AC15 w/ reverb build

[Jalmeida]

Hey folks,

Just wanted to share some of my upcoming project. I scored a really cool repro JMI AC15 chassis off ebay and thought it would be cool to make an AC15 2x12 combo with a 12DW7 single tube reverb in the quirky styling of the Vox AC30 Super Twin Reverbs.

I am in a holding pattern waiting for Princeton reverb Transformers for my current build. So perfect window to do some crafting. With minimal mods to the chassis I made this little Top Boost style sub chassis that will be hidden in the back same way the add-on top boosts and reverbs were done. Seeing I won’t be building a reverb tank like the originals I will stick with the 12DW7 style circuit driving a 4FB3A1B tank like my AC10. I am still in planning stages for what AC15 circuit integration will look like. Lots of lessons learned with the AC10 reverb build. So this one will be a better execution I hope. I would love to integrate on the EF86 channel, but am admittedly apprehensive given the fickle nature of it. And ideally if I integrate into the other channel I could still like both channels and get EF86 with reverb.

[tubeswell]

That's a pretty looking chassis. Great way to start


Just be aware of inadvertently creating ground hum loops if you start adding extra chassis bits... (not saying this will definitely happen with your build, but it can happen with seperate bits of metal in the ground return plane like that)

[Jalmeida]

That's a pretty looking chassis. Great way to start


Just be aware of inadvertently creating ground hum loops if you start adding extra chassis bits... (not saying this will definitely happen with your build, but it can happen with seperate bits of metal in the ground return plane like that)

Agree 100% with this concern. I plan to utilize common main chassis Ground points. That doesn’t guarantee no ground loops.

Ultimately 99% of the hum issues I had on my recent AC10 reverb were due to the ground scheme of the add-on reverb circuit that utilized Keystone Tag strips that had the mounting points as grounds. Ultimately I made a new board that had a common ground going to the preamp grounds and changed my reverb tank grounding. I also shortened and relocated mains ground points to a common location.

[Voxbox]

Yep certainly a good-looking chassis.


One thing I wondered about - how about placing the top of the add-on panel on the inside of the top lip of the main chassis and use countersunk screws. That way it wont push out the back panel of the cab. Would easier now rather than later!


On the AC10 - I had the pleasure of working on an original 1959 model. Lovely overdrive to my ears, and its now on my to-build wishlist. Did you manage to get a chassis like the AC15 that looks original?


Cheers, VB

[sluckey]

One thing I wondered about - how about placing the top of the add-on panel on the inside of the top lip of the main chassis and use countersunk screws. That way it wont push out the back panel of the cab. Would easier now rather than later!

Good idea.

[Jalmeida]

Yep certainly a good-looking chassis.


One thing I wondered about - how about placing the top of the add-on panel on the inside of the top lip of the main chassis and use countersunk screws. That way it wont push out the back panel of the cab. Would easier now rather than later!


On the AC10 - I had the pleasure of working on an original 1959 model. Lovely overdrive to my ears, and its now on my to-build wishlist. Did you manage to get a chassis like the AC15 that looks original?


Cheers, VB

Yep certainly a good-looking chassis.


One thing I wondered about - how about placing the top of the add-on panel on the inside of the top lip of the main chassis and use countersunk screws. That way it wont push out the back panel of the cab. Would easier now rather than later!


On the AC10 - I had the pleasure of working on an original 1959 model. Lovely overdrive to my ears, and its now on my to-build wishlist. Did you manage to get a chassis like the AC15 that looks original?


Cheers, VB

I thought about that too. But many of the Top Boost and Reverb panels added back in the 60s were bolted to the outside lip. And these usually came with a felt strip on the lip. i believe that felt would take up at least that much space making the panel thickness a non-issue. That being said I could swap it around with minimal drama.

For my AC10 build I couldn’t find a chassis like the original so I built one in similar fashion. Because I put mine in a repurposed Cambridge 30 cabinet, I made mine as close as possible to the JMI design, but added width for the extra Reverb knob and to fit the Cambridge Reverb 30(early 2000s) panel cutout. I had to add the wood slider rails on the inside of the cabinet. I also did a few other mods(baltic birch speaker baffle that lowers speakers, vents and more rounded corners) and recovered the cabinet to my liking. The chassis panel is anodized aluminum plaque material that me and a friend screen printed the legend on. The legend isn’t too fancy, but works. And I like the red anodized aluminum appearance. In hindsight I would spray a matte clear over it for protection and to subdue it a little.

[Jalmeida]

Here is a Super Reverb Chassis with add-on panel for reference.

[wsscott]

Jalmeida- Are you adding a new Reverb circuit on a separate board, and then mounting that to the existing chassis?  I ask because I built a Mag Twilighter with the help of Sluckey and others, included the Tremolo part of its circuit, but leaving out the Vibrato and Reverb circuits, but left space on the board to add reverb in the future.  I've recently been thinking of adding that circuit now, but the space is pretty tight and I've been trying to figure out how to add the circuit to the amp if I'm short on real estate.

[Jalmeida]

Jalmeida- Are you adding a new Reverb circuit on a separate board, and then mounting that to the existing chassis?  I ask because I built a Mag Twilighter with the help of Sluckey and others, included the Tremolo part of its circuit, but leaving out the Vibrato and Reverb circuits, but left space on the board to add reverb in the future.  I've recently been thinking of adding that circuit now, but the space is pretty tight and I've been trying to figure out how to add the circuit to the amp if I'm short on real estate.

The main chassis and circuit will be bassed 100% on the JMI amp. The sub-chassis will have a tagboard with a single 12DW7 reverb circuit. The circuit integration with the original AC15 circuit i am still working out.

[Jalmeida]

Jalmeida- Are you adding a new Reverb circuit on a separate board, and then mounting that to the existing chassis?  I ask because I built a Mag Twilighter with the help of Sluckey and others, included the Tremolo part of its circuit, but leaving out the Vibrato and Reverb circuits, but left space on the board to add reverb in the future.  I've recently been thinking of adding that circuit now, but the space is pretty tight and I've been trying to figure out how to add the circuit to the amp if I'm short on real estate.

This is the add-on chassis and tagboard that will have the circuit.

[Jalmeida]

Jalmeida- Are you adding a new Reverb circuit on a separate board, and then mounting that to the existing chassis?  I ask because I built a Mag Twilighter with the help of Sluckey and others, included the Tremolo part of its circuit, but leaving out the Vibrato and Reverb circuits, but left space on the board to add reverb in the future.  I've recently been thinking of adding that circuit now, but the space is pretty tight and I've been trying to figure out how to add the circuit to the amp if I'm short on real estate.

As far as space goes, it’s possible to fit a single direct coupled single 12DW7 reverb circuit on a pretty small terminal board.

[wsscott]

Could you post the schematic and layout?

[Jalmeida]

Could you post the schematic and layout?

For the baseline AC10 layout, I used the layout in the Stephen Grosvenor AC30 Service Engineer’s book. Attached is the revised schematic with reverb integration and reverb board layout. I added some additional RC filters show here. Integrating them with the factory AC10 layout is pretty easy. This drawing is how it is currently wired. I would highly suggest you replace the 1M R37 resistor with a pot to adjust the dwell. Or omit the cathode bypass capacitor to lower to reverb tank drive gain. This circuit has a ton of dwell and I find turning the reverb level past 1/2 at loud level can eventually develop a reverb howl. Adjusting the recovery doesn’t resolve this and I believe it is from the drive end. As it stands now i am playing it a bunch and will eventually replace the 1M with a trimmer to tweak it. But I have been pretty happy running reverb at 1/2 which about where I normally keep reverb.

Use in good health.

[wsscott]

Thanks.  Neat ideas.  I wanted to see what your layout looked like, just to get a sense of the size.  The photo does a pretty good job of showing it.  The amp right now with just the tremolo in it is dead quiet, and I'm reluctant to start tinkering with the circuit to add the reverb, and then potentially get hum when the reverb is active.  The effects unit I built, sort of like a Revibe, is based on the Mag 10A of Sluckey  and has great pitch shifting varistor Vibrato, and reverb.  Its just that as I increase the intensity of the reverb pot or the output pot, the hum gets louder.  Its probably normal for a reverb circuit, but I haven't been able to get the reverb to be dead quiet.  I'm using the same tank model that's in your circuit.  I don't know if it has anything to do with it being a tube reverb vs. a driver based on an output transformer.  This model tank has really high impedance, so maybe that's something that could be a factor.  The tanks with the transformers have 8 ohm input impedance and this has around 1900 ohms.  So if I decide to add a reverb circuit, want to be able to remove it easily if I don't like it.

[ac427v]

I wonder if the problem of too much dwell is tied to the point in the circuit where you connected the reverb input. Having it come before the volume control means that you have the entire signal minus the lows being amplified by the reverb driver. That signal is huge when it hits the tank and the recovery triode. Another downside of your tap point is that the wet volume is constant rather than increasing or decreasing with the dry volume.
You could try using a dual 1 meg volume/dwell control.

The more traditional tap point between the volume control and R45 might better correct those issues.

[Jalmeida]

I wonder if the problem of too much dwell is tied to the point in the circuit where you connected the reverb input. Having it come before the volume control means that you have the entire signal minus the lows being amplified by the reverb driver. That signal is huge when it hits the tank and the recovery triode. Another downside of your tap point is that the wet volume is constant rather than increasing or decreasing with the dry volume.
You could try using a dual 1 meg volume/dwell control.

The more traditional tap point between the volume control and R45 might better correct those issues.

Yeah, there was a bit of experimentation with the integration and a few other tweaks i’d like to get the amp 100% where I want it. I am pretty new to this tube amp stuff, so alot of learning curve. My electronics background is aviation. Early on the reverb was motorboating pretty bad when the reverb level was turned up a little. Then feeding back. I lowered the gain on the reverb recovery by suggestion here. That helped the motorboating.There were some other design issues that remedied the feeding boat mostly.  The feedback before was an immediate feedback. What I have now is a vibration induced howl that is induced with tank vibration only if the reverb level is above 1/2 and I play loud. Hard floors make it worse. It is a vibration howl that I believe is caused or exacerbated by the excessive dwell. 

To wsscott’s point, I chased alot of hum in this amp. Then I had a whole bunch of various hums and hisses and crackles to chase down. The majority were 60hz. I had to add additional power supply nodes to the power supply. The original AC10 runs the PI 12AXY and EL84s off one node and the EF86 and the ECF82 off the other. I added another node dedicated to the 12DW7. This reduced but did not eliminate the hum. I added another so that one node feeds the reverb drive and the other the recovery. This helped again, but had addition hum. Ultimately I utilized a layout of the original amp and where the picked the preamp ground from is a mystery. None of the pictures or diagrams show the preamp ground. The AC15 uses a odd ground location at a tube socket screw and the boards all daisy chain. The AC10 daisy chains the same way. But not knowing the ground source 8 picked the bolt that connects the aluminum vertical chassis with the lower one. I had to re-arrange grounds on my reverb board because the original circuit used tag strips that use mounting screw grounds. This alone induced alot of ground loops. And with all the changes to the reverb circuit and the bad location the board was pretty mangled and with lots of solder needing to be removed and cleaned. So I changed it out with a tagboard that picked off and uses a star scheme at the preamp ground. I also spent alot of time with alligator leads trying different ground points to fix hum. Other things I did were:

1. 100R hum pot for heaters
2. Moved Power Transformer Center tap ground to mounting bolt along with rest of mains grounds.
3. Changed Cap can to 32uf+32uf
4. Changed other power node filter after cap can to 16uf
5. Added other RC filters at EF86 before R5 similar to AC15 and AC10 SRT
6. Isolated reverb tank jacks
7. Grounded return shield from reverb tank


I still have a very small hum now. But no hum in the reverb circuit. And the hum is manageable and about the same level as my Vox AC30CC2. So I am just playing it and deliberating other design tweaks to optimize the reverb. 

Sorry that’s alot to read. And I did cover alot on the original AC10 build thread.

https://el34world.com/Forum/index.php?topic=29424.0

[ac427v]

Nope, you missed my point. Your reverb design is wrong. You are putting bandaids on the recovery circuit to try to correct a bad design. Fix it right.

Spend 5 minutes to unsolder the 500pf reverb input cap from the junction of the tube plate and the coupling cap. Resolder it to the junction of the volume pot wiper and the mixing resistor. Play music at all different volume and reverb settings before tweaking the mixing resistor value to get the right proportion of wet/dry signal. Then troubleshoot any remaining issues.
Then go back and fix your AC10 reverb which has the same design flaw.

[Jalmeida]

Nope, you missed my point. Your reverb design is wrong. You are putting bandaids on the recovery circuit to try to correct a bad design. Fix it right.

Spend 5 minutes to unsolder the 500pf reverb input cap from the junction of the tube plate and the coupling cap. Resolder it to the junction of the volume pot wiper and the mixing resistor. Play music at all different volume and reverb settings before tweaking the mixing resistor value to get the right proportion of wet/dry signal. Then troubleshoot any remaining issues.
Then go back and fix your AC10 reverb which has the same design flaw.

I did not miss the point at all. And I will do exactly what you said. The question I have is the 500pf cap value picking off after the .022uf coupling cap. Will there be enough lows and mids to vibrate the springs?

[Jalmeida]

Doh, nevermind.

[Jalmeida]

Nope, you missed my point. Your reverb design is wrong. You are putting bandaids on the recovery circuit to try to correct a bad design. Fix it right.

Spend 5 minutes to unsolder the 500pf reverb input cap from the junction of the tube plate and the coupling cap. Resolder it to the junction of the volume pot wiper and the mixing resistor. Play music at all different volume and reverb settings before tweaking the mixing resistor value to get the right proportion of wet/dry signal. Then troubleshoot any remaining issues.
Then go back and fix your AC10 reverb which has the same design flaw.

So I desoldered the cap that is at the plate and integrated it at the Normal Volume pot wiper before the 220k mixing resistor and I have almost no dwell(regardless of normal channel volume). I have not done a transformerless reverb drive before, but not having a transformer to create the current to vibrate the springs the tank needs all the current the driver triode can deliver. And the control grid signal being after pot does mean I would need more current from the 12DW7 to drive the tank. And other transformerless designs(ampeg)I have seen pull directly off the plate. I have not looked at the Magnatone and other similar. I will come up with some other ideas to tweak it and see what I get to drive the tank. Maybe some other transformerless reverb design research.

[sluckey]

And other transformerless designs(ampeg)I have seen pull directly off the plate.

That's not the point. The point is to take the reverb ***AFTER*** the volume control. And that's exactly what the Ampeg circuits do. That way, whenever you turn the volume up/down the reverb input is also turned up/down, thus keeping the wet to dry ratio somewhat equal. Otherwise, everytime you turn the volume you would also have to turn the reverb to maintain the same wet/dry ratio.

[Jalmeida]

And other transformerless designs(ampeg)I have seen pull directly off the plate.

That's not the point. The point is to take the reverb ***AFTER*** the volume control. And that's exactly what the Ampeg circuits do. That way, whenever you turn the volume up/down the reverb input is also turned up/down, thus keeping the wet to dry ratio somewhat equal. Otherwise, everytime you turn the volume you would also have to turn the reverb to maintain the same wet/dry ratio.

True, but the Ampeg designs I have seen have an additional triode stage between the volume and the reverb driver stage.

I am missing 100% of the point here obviously. I am not worthy. 

Carry on with your day.

[sluckey]

You're not missing the point. And I understand your point too. I agree that adding a dwell pot is needed to avoid the reverb howl and allow you to set the reverb pot to a higher number.

[ac427v]

Please try the second part of my suggestion before quitting.

Tweak the mixing resistor (currently R45 220k) to a much larger value a la Fender. Try 3 meg ohms. That will direct more of the signal to the reverb driver and allow less signal to the dry circuit. The result may be usable reverb and a very clean-toned channel. I think any better outcome is impossible without more signal amplification than is possible with a 1 triode preamplifier design. If you have seen or heard one that does work well (Swart???) please share it!

[Jalmeida]

Please try the second part of my suggestion before quitting.

Tweak the mixing resistor (currently R45 220k) to a much larger value a la Fender. Try 3 meg ohms. That will direct more of the signal to the reverb driver and allow less signal to the dry circuit. The result may be usable reverb and a very clean-toned channel. I think any better outcome is impossible without more signal amplification than is possible with a 1 triode preamplifier design. If you have seen or heard one that does work well (Swart???) please share it!

Definitely not giving up. Was just going to mull over and have a well formulated plan to tweak.

I will give this a try. The normal channel is already alot quieter than the EF86. I am admittedly concerned with how much more I am going to loose changing that resistor to a 3Meg without another gain stage.

And while having the reverb level stay high and not changing with volume wasn’t a crippling factor honestly. It just makes work more like a studio type reverb. It is quirky but manageable.

If I had my druthers i’d want it to work with integrated before the volume knob though like you suggest. But the single tube reverb like this is a compromise setup at best I fear. A complete amp redesign integrating the reverb would get a better working reverb at the cost of the characteristics of what makes an AC10 or 15 a really cool amp. And even Vox couldn’t make a reverb AC10 or 30 that wasn’t a completely different amp hence their hodge-podge scabbed on solution with their own reverb spring design that only kinda works and feeds back. Your suggestion will be tried and I will also come up with some other ideas as well.

[sluckey]

I will give this a try. The normal channel is already alot quieter than the EF86. I am admittedly concerned with how much more I am going to loose changing that resistor to a 3Meg without another gain stage.

I suspect it will be whisper quiet at the speaker.

[Jalmeida]

Please try the second part of my suggestion before quitting.

Tweak the mixing resistor (currently R45 220k) to a much larger value a la Fender. Try 3 meg ohms. That will direct more of the signal to the reverb driver and allow less signal to the dry circuit. The result may be usable reverb and a very clean-toned channel. I think any better outcome is impossible without more signal amplification than is possible with a 1 triode preamplifier design. If you have seen or heard one that does work well (Swart???) please share it!

I will have time tomorrow to dive back in on this, but in the meantime I did some more reverb circuit research. The Benson Monarch has a 12DW7 transformerless reverb that is integrated pretty simply and similar to mine. However the recover stage feeds after the volume pot into another triode of a 12AX7. This is not apples for apples not having another gain stage to recover, but does give me some other ideas to try in addition to your 3meg mixing resistor. Especially if the drive channel becomes too attenuated with a 3Meg mixing resistor.

So, if the 3meg resistor doesn’t work is there any reason I couldn’t move the mixing resistor to introduce wet signal before the pot after C11? I would still need a dwell pot.

This isn’t to avoid the suggestions as I fully intend to play around more. I am just anticipating possibly over attenuating the normal channel with the massive mixing resistor without another gain stage.

[ac427v]

I do admire your persistence, you are working hard on this to make it work! I am interested in hearing how it sounds if you move the mixing resistor to the plate of the preamp triode. That would place the entire reverb circuit before the volume control. I've never seen it done but you are trying to eek out every mv of signal to drive the reverb tube and the pi

I am not a fan of dwell pots for the same reason. They work by sending precious millivolts of signal to ground. You can't afford that!

I noticed that you have a .02 coupling cap (C27) after the recovery triode. Typically a .002 or .003 is used to prevent lower frequencies from the tank from being amplified by the power amp. That may contribute to the "howling" you mentioned in the original post. Please keep us posted on your creation. I'm learning a lot from your efforts!

[Jalmeida]

I do admire your persistence, you are working hard on this to make it work! I am interested in hearing how it sounds if you move the mixing resistor to the plate of the preamp triode. That would place the entire reverb circuit before the volume control. I've never seen it done but you are trying to eek out every mv of signal to drive the reverb tube and the pi

I am not a fan of dwell pots for the same reason. They work by sending precious millivolts of signal to ground. You can't afford that!

I noticed that you have a .02 coupling cap (C27) after the recovery triode. Typically a .002 or .003 is used to prevent lower frequencies from the tank from being amplified by the power amp. That may contribute to the "howling" you mentioned in the original post. Please keep us posted on your creation. I'm learning a lot from your efforts!

Yeah, I know I jumped in the deep end on this amp. And giving up means nothing learned. I reached my wits end chasing hums earlier on. But walking away when frustrated and coming back to it is really the only way.

For the dwell pot, I just need a tucked away trimmer pot where I can set the happy place and leave it. I could play with the 1Meg resistor value and leave it fixed. The trimmer pot is easier and allows me to play it a bunch tweak it if needed etc. I suspect this will fix the howl. Alternately I could try different value coupling caps in lieu of the .022uf. But in my mind a .022uf cap couldn’t cause the howl from the recovery side. I would expect a squeel. The howl with the original circuit integration from earlier in this post in all it’s behavior almost has to be a drive side issue with too much dwell. Because it was accumulative the louder i played and louder higher the reverb level was set. I stop playing and the howl would continue. If I turn the reverb level off it would stop but as soon as I turn it back up it would commence. UNLESS I waiting for the reverb springs to stop vibrating and then turn it up. And I thing the volume created a vibration that exasperated all of that. I at least want to rule out the too much dwell issue. And moving the the drive where you suggested proved that. The squeeling feedback currently after moving the drive integration I am thinking is from the mixing resistors not being correct as you have said. Again this is new territory to me too, so this is my theory.

On the AC15 build I will use a breakboard to test and prototype the circuit integration and values. I should have done that on the AC10 too.

[tubenit]

Just food for thought and discussion.  This is an unproven DRAFT idea of how I would probably approach an AC-15 & reverb build.  Not saying this should be done this way or that it would be "better".  Perhaps just another approach for consideration?


OR if you didn't want/need two channels ........... and are willing to use a 5879 instead of an EF86?   This is also a DRAFT unproven idea although I've built a HoSo56 that is similar to this some years ago. 


With respect,  Tubenit

[Jalmeida]

Just food for thought and discussion.  This is an unproven DRAFT idea of how I would probably approach an AC-15 & reverb build.  Not saying this should be done this way or that it would be "better".  Perhaps just another approach for consideration?


OR if you didn't want/need two channels ........... and are willing to use a 5879 instead of an EF86?   This is also a DRAFT unproven idea although I've built a HoSo56 that is similar to this some years ago. 


With respect,  Tubenit

Thanks for sharing. I think the cleanest would be if I redesigned the circuit instead of trying to scab on a circuit like I had been doing. However, I wanted preserve all the original JMI AC15 features(quirks included) which makes it a way more fickle path what I am trying to do. I really like the EF86 channel and the tremolo. And I do believe that of the two channels, the EF86 is the best path, albeit problematic. Feeding a EF86 preamp signal into a 12DW7 AU side to drive and then recovery may for the reverb will present even more issues than what I am doing on the AC10. While I would almost prefer driving from the 12AX7 VIB/Trem channel, i’m afraid that would mean a complete redesign just to get the VIB/Trem circuit to even remotely work. In my mind changing up the preamp tubes may change too many of the amp characteristics. Which the whole point of my exercise with the 12DW7 silliness is to avoid as much of that as possible. I really am a bonehead trying to do all this. Any engineer would look at this as an exercise in futility. But I look forward to a challenge 

I will keep studying those circuits and brainstorming. Luckily the AC15 build will be approached differently and may very well take a few years at that to get nailed down.

I didn’t any time to tinker with the AC10 last night, but will tonight. I have a few more ideas to try in addition to what was suggested here too.

[Jalmeida]

Please try the second part of my suggestion before quitting.

Tweak the mixing resistor (currently R45 220k) to a much larger value a la Fender. Try 3 meg ohms. That will direct more of the signal to the reverb driver and allow less signal to the dry circuit. The result may be usable reverb and a very clean-toned channel. I think any better outcome is impossible without more signal amplification than is possible with a 1 triode preamplifier design. If you have seen or heard one that does work well (Swart???) please share it!

So I posted this on the original AC10 build post for continuity.

But here is an update.


As for the reverb circuit, i have the reverb working and have tried pulling the dry signal the reverb drive from the other side of the Normal volume pot so that the drive is goes up and down with volume. I have tried multiple suggestions to try to get the reverb so that it doesn’t start the slow and perpetuating feedback induced with the reverb level above 1/2 up.

Here’s what I have tried and the results:

1. Changed C11 from 500pf to .022uf
2. Changed C24 from 1uf to .22uf.
  - These two cap changes made the howling feedback to a squeeling feedback.

3. Moved C24 dry signal from V2 plate over to between Normal volume wiper and R45.
- The reverb drive was very weak. Feedback squeel starts immediately.

4. Temporarily clipped in resistors to ground at R37 around the existing 1M varying from 750K all the way down to 330K.
-Feedback starts slower each resistor value down I go, but is still present.

5. Replaced R45 with 3.3M(ala Fender).
- extremely weak reverb drive (almost none at all)
 
6. Tried a 750K for R45.
-still weak reverb

7. Moved C24 dry signal feed back to plate.
- feedback seems to be gone with reverb. Need to fix a few other things before I am certain of the reverb being fixed. The dwell also seems better likely due to the capacitor changes not piping so many lows to the tank.

In summary…

I believe that the Swart style reverb circuit is a compromise. If I had a cascaded gain stage I believe I would be able to tie the dry signal drive after the volume and would be good with another gain stage before the PI. But because I go right into a PI from the reverb re-integration, I needing all signal I can to drive this with one tube.
So I will keep the reverb integrated at the plate and it will function like a stang alone reverb where I mix dry and wet signals. The nature of scabbing on a one tube reverb like I am I suppose versus adding another recovery tube. 🤷🏻‍♂️

 

[ac427v]

Yeh, Proving what doesn't work is not nearly as much fun as proving what does work.
The one idea I shared that doesn't seem to have appealed to you is changing C27 from .02 to .002.

That should help with the original howling, squealing feedback problem. From your description, I think you are amplifying too many low frequencies from the springs and passing them on to the power amp and speaker where they vibrate the springs even more. Causing feedback.
Good luck!

[tubeswell]

You have got the pan in a bag ay?

[Jalmeida]

Yeh, Proving what doesn't work is not nearly as much fun as proving what does work.
The one idea I shared that doesn't seem to have appealed to you is changing C27 from .02 to .002.

That should help with the original howling, squealing feedback problem. From your description, I think you are amplifying too many low frequencies from the springs and passing them on to the power amp and speaker where they vibrate the springs even more. Causing feedback.
Good luck!

I didn’t play with C27. Right now it seems the with the dry signal at the plate and the 750K mixing resistor there is no feedback. I didn’t play with the C27 capacitor after the recover because to my ears changing the 1uf C24 cap to .22 and C23 to .022 seems to tame the lows considerably. I guess I just am not understanding how C27 is effecting what is getting amplified being after the recovery stage. I have some more 68K resistors coming to replace the at the inputs because any time I move my input coax the skinny 1/2w ones break. When I change those I can play with that capacitor.

[Jalmeida]

You have got the pan in a bag ay?

Reverb tank is in a bag. Although I mounted through the bag and through the rubber dampened tank mounts too. I considered screwing down the bag with the tank inside loosely like fender does. Maybe put a piece of carpet under if I still have feedback.

[wsscott]

Tubeswell-I just wondered if you really need a bag, and what does it do for the reverb?  Would it affect low pitch (60hz type) hum? I've got a Gretsch/Supro 6165 Variety Plus, year 1966, amp that has a cardboard "box" cover over the tank, but nothing else.  And I haven't read anything that definitively says a bag is necessary.  Thoughts?

[wsscott]

It’s a Gretsch/Valco, not Supro even though Valco also built Supro’s.

[ac427v]

Everything coming out of the recovery stage gets amplified by the phase inverter and power amp and propagated by the speaker. Your circuit amplifies low frequency noise from the springs and the speaker transmits the sound back to the springs to cause acoustic feedback (howling). The recovery stage coupling cap cuts low frequencies to correct this.

[Jalmeida]

 

Everything coming out of the recovery stage gets amplified by the phase inverter and power amp and propagated by the speaker. Your circuit amplifies low frequency noise from the springs and the speaker transmits the sound back to the springs to cause acoustic feedback (howling). The recovery stage coupling cap cuts low frequencies to correct this.

Oh, I see what you’re getting at now. Yeah I was thinking just on the recovery stage and not on the successive stages.