Garnet 15R Tube reverb unit

[Toxophilite]

Hi
I recently picked up the chassis of a Garnet 15R tube reverb unit.
It came with a tiny 2 spring tank
I recapped it as the filter caps were quite unhappy and hooked it up to a Gibbs 2 spring long tank, as the one it came with (which I think might not have been original) , whilst it produced tonnes of reverb , also produced lots of ugly metallic feedback (I also tried it with a small 3 spring tank that sounded decent but the Gibbs was better)
The reverb sounds good now but there's a few things I'd like to address if possible.


-It really preamps and warms up the signal from the guitar. I can live with the extra volume but i don't like how muddy and woofy it sounds compared to the sound straight into my deluxe lite conversion. I tried putting 12AT7s and 12AU7s in the preamp section and got less volume but also less reverb,


- I'd like to get more reverb. The Gibbs tank sounds great but not super loud. It would be nice to go beyond a tasteful amount of reverb into full on surf /effect territory occasionally(ala my Super Reverb) Can I accomplish this?


- Maybe a way to mix the dry and wet signals, and or adjust the dry one more


- I've noticed there seems to be a bit of sizzly distortion to the reverb too could that be a tank impedance mismatch?


the schematic is chopped in the top right corner
There is a .1mfd cap between the 12au7 plate and the 1 meg lin pot (labelled 'colour' basically a reverb tone pot)

[PRR]

What are the impedances of the tanks you tried? (Give DC resistance if the Z can't be found from the part number.)

What voltages do you get at tube pins?

[Toxophilite]

Some crazy voltages on the plates of the 12au7  280 pin 1(straight to it from the B+ see schematic) and 334 pin 6
on the 12AX7  250 pin 1 and   154 pin 6


the tanks:


-Japanese tank that came with it (small 2 spring) measures - 3.5 ohms IN    2.5k out


long Gibbs 1122 tank from hammond L-100 measures  3.8ohms  IN     173 ohms OUT

[PRR]

How about grid and cathode voltages? (The plate voltages are not "crazy", as you note, they are essentially B+ so not much use.)

The 4-ohm input tanks are reasonable for a transformer-coupled plan. The 1122's output may be a bit low-level, but get the rest of it happy first.

[Toxophilite]

12AX7
pin 2- .6 mv   
     3   2.9 v


pin 7 - 3.2 mv
 pin 8 - 1.3 v


12au7
pin 2 - 56v
    3  - 80v
   
pin   7 - 51mv
       8 - 15v


My limited experience makes the high voltages on the 12au7 plates seem "crazy" I know it's B+ but I've only encountered a few amp preamps and the range was more in the 100 -200 range on those.

[Toxophilite]

the power supply has been recapped(filter caps) and the resitors there too as one had drifted.
I also replaced  the other electrolytics on the cathodes and the .1 mfd non electrolytic caps
I haven't done the smaller ceramic caps or the  2 .047 caps
Most of the resistors seem to read bang on or very close though I haven't checked each and everyone out of circuit.
THe tubes are good

[Toxophilite]

I've noticed a few differences between my unit and the schematic that Garnet very nicely sent me


 On my unit :


-The resistor to the right of the line going to the .1 mfd cap and then to the output is 10k not 1.5k like like the schematic


- The line coming back from the reverb tank goes straight to the ft switch and then straight to the 1meg linear reverb control pot. There is no 100k resistor to ground, nor is there a .001 mfd cap before the 1 meg pot


- there is no .001 bypass cap on the 470k mixer resistor coming from the input to the grid of the 12ax7 (where the reverb tank signal is sent through a 470k resistor and 470pf cap)


- also on the schematic the filter cap at point 'C' is 30mfd. on the can cap I replaced under chassis it was 20 mfd


These could have been design changes or someone monkeying with the unit


Any ideas?

[Toxophilite]

Any insights into how I might obtain more reverb out of this unit?
And  flatter frequency response? (more treble and less heavy mid range and bass)


THanks

[Willabe]

You might not be able to get more verb out of as it doesn't have as much gain going to the tank as a Fender.

Both the Fender AB763's and the standalone have more gain stages before the verb tank than the Garnet.

But you could try changing the cathode (K) R on the verb recovery stage from 5.6K to 1.5K might give a little boost of verb? Maybe take the 470K grid R on the recovery tube 1st? (Pull the 470pF bypass cap with it.)

As for less bottom, more treble try changing the K bypass caps from 25uF/25v to 5uF or even down to 1uF.

I'd try the K bypass cap on the verb recovery stage 1st.


                Brad        

[PRR]

Get a tank with part-number something like *AC*** (8 Ohm input, 10K output), and possibly *AC3*** (long decay).

However AntiqueElectronics is not showing any AC3, so it is not common. AB3 (2K output, long) is a very standard spec, such as:

MOD 4AB3C1B -- Long (16 ¾") 2 Spring Unit , Long Decay · Input Impedance 8 Ohms, Output Impedance 2250 Ohms  $18.50
https://www.tubesandmore.com/products/reverb_tanks?page=0%2C0

It may require a complete re-think, maybe more tubes, to be a "great all-purpose" reverb. It has most of the infrastructure to be a Fender stand-alone reverb, a known-good plan. Use 6AQ5 for the power-bottle instead of 6F6 (or El84 as in modern copycats).

[Willabe]

You could also try lifting 1 end of the 680pF cap that's across the verb transformer and see if that gives you more treble on the verb.


                 Brad   

[Toxophilite]

I was thinking about a rebuild
As is obvious from looking at the chassis Garnet had a standard chassis that they used for many applications
This one has all the holes etc that one would need for a couple different types of amps or a fender style reverb. It's actually pretty funny
It almost looks like they were running low and grabbed one from an amp pile and re-stencilled the front as it looks like it's had things removed (that are NOT on the original schematic)
There's a space for the output transformer, maybe a tube recitifier or another can cap. maybe a 6L6 or pair of 6V6s
and then a couple more holes for either more preamp tubes or a pair of EL84s


-I should've specified. the reverb sound is bright enough and it even has a colour(tone) control to roll of the high end
It's the dry sound that's changed dramatically, a LOT of gain and midrange, high end cut. LOUD and mushy


- i was under the (perhaps mistaken) impression that the 2  470k resistors going to the grid of the 12AX7 from the reverb and the dry sound were mixing resistors. Yes? No?? Could I put a smaller one on the reverb say 220K to achieve more effect or do they need to be the same size?


- as far as gain, from what I can see, the dry sound goes in one side of the 12AX7, comes out, goes into the 12au7 and then into the transformer and tank. out into a pot then back into the other side of the 12ax7 with the dry sound right from the 1/4" input mixed in , then out into the other side of the 12au7 and out to the output. That's how i see it, once again I could be very wrong. It does boost the guitar signal ALOT!


For fun I tried swapping out the 10k resistor to ground (just before the the output)for a 2.2k(closer to the schematics 1.5k) and of course got a dramatically reduced signal(more like plugging straight into my amp)but still with unfortunate midrange emphasis and lack of top end.

[Willabe]

Could I put a smaller one on the reverb say 220K to achieve more effect or do they need to be the same size?

Yes you can make either of them what you want.

Maybe take the 470K grid R on the recovery tube 1st? (Pull the 470pF bypass cap with it.)

 

As for less bottom, more treble try changing the K bypass caps from 25uF/25v to 5uF or even down to 1uF.

I'd try the K bypass cap on the verb recovery stage 1st.

Since you figured out that both the wet and dry signal are being mixed at this tube using a smaller value K bypass cap will lessen the bass from the dry signal. It probably wont effect the wet signal much because the low end is striped out by using small value coupling caps so the verb is not muddy.

You can also make the .001 bypass cap that's on that 470K dry signal mix R smaller, this will cut the dry signal mix. 

If this gets your treb/bass problem corrected then fiddle with that 10K R (that you changed to 2.2K) until you get the right amount of signal to feed your amp. Or make it a 25KA pot.

You could also make the 470K dry signal mix R a 1MA pot. Garnet did not give the player a way to adjust the wet/dry mix, a pot there will do that for you.


                               Brad       

[Willabe]

Since your not sure what tank came with that unit originally, I'd still try playing with the 680pF cap across the verb trany primary.

Garnet tweaked that value with a certain tank that is unknown to you.

Fender sometimes used a 120pF cap in the same place 680pF is getting kinda big when it comes to bleeding off high treble end. There's a lot of chime/sparkle up in that range.

You might like some or all of that back?


                  Brad     

[terminalgs]

- i was under the (perhaps mistaken) impression that the 2  470k resistors going to the grid of the 12AX7 from the reverb and the dry sound were mixing resistors. Yes? No?? Could I put a smaller one on the reverb say 220K to achieve more effect or do they need to be the same size?

ignoring the tone of the reverb for the moment, and focusing on gain levels.  It sounds like the issue is the max achievable ratio of dry:wet doesn't have enough 'wet'.   


You are on the right track to fiddle with those 470k mix resistors.  grab a 1M pot  and temporarily replace the two 470K's (center lug on grid, outer lugs on far sides 470K's), and then see if a different mix will get you want you want.


You could also elect to mix the dry and wet on the other side of the recovery tube,  that way the recovery tube's output (as governed by the level pot) is 100% wet, and your dry is closer to unity (just under) if you feed it directly to the CF's grid.  that would require a good bit of reworking between the recovery triode and the CF,  and would end up looking more like the Fender 6G15 in that area...


most tube recovery reverbs feed the grid directly from the tank with no attenuation between.  Those are tanks with 2K-10K Zouts.  maybe the original Garnet  spec'd smaller tank produced an output with enough umph (voltage&current) to stand up to the level pot attenuator before recovery?

[Toxophilite]

Well I've been monkeying around
i tried a 1 meg mixing pot between the 2 470k mixing resistors and the grid of the recovery stage
That still gave me rather muted top end despite getting a better mix of dry to reverb, it also did some funny things with the reverb control pot, buzzing, strange drops in volume etc.


I tried bypassing both 470K resistors and just using the 1 meg pot . That produced ALOT more high end, some feedback and other oddities and even more crazy behaviour from the 1 meg reverb control pot


I tried running the dry sound into the grid of the 12au7 and bypassing the 12ax7 recovery stage..it was SUPER quiet then, not useable


The best results I've had was by taking the bypass 470 pf cap off the reverb return line. This made the guitar dry signal get instantly brighter, much more like it's straight-into-the-amp sound but with more gain. (i don't mind the gain as much as I mind the muddy honky dead sound)


Then as a further experiment I replaced the 470k resistor on the reverb return line to 270K and that gave me more reverb


So basically I removed the 470k resistor along with it's 470pf bypass cap and put a 270k resistor in their place.


Willabe- my unit differs from the schematic in that it has NO .001 caps anywhere. I did think of adding them to see what they do.

[Willabe]

Willabe- my unit differs from the schematic in that it has NO .001 caps anywhere.

There's your dry signal mud!   

Both of the mix R's should have a smaller value cap across them (in schemo) to tune the treble/bass response.

A signal through a large series resistance kills top end, the larger the series resistance the more top end you loose, so he added those 2 caps. If you keep that 470K R as the mixer for the dry you probably need some small value of cap in parallel with it so the high treble will bypass that R. It's a hi-pass filter, cap passes hi's, R restricts what's lower than the caps 6db roll off knee.

Fender AB763 had a 1M (2.2M, 3.3M in some models) with a 10pF cap across to add back the top end lost by the R.

There's more than 1 way to skin a cut, but which will sound best? There's several things to try and see what works best.

If it were me, I'd put the 470pF cap back across the wet mix R AND put the cap back across the dry mix R like Garnet did. You need it for the high end information in the reverb.

Try .001, if too much bottom end mud on dry signal, try 500pF (470pF) still to much try 250pF, ect.

And I'd still try changing the K bypass R to a smaller value on the verb recovery tube. Garnet might have used all 25uF/25v bypass caps because he had them in stock and tuned the circuit with coupling and bypass caps.

These are 2 easy ways to trim the dry signal bass end. It might work best to do both.

It seems like somebody was in that unit before you and tweaked for what they wanted, like a gain boost?

If you like the gain boost change the 10K tail R on the CF to a 10K or 20K pot and put a 1K R between that pot's 'O' end and ground so when you turn the 'gain boost' pot all the way down the circuit is back to stock.

The best results I've had was by taking the bypass 470 pf cap off the reverb return line. This made the guitar dry signal get instantly brighter.

That's because there's a 2nd path to ground for the dry signal's treble. With that 470pF cap in place the dry signal can bypass the 470K mix R and go to the verb pot, then to ground, the lower you have the verb pot turned down the easer it is to bleed off the dry signal to ground.

Without that 470pF cap the dry signal will always 'see' at least 470K between it and ground, no matter where the verb pot is set.


                        Brad          

[Toxophilite]

Oh and I tried disconnecting the 680of cap on the reverb transformer input and ut's really hard to tell if it made a discernible difference.

That's neat stuff. i didn't remove any of those .001 caps, they just weren't there!
With what I've done:

I can get the dry sound sounding alot better and a some more reverb but nothing like the ventures-esque glory that is achieved by the super reverb on 10
Humourously i only really use that setting on 2 songs. Shot in the dark by Mancini and a song of middle-eastern sounding beguine of Mine called Frankincense which has a sitar-esque solo.

I'll monkey with your suggestions regarding the bypass caps

The full on fender reverb unit looks pretty simple. It's only a 3 tube affair with not many components

[Willabe]

I didn't remove any of those .001 caps, they just weren't there!

I bet it left the factory with them (some value anyways).

Oh and I tried disconnecting the 680of cap on the reverb transformer input and ut's really hard to tell if it made a discernible difference.

Ok, it was worth trying.

I can get the dry sound sounding alot better and a some more reverb but nothing like the ventures-esque glory that is achieved by the super reverb on 10

Well, that's going in the right direction.   

You can try this;

Temp in a 1M pot for each of the 470K mixer R's AND put the bypass caps across them. Wire them like an adjustable R, use 1 end and wiper only, then put cap across these 2 connections. Play around with the dry side bypass cap value after you find a nice dry/wet signal mix.

Hopefully you can increase the dry R and decrease the wet R to get more verb. Less dry will give move verb without even increasing the verb.

Then if you want to try for more verb, you might be able to change the verb recovery stage K R to say 1.5K, or even 820 ohm, along with changing it's bypass cap to, like 5uF or even less. This will give a little more gain without out amplifying the low end.

You might even be able to change that recovery tubes plate R from 100K to 180K or 220K.

Doing all these things together might get you some more verb. Try different combinations of these things to see what works best.


                          Brad     

[Willabe]

I tried running the dry sound into the grid of the 12au7 and bypassing the 12ax7 recovery stage..it was SUPER quiet then, not useable.

Do you mean, the volume output level was too low? For the dry? The wet? Both?

Did you take out the 470K dry mix R when you did this? You should have had a TON of reverb then, even if you took out the 470K dry mix R.

The signal from the verb tanks output is less than a guitars output from it's PUP's. So, Garnet has the 470K dry mix R to also knock down the dry signal to the mix/recovery gain stage grid because the dry gets amplified there as well as the wet signal.

Fender's stand alone unit has NO series R from the input jack to the CF's grid for the dry signal. (If you run the input signal straight into the CF's grid it doesn't amplified.)

Either way, if you did or didn't try it again but also move the output wire that goes from the junction of the 22K and 1.5K R's to the K so it goes straight from the K with NO resistance to the .1 cap. And bypass/lift 1 end of the dry mix R (470K) and its bypass cap.

See if that works.

                   Brad     

[Toxophilite]

Hi Willabe
Thanks for all the cool ideas
When I did run the dry to the 12AU7 I did have tonnes of reverb, in fact it would start howling at around 6 on the reverb pot.
I ran the dry signal straight from the 68k resistor but it was very very quiet, much quieter than straight into my amp.
 One thing I did that may have been a problem was i moved the reverb control pot to between the 12ax7 and the 12au7 and ran the tank right into the 12ax7 grid.
I did then try it in the original configuration but running the dry signal straight to the 12au7 with similiar results

Out of curiousity what do you mean by "K"    and  "CF" if was arranged differentky I might start to get hungry!!(KFC)


And when you refer to the fender stand alone unit are you referring to the original 3 tube 60s version of the fender unit or the post 70s 4 tube version?


Also has anyone compared these units? the 60s version certainly looks like an easier build

[Willabe]

Out of curiousity what do you mean by "K" and "CF"

But you could try changing the cathode (K) R on the verb recovery stage

K= cathode, CF= cathode follower. 

And when you refer to the fender stand alone unit are you referring to the original 3 tube 60s version of the fender unit

Yes, 6G15. That's the classic (pre CBS) stand alone model.

I'm sorry but I'm not understanding some of your answers to my questions. (I'm not sure if your not understanding me or are just skimming through what I'm posting?) I'll try again.

I ran the dry signal straight from the 68k resistor but it was very very quiet, much quieter than straight into my amp.

Do you mean, the volume output level was too low? For the dry? The wet? Both? Did you take out the 470K dry mix R when you did this?

So no dry signal 470K mixer R? And by very quite you mean very low signal output volume, not very low noise? And what's to low in volume, if that's what you meant, the dry or the wet or both?

One thing I did that may have been a problem was i moved the reverb control pot to between the 12ax7 and the 12au7 and ran the tank right into the 12ax7 grid.

I did then try it in the original configuration but running the dry signal straight to the 12au7 with similiar results

So you put the verb pot back before the recovery tube wired to it's grid? And was the 470K dry mix R still in the series going to the CF's grid or was it taken out/bypassed?

When you moved the 1M verb pot after the verb recovery tube did you leave the wet 470K mix R in place from grid to ground as a grid leak (grid return) R? On the 6G15 Fender left the recovery stage grid leak out as the verb tanks output transducer coil acts as the R. Most guys add a 220K grid leak there.

It would probably be best to try any of these things 1 at a time and see how they sound/work. If 1 sounds good, leave it in place, then try another.


                           Brad     

[Toxophilite]

I appreciate the help and I try to your response as best I can
i think the internet , email ect is fraught with opportunities for miscommunication, especially when discussing these sorts of things


My response regarding the volume of reverb etc was about trying the dry sound straight to the 12au7 and skipping the 12ax7 gain stage where it was originally mixed with the reverb. I didn't use the 470k resistor , I took it straight from the 68k resistor that both inputs have on them .(Perhaps that was a mistake) the results of this were:


-Tonnnes of reverb, but very little dry sound, I had to turn my amp up to 4 to get a quiet sound which on my super reverb would normally be too loud for a very loud drummer!
I tried this with teh reverb control pot moved between the 12AX7 and 12au7 and with it in it's original postion right after the ft switch and tank


I've also discovered that different values at the 2  470k mixing resistors tends to make the reverb volume control into a strange mixing control with the loudest dry sound being achieved in the 12 noon position (or around 5 on a dial). THis is of course more noticeable the lower you go in value and i have only adjusting the wet sound so far.




I did try a 1meg pot in that position with much the same results as putting in different resistors, some satisfactory results but also strange results from the reverb control pot.




Sorry for the long response I just wanted to make sure I was being clear about what I was saying.
Once again I really appreciate the help and insights!

[Willabe]

I appreciate the help and I try to your response as best I can
i think the internet , email ect is fraught with opportunities for miscommunication, especially when discussing these sorts of things

Yes this is true and happens.   

My response regarding the volume of reverb etc was about trying the dry sound straight to the 12au7 and skipping the 12ax7 gain stage where it was originally mixed with the reverb. I didn't use the 470k resistor , I took it straight from the 68k resistor that both inputs have on them .(Perhaps that was a mistake)

Ok, so no (dry) 470k mix R, good. You could try moving the that wire from the 2x68K R's junction to the input jack tip, but I don't know there will be much difference?

I've also discovered that different values at the 2  470k mixing resistors tends to make the reverb volume control into a strange mixing control with the loudest dry sound being achieved in the 12 noon position (or around 5 on a dial). THis is of course more noticeable the lower you go in value and i have only adjusting the wet sound so far.

I did try a 1meg pot in that position with much the same results as putting in different resistors, some satisfactory results but also strange results from the reverb control pot.

Did you put bypass caps across the 2 mix R's when you did this?

Sorry for the long response I just wanted to make sure I was being clear about what I was saying.

Not a problem, thank you for explaining.


                       Brad        

[Willabe]

Here's a 2nd version of the Garnet reverb.

Looks like he changed the output and got rid of the mixer R's into the recovery gain stage.


                Brad     

[Toxophilite]

Yes I have that schematic too. Garnet sent me a PDF with 3 different incarnations
Mine is the middle one. 73-74
I guess in that one he mixes them right at the output with the dry signal taken after the first 12ax7 gain stage but before the 12au7 and the tank. It might be worth it, trying that wiring scheme. Apparently it was done for a reason.
I wonder what the advantages/dofferences would be? If any.
I guess one would put a pot on the dry output of the 12ax7 that goes to the output.

[Willabe]

Yes I have that schematic too. Garnet sent me a PDF with 3 different incarnations

Oh man, your holding out on us.       Will you please post the 3rd version that we haven't seen yet? (The earliest version?)

Apparently it was done for a reason.
I wonder what the advantages/dofferences would be? If any.

Yes I think maybe he was hearing from players about how it sounded and he changed it? 

Differences? How about the overall wet/dry mix and total volume output? That's most of what he changed, the post tank circuit. Just what your having problems with.

Although it seems someone already moded that unit before you and to be fair it would probably be best to set it back to stock 1st and see how it sounds. Then you could mess with it if it still wasn't what you like.


                  Brad     

[Toxophilite]

That's funny
i actually tried to post the PDF I was sent but it was too big, and as a result I lost the entire contents of the post I had written and had to write it again! So I figured out how to export one page of a pdf as a jpeg and only posted the relevant one
Here's the 1st one as requested

[Toxophilite]

Well I have it relatively close to stock


-I have a 4.7K resistor instead of the 10k that had replaced the 1.5 K (I see this as affecting the overall output not the reverb level per se)
- I have a 330 pf bypass cap on the dry signal being mixed in with the wet so there is a bypass cap, not it's true the original .001 but I don't have one of those on hand and the electronics store that has caps with the voltages needed for tube amps is about 30 min drive (lazy me)
- I also don't have the .001 on the reverb line prior to the reverb level 1 meg pot .
I think as was mentioned previously my problem with reverb level probably has a fair amount to do with my tank,which though it is a nice Gibbs tank, it's not the original. Though it was fun messing around with the unit in general and I learned a bunch of stuff.


What I don't know (and maybe I'll ask Garnet) is what the specs were on the original tank as I'm oretty sure the one that came with it was not original. I could be wrong, it does produce tonnes of reverb,,,just tonnes of evil sounding reverb!

[PRR]

> the dry sound that's changed dramatically

There's a LOT of stuff happening in the "dry" path.

Contrast to the Fender, which has one huge-headroom cathode follower.

You got plenty of PT and space. Garnet is not as Collectable as it should be. Build-quality looks like "we need this toNIGHT!!" I say build the Fender Standalone boinger in there. True, that may be over-the-top for your style, but it has two Wet knobs, you can dial-back.

Or the Garnet 1975, which is better than what you have, though still a bit messy on the dry path.

[DummyLoad]

I say build the Fender Standalone boinger in there. True, that may be over-the-top for your style, but it has two Wet knobs, you can dial-back.

second that.

--pete

[Toxophilite]

Yes
I've been leaning that direction myself
It'd likely be a fun build too, low component count!

Can I use the transformer that's in there, or is it only for a 12AX7 style driver?

I have a few single ended hefty looking output transformers used for EL84's  They all seem to be 4ohms output impedance
BUt i think I asked this question before and i was told it's nota  big deal


Thanks

[sluckey]

I have a few single ended hefty looking output transformers used for EL84's  They all seem to be 4ohms output impedance
BUt i think I asked this question before and i was told it's nota  big deal

I would use one of those. I'm using a 4Ω OT to drive my 8Ω reverb tank. Plenty of nice sounding verb.

[terminalgs]

You have lots of leeway with a reverb tank.  When you connect the OT's secondary to the tank's input, you are connecting it to an inductor. An Inductor's impedance rises in a linear fashion as the frequency rises.  at some freq. the tank is an 8ohm tank, but at some lower freq. it has a 4ohm impedance.  issues come up when you try to mate a 8ohm tank to a 10K OT, or a 2K tank to a 4ohm OT. 

[Toxophilite]

Hmm
Here's what I have and maybe you can tell me if it's going to work half decent
My Gibbs tanks measures about 2.3 ohms on the input with a multimeter


My output transformer I'm thinking of using is this one (from an ebay add)
this transformers where used with Alnico Speakers 5 Ohm Impedance. The driver was original a EL84  single ended
the type is BV 9060-503 (bigger Version)   
The transformers are made for SE Amps with EL84 tubes -

suggested Impedance for speaker  is 4-8 Ohm (original 5 Ohm)


the measurements are ca.: 60x50x22 mm
made for tube amps with 8-10 Watt