6G15/5G15 Reverb - trashy thin distortion

[theunrulychef]

Finished my standalone reverb this week. So far, it's pretty quiet, though I'm getting the dreaded ground loop noise when I crank up the amp. I'll deal with that issue later. Currently, my problem is that the reverb sounds thin, trashy & a little distorted when the dwell & tone are higher. I can get a decent low level reverb, but when I want to go all out, super lush & spacey, it just sounds really bad. I've tried a few different things, but I could use a little direction on how to trouble shoot further.

I've swapped the tubes for at least 2 other known good ones each - only have one NOS 6K6 so I swapped with 2 other 6V6 I had. I connected it to another known good tank & it still sounded like trash. I have had the tank out of the cabinet & mounted inside. No bag, though the alternate tank I tried was in a bag. I re-flowed new solder on all of my solder connections again really quick just to make sure I didn't have a bad joint. I tested all of the resistors & they're within spec. The exception is that my DMM won't test the 2.2 megs so I tried to read them by paralleling another 2.2 meg resistor & it was reading a little high - around 1.7 megs instead of the 1.1 I was expecting. I don't know of a way to test the caps while mounted & I don't have a capacitance meter anyways (I didn't bridge caps of the same value across each one or anything).

PT, OT, choke & tank are all Weber units. The tank sounds fine connected to my Princeton Reverb, though not as nice as the old loosey goosey vintage tank - it's tighter.

It sounds fine when the mix is set to bypass, which is odd because the voltage at the screen of that 12AX7 triode is the only voltage that seemed to be really off (68 vdc instead of the schematic's 120 vdc). The PT is throwing out a little too much voltage, but nothing too crazy. I am thinking I'll order some zeners to play around with lowering it down. For now, here's the other voltages:

PT putting out 274 vac instead of 240 on the schematic
6.1 vac on the heaters
B+ 350
B+1 = 345 vdc (295 schematic)
B+2 = 290 vdc (250 schematic)

V1 - 12AX7
Pin 1 (plate) = 173 vdc (160 schematic)
Pin 3 (cathode) = 1.4 vdc (1.2 schematic)
Pin 6 (plate) = 290 vdc (250 schematic)
Pin 7 (screen) = 68 vdc (120 schematic)
Pin 8 (cathode) = 145 vdc (130 schematic)

V2 - 12AT7
Pin 1 (plate) = 127 vdc (105 schematic)
Pin 3 (cathode) = 2.25 vdc (1.8 schematic)
Pin 6 (plate) = 127 vdc (120 schematic)
Pin 8 (cathode) = 2.4 vdc (2 schematic)

V3 - 6K6
Pin 3 (plate) = 338 vdc (285 schematic)
Pin 4 (B+1) = 345 vdc (295 schematic)
Pin 8 (cathode) = 26 vdc (24 schematic)

Whew. Sorry for the long post. I'm kinda stuck here. I'm getting to the point that I really need to buy a scope & a reliable DMM - getting past the total novice point, but still have so so much to learn (from all of you guys, thanks!).

Thanks in advance for your help,
Jay in Philly

[Diablo]

I think you should try to figure out why pin 7 on V1 is so low a voltage. Since pin 1 is in the right ballpark, it says that B voltage is fine. You may have a bad 2.2M resistor from B to pin 7 or the 2.2M resistor from pin 7 to ground may be off value, or you may have a bad solder joint.

[tubeswell]

The PT is throwing out a little too much voltage, but nothing too crazy. I am thinking I'll order some zeners to play around with lowering it down. For now, here's the other voltages:

PT putting out 274 vac instead of 240 on the schematic
6.1 vac on the heaters
B+ 350
B+1 = 345 vdc (295 schematic)
B+2 = 290 vdc (250 schematic)

345 B+ is nothing to worry about. The 6G15 I built recently has a 342 B+ and it works perfectly fine.

V1 - 12AX7
Pin 1 (plate) = 173 vdc (160 schematic)
Pin 3 (cathode) = 1.4 vdc (1.2 schematic)
Pin 6 (plate) = 290 vdc (250 schematic)
Pin 7 (screen) = 68 vdc (120 schematic)
Pin 8 (cathode) = 145 vdc (130 schematic)

FWIW, Pin 7 isn't a screen, its a signal grid. My guess is that this 68VDC is probably the result of your biasing resistors not being within ballpark values. Try putting some other 2M2 ones in (maybe replace both of them). You might want to think about elevating the heater (if you haven't already done so). The heater-to-cathode voltage on the CF stage is probably getting too high for 'safe' operation. But maybe see how it responds to a couple of new 2M2 first.

V2 - 12AT7
Pin 1 (plate) = 127 vdc (105 schematic)
Pin 3 (cathode) = 2.25 vdc (1.8 schematic)
Pin 6 (plate) = 127 vdc (120 schematic)
Pin 8 (cathode) = 2.4 vdc (2 schematic)

you could always try a good quality 12AT7 (like a GE 12AT7WA black plate). Or if that doesn't do the trick maybe a 12k or 15k supply resistor to V1A to bring the plate down a smidgeon closer to the 6G15 schematic, which might help the tone a wee bit. V1B doesn't look too far from normal.

V3 - 6K6
Pin 3 (plate) = 338 vdc (285 schematic)
Pin 4 (B+1) = 345 vdc (295 schematic)
Pin 8 (cathode) = 26 vdc (24 schematic)

These aren't much to worry about. My 6G15 is running 342 on the screen (pin 4) and 338 on the plate (pin 3) and 25V on the cathode - much of a muchness - and as I said before it sounds fine.

Have you got all the correct resistor and cap values in the reverb channel. Is the tranny wired up the right way around (or back to front)? and is the pan plugged in back-to-front maybe?

[loudthud]

The 68V on pin 7 of V1 is caused by your meter altering the circuit when you are taking the reading. When you connect your meter it changes the biasing on the grid pulling it negative, that reduces the current through the tube. The reading on the cathode indicates the tube is operating properly. If you measure the grid with reference to (put the black lead on) the cathode, you will read something like -1.5V.

The same issue applies when measuring the grids of an LTPI.

[theunrulychef]

Gotta love this forum. I measured the grid in reference to the cathode & got -0.85 vdc. Is it safe to assume it's a little high because all of my other voltages are a little high? It's no wonder my voltage readings weren't getting any better when I swapped those 2M2s out late last night.

"FWIW, Pin 7 isn't a screen, its a signal grid." D'oh. I even was sitting there at the keyboard going "screen", "grid", "screen grid". I knew I should've looked up my basic tube terminology.

"you could always try a good quality 12AT7 (like a GE 12AT7WA black plate)...." That's one thing I was thinking - the 3 12AT7s I have are all Chinese made. Though, the one I've been using in my PR didn't sound any better in the 6G15 - the reverb in that amp is so good it sounds like magical golden thunder (or something reeeaaaally awesome like that). I just got an order of tubes & was trying to save a buck hoping I could get away with one of the AT7s I already had. Perhaps that & the tight sounding tank are the culprits? What pitfalls might arise if I put in a NOS 12AY7 or 12AU7? I've got quite a few of those around. Would playing a drum machine through the tank for a while to loosen it up a little, or is that too hair-brained to work?

I'll try playing around with the supply resistor to V1B as you suggested as well.

Thanks for your responses,
Jay

[theunrulychef]

So I didn't play around with the supply voltage resistor yet, but I did do some more tube swapping. A 12AY7 is a teency weency bit better, but still too brittle & distorting on the higher frequencies. A 12AU7 sounds much more in control & smooth while retaining the high end. It doesn't do the super crazy reverb that I was hoping for, but it's much more usable than with the 12AT7s. How concerned should I be about running an AU7 in an AT7s place in this application? It's running a little hotter than the AT7, but not too much. If I'm talking about 50% tube life, that's a sacrifice I can live with for now until I get some nicer AT7s & can experiment a little more. On the other hand, if you tell me I'm likely to burn it out in a couple hours, I'd be more inclined to go back to the AT7 or AY7.

Thanks,
Jay

[loudthud]

On the 6G15 schematic there is a voltage divider on the first driver stage plate with a 10K and 100K. Try moving the .01 cap to the plate side of the 100K.

[tubeswell]

On the 6G15 schematic there is a voltage divider on the first driver stage plate with a 10K and 100K. Try moving the .01 cap to the plate side of the 100K.

Or swap the 10k for a 12K or 15k (or a 20k-50k 1/2W trimpot wired as a variable resistor in series)

[Diablo]

I was playing around with my Tore-Vibe 5H15T, and you may find the information useful because of the similarities to your reverb. I hooked up the power to a variac and monitored B+ right after the choke - point A on the schematic. Changing B+ from 356 down to 290 volts did not have a big effect on tone. My conclusion is there is no reason to get a Zener to drop the volts. I swapped V1 (12AX7) to a 5965 tube (gain 47), and that also didn't have a big effect on tone. I swapped a V2 (12AT7) to a 5965 tube, and also no big effect. Swapped in a 6V6GT in place of the 6K6GT, no big effect.

I find that cranking up the dwell creates a fuzztone effect, and I don't really want that distortion. I want it clean. I also find that if I turn the volume knob down a little on the guitar, I can make the fuzztone go away. This was with a Les Paul (humbuckers). I then switched over to a Strat, and found there was no fuzztone effect with the guitar volume up all the way. So it seems to me that the Fender standalone reverb was designed for Fender single-coil pickups. Come to think of it, all I ever see are Fender guitars for surf music....

So, I've got to figure out the best way to keep from clipping the signal with my Les Paul without turning down the volume knob on the guitar.

So here's a followup. Subbed in a lower gain tube for the 12AT7 - used a 5963 (12AU7 equivalent). And now the fuzztone is completely gone at full dwell. I don't have as much reverb as before, but I wouldn't use the complete range anyway. Perhaps something with a little more gain than the 5963 could be used, but right now I like it the way it is.

[loudthud]

You could get more headroom out of that first drive stage if the cathode bypass cap was removed. Using a 12AU7 I would think plate and cathode resistors should be changed for better bias and headroom, and the voltage divider wouldn't be needed.

[Diablo]

You could get more headroom out of that first drive stage if the cathode bypass cap was removed. Using a 12AU7 I would think plate and cathode resistors should be changed for better bias and headroom, and the voltage divider wouldn't be needed.

Thanks Loudthud for the advice. So if I go back to the 12AT7, I can just lift one leg of the bypass cap for a test?

Do you have suggestions for plate and cathode resistor values to use with a 12AU7? I'm not knowledgable enough to figure those out myself.

[theunrulychef]

I just got done doing both loudthud & tubeswell's earlier suggestions plus a couple other things. First, I put back in the 12AT7 I was using before.

Moving the .01 cap over to the plate side of the voltage divider darkened the tone somewhat, but really put a distortion on the reverb. Actually sounded kinda bluesy, but not what I'm looking for in this unit.

Putting the .01 back, I played with the 10k value, going all the way up to 56k. None of those had the desired effect either. I really wanted to alter the tone going into the tank, as I could still hear that trash-can vibe after the stock tone knob did it's thing.

I decided to do a low pass filter going into the dwell control with a 50k pot & .01 cap. That worked well to mellow out the tone going into the tank, but seemed like it stole the balls of the reverb somewhat. Then I tried running off different cap values from the voltage divider junction to ground. That seemed to work a bit better & felt like it kept a little more "oomph" in the circuit. After going back & forth from .0033 to .1, I settled on .047. That kills the trash-can vibe but I can still get a bit more out of it.

I was just writing to describe what Diablo was talking about. Getting that fuzz on the front end of the reverb when the dwell's set too high. I want to be able to get alot of reverb delay out of it, but I hate that fuzz on the attack.

I also liked the tone on the 12AU7, but felt like there wasn't as much reverb on hand as I'd like from a unit that was designed to give me... reverb. I did also put back in the 6V6 to see if I couldn't bring up the reverb signal to level with the bypassed signal, but it just made the distortion on the pan more pronounced. It may have upped the output level slightly, but I can turn up the amp a click instead of having more distortion in the reverb.

I think I'm done with the iron for tonight - would be nice to play the damn thing for a bit. Please let us know how removing the bypass cap works. I'll have to try that tomorrow if you get some decent results.

Thanks again everyone for their help. It's getting closer.
-Jay

EDIT: I also tried a 12AY7 in place of the AT7 after my cap addition. It didn't really kill any of the unwanted fuzz, but felt a little lifeless so I put the AT7 back in.

[loudthud]

I think one of the things that hurts the drive circuit is the low B+. But there should be a way to get enough headroom to get rid of the fuzz.

Looking at the Resistance Coupled Amplifier section of the RCA tube manual a couple of good combinations of plate and cathode resistors for a 12AU7 are 47K/1500 and 100K/3000. These are listed for 300V B+ but will be close for +250V. I would think the voltage divider could go with the low gain of the 12AU7.

[Diablo]

Thanks Loudthud for sticking with us. I tried lifting the bypass cap on the 12AT7 (V2A), and it cut down a good bit of gain and reverb, but still left some of the fuzztone at high dwell. I swapped in the 5963 tube, and the dwell and mix needed to be maxed out for even "normal reverb", but no fuzztone.

Does increasing the cathode bias resistor value increase the headroom? I'm assuming I leave the plate resistor as is, and try to run the 12AT7.

[tubeswell]

Does increasing the cathode bias resistor value increase the headroom? I'm assuming I leave the plate resistor as is, and try to run the 12AT7.

Up to a point it does, but you really have to increase the B+ if you want the gain stage to go 'cleaner' all other things being equal. Remember that the rk value affects the bias, therefore provided that it is biased to ensure the idle point is roughly in the middle of where the grid voltage curves are the most evenly spaced along the load line (i.e. usually somewhere just to the left of the middle of the load line), then you will have the most available headroom for that particular voltage. See the attached pic (which is borrowed from Mr Valve Wizard)

The grid curves (the curvy black lines) show that the grid is biased at -1.5V w.r.t. the cathode (where the green lines meet the blue line, and the -1.5V black curvy line) with an HT/B+ of 300V with a 100k plate resistor. At this point the idle condition is that the plate is about 195V (where the vertical green line meets the horizontal/bottom axis), and the plate current is just under 1.1mA (where the horizontal green line meets the vertical/side axis). A grid voltage swing of 2V P-P will make the plate voltage swing from about 130V to about 250V (you get that from looking at where the vertical green line would move to if you slid it along the blue line by 1-volt's worth of curvy black lines either side of the -1.5V idle voltage).

Now if the rk gets too big, the bias voltage will move further to the right and ultimately (if rk is too big), the bias point will end up being down at the 'tail' end of the load line where the plate voltage is highest but the plate current is lowest, so the valve will operate in plate cut-off for about half of the grid signal swing.

So if we looked at that example pic again, if you increased the idle voltage so that the tube was idling at say -3V, then when you put a 2V P-P swing on the grid, there is hardly any where for the voltage to swing to (i.e. hardly any blue line left to the right of where the -3V curvy black line intersects with the blue line). The plate is in nearly in cut-off for the whole time that the grid is swinging down to the -4V part of the grid voltage swing. So that part of the signal won't get amplified by the plate and the signal will be clipped. The only way to overcome that would be to increase the B+. (But there is a limit to how much B+ voltage the tube can actually take without self-destructing)