should not go to pin 3, the wire becomes the new output wire.

We can clean-up a few things easily but for now just try the above.

You might need to adjust the reverb level trimpot.  I'm expecting you might need to crank
that up full.

When you bypass the LM3080 there might be a bit of bass cut, because of the small output cap - we'll will deal with that later.

The result you should expect from that is
- a clean reverb, no distortion
- the reverb level is about equal to the original signal

[11-90-an]

@Rob what do you think would be more suitable for this application? passive mix? or active mixer?

[Rob Strand]

@Rob what do you think would be more suitable for this application? passive mix? or active mixer?

Both can be made to work.

I do have a couple of concerns about the passive mixer.   The first option is at the output of IC4 on the TX 30 schematic.   The passive mixer might affect the clipper.  So to fix that we could move the passive mixer to after the clipper.     Also, we would prefer not to have the 10k R72 loading down the passive mixer.   So that means maybe put the passive mixer between R72 and C30.   Now we still have a problem the passive mixer is going to cause a signal loss of 0.5 (-6dB).   The limiter level is tuned to match the power amp.  If we put attenuation after the limiter the signal getting to the power is going to be too low.    To get around that we would have to play with the limiter level.

For the active mixer we would put it before IC4 and simply use IC4 to mix in the reverb.   We would resistively mix the inputs to the reverb from each channel then take the output of the reverb and mixed that with the original signal at IC4.   I think this method is a lot less intrusive.   That was my original plan to graft in the reverb circuit.

[razabri]

I did some calculations.

So the 4k7 on the tank output  is *just* OK.  I wouldn't worry about it at this point.

The biggest problem is the 2.2k input resistor.  It needs to be increased to about 1Meg.

Yes! it's *way* off.

The 100nF (0.1uF) cap that goes connects to the 2.2k is OK for now.

Try just changing the resistor!

So a couple of things:

-  Have you tried running with the LM3080 bypassed with wires?

- On your "wire" pic, you have show the 68k lifted or cut on one end, that's OK
  However, I don't understand the  wire that goes from that point to pin 3 of the opamp.
  That wire should not go to pin 3, the wire becomes the new output wire.

We can clean-up a few things easily but for now just try the above.

You might need to adjust the reverb level trimpot.  I'm expecting you might need to crank
that up full.

When you bypass the LM3080 there might be a bit of bass cut, because of the small output cap - we'll will deal with that later.

The result you should expect from that is
- a clean reverb, no distortion
- the reverb level is about equal to the original signal

I was coming to the same conclusion, as for the values of those components. Found this one on the web:

And I've looked at it's values - the 4k7 is the same as with the ELKA, so I left it at peace and it really just sounds best from the bunch. There's that 1M you mentioned too and the cap of 0.1uF too.
I haven't found any of the 1M, but I did try out 220k and 560k and with value increased it does sound much better, but it's fainting - guess that with 1M it would be hardly audible. And I did crank it up all the way. Maybe we'd need to make a return output higher somehow, as it's not really equal this way to the clean signal.

About the 3080 bypass mod - I made a mistake of connecting that pin, as I thought it was the pin 7, as you suggested before - the IC is upside down.

[11-90-an]

pull out the lm3080 and save it for later... (If I'm not mistaken it's socketed...) It will be very useful later on...

[Rob Strand]

And I've looked at it's values - the 4k7 is the same as with the ELKA, so I left it at peace and it really just sounds best from the bunch. There's that 1M you mentioned too and the cap of 0.1uF too.

It's hard to draw parallel without a schematic.   The reason you need 1M is because the gain of the TDA1904, and the fact the signal before the power amp is relatively strong.

The reverb units which use only use opamps, like the link you posted, tend to be for 600 ohm reverb tanks.   Some use parallel opamps to drive 150ohm tanks.  However, for 8 ohm tanks you need a power amp (TDA1904, LM380, LM386), or, an OPAMP with added transistors.   No many circuits have the "extra" opamp before the power amp.

Do you have a multimeter?   You can check what you tank DC resistance is.  I'm pretty sure it will be an 8 ohm type, they typically have DC resistances on the drive coil of 0.8 ohms.

[razabri]

Still no multimeter here, waiting for the next paycheck... 

You're probably right, as you have been all the way and we'll check that value soon.

I did try to jump that hole from 68k to 7th pin of the 082, but there's no signal at all - have tried it with and without the jumping 10k from 1 to 8 - no luck.

However, once I put it back as it were and with the 560k as input resistor, I tried to connect it with the 12v adapter again - it worked like a charm and I got a much better output, close to the clean signal. I realized that I must have reversed the polarity before and burned that resistor, as the circuit has no polarity protection diodes.

About the mix, how would you feel about this, rather than tapping into the amp's circuitry:



I might be able to build this mixer and add clean to one and reverb to the other channel and adjust the two - I'd probably use trimpots. It does run on 9v, but I hope that 12v would be ok too for it. Than just go into that INSERT jack.

If the solutions you've mentioned are better/simpler, I'd go ahead with those too...

[razabri]

Here's another drawing, flipped correctly so that I don't confuse myself again about the ICs orientation.

[Rob Strand]

I did try to jump that hole from 68k to 7th pin of the 082, but there's no signal at all - have tried it with and without the jumping 10k from 1 to 8 - no luck.

Hmm, something not right.  I has to work  .

The bridge wire you added from pin 7 of the opamp isn't required.
Pin 7 goes to the 0.01uF cap.
You want to take the reverb out from the *other side* of the 0.01uF cap,
basically that's the side of the 0.01uF cap that used to go to the 68k.


The 0.01uF cap is quite a bit too small.  It needs to be increased to something like
47n to 220nF, depending on how you want the low-end of the reverb to sound.

However, once I put it back as it were and with the 560k as input resistor, I tried to connect it with the 12v adapter again - it worked like a charm and I got a much better output, close to the clean signal. I realized that I must have reversed the polarity before and burned that resistor, as the circuit has no polarity protection diodes.

The LM3080 could add some distortion.  That's a good reason to bypass it.

As for 560k, that's good news.  I honestly think you need to got to 1M.   If you find the reverb signal too weak we need to increase the recover amp gain.  The first step would be to adjust the trimpot for maximum output.   If that's not enough we need to increase the gain by changing a resistor.

About the mix, how would you feel about this, rather than tapping into the amp's circuitry:

I'm pretty sure you don't need to add a mixer if you use the external audio out/audio in jack.

The way that jack works is it takes the guitar and mic channels, sends them to the audio out.
*However* while the jack is in the guitar and mic signal still pass through to the internal mixer and power amp.
The audio in is mixed back with those existing signals.

So the way that jack work already doesn't the mixing.  I don't know why the amp is designed like that but it's
exactly what you want for the reverb.   You just use the reverb board as is and splice it between
the audio out and audio in on the external jack.

I might be able to build this mixer and add clean to one and reverb to the other channel and adjust the two - I'd probably use trimpots. It does run on 9v, but I hope that 12v would be ok too for it. Than just go into that INSERT jack.

Since the audio out contains both guitar and mix mixed aleady you actually won't gain anything but adding that extra board.

If the solutions you've mentioned are better/simpler, I'd go ahead with those too...

If you can get the LM3080 bypass working and maybe change that 0.01uF cap you are pretty much there.

[razabri]

Seems I didn't understand the 3080 bypass mod, but I think I got it now. There are two holes that are just like made for bypassing it and didn't have any component there, so I placed jumper over there and disconnected the powering 10k from pins 1 and 8. Got he signal through!



I'm not sure about that output cap, I seem to don't have 47n, nor 220n at the moment, so I placed that 100n (.1uF) and it's ok - increasing should give more or less low-end?

About the mixing - I think I understand what you're saying, but I'm not sure how will I adjust the level of the effect, as the mixing INSERT jack doesn't have it's own control pot or anything - should I add a volume pot at the effect board? S that I can lower it thus letting the clean signal through, as I like it somewhere half/half...

[Rob Strand]

[have any component there, so I placed jumper over there and disconnected the powering 10k from pins 1 and 8. Got he signal through!

That's pretty much it.  You should also remove the 100nF (0.1uF) capacitor to the right of the link.   Otherwise the signal feeds back into the *output* of the LM3080.

Instead of a link a 100ohm to 1k ohm resistor would be better.  This is for technical reasons to avoid oscillations when the opamp drives a cable.

'm not sure about that output cap, I seem to don't have 47n, nor 220n at the moment, so I placed that 100n (.1uF) and it's ok - increasing should give more or less low-end?

Larger is more low end.

You could re-use the 100nF cap removed in the previously.

About the mixing - I think I understand what you're saying, but I'm not sure how will I adjust the level of the effect, as the mixing INSERT jack doesn't have it's own control pot or anything - should I add a volume pot at the effect board? S that I can lower it thus letting the clean signal through, as I like it somewhere half/half...

It's best to just put a pot at the output of the Reverb.   That will let you adjust the reverb level from 0 to full.
You won't be able to get *only* the reverb signal.  It's not much use have only reverb anyway.  The reason is the amp mixes the mic and guitar signal internally.   There's no way to remove those signals via the external connector [I shouldn't say no way as there is an evil way.]

[anotherjim]

If you want a switch to bypass the 3080, have the 100nF output cap always in the feed to the output and with an SPDT switch (changeover) to select the feed into the output cap from either the 3080 output pin6 (as it is) or the TL082 output pin7. So the switch lug 2 to the 100nF lower leg and lug 1 & 3 wired from each of those two outputs.

[razabri]

[have any component there, so I placed jumper over there and disconnected the powering 10k from pins 1 and 8. Got he signal through!

That's pretty much it.  You should also remove the 100nF (0.1uF) capacitor to the right of the link.   Otherwise the signal feeds back into the *output* of the LM3080.

Instead of a link a 100ohm to 1k ohm resistor would be better.  This is for technical reasons to avoid oscillations when the opamp drives a cable.

'm not sure about that output cap, I seem to don't have 47n, nor 220n at the moment, so I placed that 100n (.1uF) and it's ok - increasing should give more or less low-end?

Larger is more low end.

You could re-use the 100nF cap removed in the previously.

About the mixing - I think I understand what you're saying, but I'm not sure how will I adjust the level of the effect, as the mixing INSERT jack doesn't have it's own control pot or anything - should I add a volume pot at the effect board? S that I can lower it thus letting the clean signal through, as I like it somewhere half/half...

It's best to just put a pot at the output of the Reverb.   That will let you adjust the reverb level from 0 to full.
You won't be able to get *only* the reverb signal.  It's not much use have only reverb anyway.  The reason is the amp mixes the mic and guitar signal internally.   There's no way to remove those signals via the external connector [I shouldn't say no way as there is an evil way.]

Evil way! Hahahaaa, that's a good one Rob, pretty devlish. Well, I've tried that connector, as the plan was to connect the thing from the inside - to solder on the tip, sleeve and ring of the jack itself, and it works pretty well (at least now, from the outside of it). The mix is alright by my ear and it's kind of well balanced between clean and reverb, but I should roll the effect back a bit, I think.

I did everything you soggested - the link is now replaced with a 1k resistor, the output cap is now 0.1uF and I removed the one that was connecting the 3080. I'm not sure if I'll keep the value of the output cap like so, as it might need a bit more bass to the reverb, we'll see.

Still haven't found a 1M resistor for the input, but I did found a 1.5M one and tried it like that. The effect is pretty much lowered and the 1M is probably a better match. I did push the driver all the way and it would need more on the return amp to get something usable with the guitar amp - through the PC it does sound well, but into the Taxi it would need to be stronger to equal the mix with the dry signal. I'll probably go and buy some 1Ms', as there's no use of pushing it like this. On the other hand, when the driver is rolled all the way up, the device is more prone to picking up electric noise from the surrounding, but the signal has less chance of being distorted, as with a lower value of the resistor, 560k, or even 220k it's quieter, as I don't need to push the driver, but it's more prone to distort the sound, as you hit the strings harder - guess I'll need to find the sweet spot 

Thanks for following all of it, I couldn't get anywhere without you and the guys - the rest of it, when I set the sound right, is to mount it and I was thinking to power it via the amp's circuit, so the effect and the amp can be started with the same switch.

[razabri]

If you want a switch to bypass the 3080, have the 100nF output cap always in the feed to the output and with an SPDT switch (changeover) to select the feed into the output cap from either the 3080 output pin6 (as it is) or the TL082 output pin7. So the switch lug 2 to the 100nF lower leg and lug 1 & 3 wired from each of those two outputs.

Good suggestion, I'm not sure if I were to have any use of the 3080 part of the circuit and I've bypassed it for now, but it's good to know and perhaps I'll tweak with it too sometime. Thanks!

[razabri]

I've found that maybe I can adjust the effect level to the fixed spot, before mounting it, with the use of it's own trimpot. Perhaps I don't even need a volume pot... But if so, how would I connect it and what value would be best?
Is it alright to use this diagram as a guitar wiring standard?

[11-90-an]

I've found that maybe I can adjust the effect level to the fixed spot, before mounting it, with the use of it's own trimpot. Perhaps I don't even need a volume pot... But if so, how would I connect it and what value would be best?
Is it alright to use this diagram as a guitar wiring standard?

Yes, that can be used
A 100k pot would be ok.

[Rob Strand]

Well, I've tried that connector, as the plan was to connect the thing from the inside - to solder on the tip, sleeve and ring of the jack itself, and it works pretty well (at least now, from the outside of it). The mix is alright by my ear and it's kind of well balanced between clean and reverb, but I should roll the effect back a bit, I think.

Sounds good.   Adding the reverb pot will be the icing on the cake.

I did everything you soggested - the link is now replaced with a 1k resistor, the output cap is now 0.1uF and I removed the one that was connecting the 3080. I'm not sure if I'll keep the value of the output cap like so, as it might need a bit more bass to the reverb, we'll see.

OK cool.   You might get a bit more bass but on the whole there's not a lot of low-end filtering as it is now.
I think anything larger 1uF isn't going to produce any noticeable change.  The 100nF is at the point where is cuts some bass, probably less than a fender amp does.

On the other hand, when the driver is rolled all the way up, the device is more prone to picking up electric noise from the surrounding, but the signal has less chance of being distorted, as with a lower value of the resistor, 560k, or even 220k it's quieter, as I don't need to push the driver, but it's more prone to distort the sound, as you hit the strings harder - guess I'll need to find the sweet spot 

Exactly!   It's a delicate balancing act.   I always try to drive the reverb tanks hard to reduce noise etc.   As it is now, it's probably close to the ball-park.  Tuning of small details can only be done by playing the unit.   I have a feeling 560k is going to overload the reverb drive amp when you crank the amp.

If you need more gain in the recover amp you can increase the 100k in series with the trimpot with say 220k.

A 100k pot would be ok.

I'm in favour of a 10k pot.    The reason is the amp input impedance is about 9.1k so a 100k pot will cause the load to vary on the reverb tank and that's going to cause the LF roll-off to shift depending on the reverb pot setting.   The low pot value reduces the filter shift.

When the reverb pot is dialed down.   The opamp sees about 100k  so the that forms a high pass filter with the 100nF output cap of 1/(2*pi*100k*100n) = 15.9Hz.   When the reverb pot is set to full the opamp sees 100k // 9.1k = 8.3k ohm so the high-pass filter moves up to 1/(2*pi*8.3k*100n) = 192Hz  ; a big change.

With the 10k pot the load varies from 10k to 10k//9.1k = 4.67k  and the cut-off varies from  159Hz to 341Hz  much less change, still not great perhaps.

We simply need to choose the cap to set the frequency.   A 1uF cap will drop the cut-off to 15.9Hz and 34.1Hz.  Which is pretty much doing nothing.   Maybe 220nF to 470nF is more on the money.

[11-90-an]

perhaps that "electric noise from your surroundings" is because the board itself isn't in a shielded enclosure yet...

[razabri]

Well, I've tried that connector, as the plan was to connect the thing from the inside - to solder on the tip, sleeve and ring of the jack itself, and it works pretty well (at least now, from the outside of it). The mix is alright by my ear and it's kind of well balanced between clean and reverb, but I should roll the effect back a bit, I think.

Sounds good.   Adding the reverb pot will be the icing on the cake.

I did everything you soggested - the link is now replaced with a 1k resistor, the output cap is now 0.1uF and I removed the one that was connecting the 3080. I'm not sure if I'll keep the value of the output cap like so, as it might need a bit more bass to the reverb, we'll see.

OK cool.   You might get a bit more bass but on the whole there's not a lot of low-end filtering as it is now.
I think anything larger 1uF isn't going to produce any noticeable change.  The 100nF is at the point where is cuts some bass, probably less than a fender amp does.

On the other hand, when the driver is rolled all the way up, the device is more prone to picking up electric noise from the surrounding, but the signal has less chance of being distorted, as with a lower value of the resistor, 560k, or even 220k it's quieter, as I don't need to push the driver, but it's more prone to distort the sound, as you hit the strings harder - guess I'll need to find the sweet spot 

Exactly!   It's a delicate balancing act.   I always try to drive the reverb tanks hard to reduce noise etc.   As it is now, it's probably close to the ball-park.  Tuning of small details can only be done by playing the unit.   I have a feeling 560k is going to overload the reverb drive amp when you crank the amp.

If you need more gain in the recover amp you can increase the 100k in series with the trimpot with say 220k.

A 100k pot would be ok.

I'm in favour of a 10k pot.    The reason is the amp input impedance is about 9.1k so a 100k pot will cause the load to vary on the reverb tank and that's going to cause the LF roll-off to shift depending on the reverb pot setting.   The low pot value reduces the filter shift.

When the reverb pot is dialed down.   The opamp sees about 100k  so the that forms a high pass filter with the 100nF output cap of 1/(2*pi*100k*100n) = 15.9Hz.   When the reverb pot is set to full the opamp sees 100k // 9.1k = 8.3k ohm so the high-pass filter moves up to 1/(2*pi*8.3k*100n) = 192Hz  ; a big change.

With the 10k pot the load varies from 10k to 10k//9.1k = 4.67k  and the cut-off varies from  159Hz to 341Hz  much less change, still not great perhaps.

We simply need to choose the cap to set the frequency.   A 1uF cap will drop the cut-off to 15.9Hz and 34.1Hz.  Which is pretty much doing nothing.   Maybe 220nF to 470nF is more on the money.

That's great, Rob, that you did the calculations and all - I think I have a 22k pot that's close enough to try out how it works. I'm not sure about the cap you're mentioning - is that the output one? Now it's 100n and I should go from 220n to 470n, as I understand.

Also, I'm not sure about what are you saying about the 100ks' in series and trimpot of 220k 

There's that noise issue too... I don't really know if it's lacking shielding or is it the circuit itself, as when the trimpot is rolled down, there's no noise and no effect, but cranked full up it is a bit noisy. I get that reverb tanks can be a bit noisy, and I'm not sure if the noise here is coming from the fact that it's, as you said, power driver and not a current driver... I lack the knowledge to understand what's the difference and how maybe that 4w 1904 amp is itself a bit noisy, or it's that not-so-good of a choice TL082 as the recovery. Maybe it's because my power adapter is not giving very clean power and it'll pass once I switch to the battery. When I was recording some tests, I could clearly hear that it's picking up noise that's coming from my laptop's hard drive and rolling the trimpot increased this noise. There was some more noise that was from the lack of grounding and I'm not sure but it seems like it happened once we bypassed that 3080, but then I just connected the ground of the circuit to the chassis of the springs and it stopped.

Anyway, I think that it's time to try and mount it, as things can vary on the position and power and I'll try and do so in a way that I can still easily adjust the details.