Component values for triode half of 6GH8 gain stage?

[octal]

I'm using a 6GH8 (pentode side) for a cap coupled reverb driver. (Using the "Channel Road Amps" schematic.) I'd like to use the triode side for reverb recovery, but the tube's specs (rP 5400 ohms, transconductance 8500 micromohs) aren't close to any of the common triodes I'm familiar with. I was thinking about using the "set the plate load at 2X rp and then adjust cathode resistor until plate sits at 2/3 the HV method, but that gets me an 11K plate load, probably tons of current going through the tube, etc. Does anyone have any suggested values for the plate and cathode resistors? B+ is about 250V. I guess I could try to work out loadlines, but I'm hoping someone has been down this road before. Thanks!

[HotBluePlates]

I'm using a 6GH8 ... the tube's specs (rP 5400 ohms, transconductance 8500 micromohs) aren't close to any of the common triodes I'm familiar with. ...

6GH8 data sheet

The top graph of page 5 shows "the tube's specs" change a lot with plate current (and also plate voltage).  Amplification Factor will fall from the ~46 you expected to ~20 down around the 2-3mA of plate current you'd like to be using.

Not really a good fit for a Reverb Recovery gain stage, where a 12AX7 is closer to the amount of voltage gain needed.

I'm using a 6GH8 (pentode side) for a cap coupled reverb driver. (Using the "Channel Road Amps" schematic.) I'd like to use the triode side for reverb recovery ...

It's polite to point the Peanut Gallery to your sources. 

Channel Road Amps Reverb Driver article

I don't know if Swampertech ever built his amp or if 2deaf tried building the recovery stage shown.  100kΩ and 1.5kΩ for the 6GH8 triode side will probably work about as well as having a half-12AU7 using those values.

[octal]

Thanks for the speedy response HBP! I do see some strange "bunching" in the triode curves for this tube.  Is there a triode/pentode that would work better in this application? I'm at my limit with filament current draw, so using a 6gh8 for the driver and an extra 12ax7 for recovery would be a bit much.  I guess I could just ditch the pentode driver and go for a 12dw7... lo mu side for the driver and the high mu side for the recovery.

[HotBluePlates]

... Is there a triode/pentode that would work better in this application? I'm at my limit with filament current draw ...

I'll defer to others (perhaps 2deaf?) with more experience setting up the pentode driver.

I wonder if you should just use some other common sharp-cutoff pentode, and perhaps a low-noise FET for recovery if there's no budget left for a tube-heater.  We use FETs all the time in boost pedals with triode-like circuitry.  The weak signal communing out of the reverb tank is not unlike the signal from a guitar pickup (though generally even weaker).

[sluckey]

I've heard this good sounding reverb lately. Sound clip either here at Hoffman's or TAG, don't remember which thread. But I liked it enough to save. I downloaded it in January. Take a look...

[octal]

Hmm. It looks like the 6bm8 is triode strapped in the good sounding reverb.  I have to question- why attenuate the output of the reverb tank before amplifying it? Seems like it would be better to get the signal level out of the dirt and then cut it down to size post amplification. But, I do have 6BM8s on hand and the triode section is listed with an R-C amp chart in my Mullard book. The extra current draw of the 6bm8 filament might be pushing things. Definitely one to consider, and I believe you that it sounds good. Could also just do a SS recovery circuit like HBP suggested. I've got some things to mull over, I guess

[2deaf]

I don't know if Swampertech ever built his amp or if 2deaf tried building the recovery stage shown.  100kΩ and 1.5kΩ for the 6GH8 triode side will probably work about as well as having a half-12AU7 using those values.

I never bothered to try the triode for a recovery stage.  It's just like you say, it wouldn't have anywhere near enough gain.

I did build one where the triode and a transformer drove the tank and the pentode was the recovery stage.  That worked pretty good in a small amp at home.

[tubeswell]

...6GH8 ... triode side for reverb recovery, ...thinking about using the "set the plate load at 2X rp and then adjust cathode resistor until plate sits at 2/3 the HV method, but that gets me an 11K plate load, probably tons of current going through the tube, etc. ... B+ is about 250V.

Yes it does and Voltage gain is about 25 - See 11k load line example

Does anyone have any suggested values for the plate and cathode resistors?

As HBP has said, you can't do much with that transconductance - the grid curves are steepish, which limits the amount of plate voltage swing you can get for a given amount of grid voltage swing.


Attached a couple of other loadlines where I jacked up the B+ to 350 and made the load 33k. As you can see, even if you warm the bias up (to shift the idle point to the left, where the intervals in grid voltage are slightly larger), the voltage gain doesn't change much.

[Swampertech]

Hello HotBluesPlates, and thanks for the mention. Octal, I am attaching the preamp stage to my amplifier, "The Tremor-Verb". I needed a way to cut down on the 9 pin tubes in this PR chassis, so I did research on the 12DW7. I found, as you did, the Channel RoadAmps information. I plugged most of their design into my work, and then posted the schematic here for the experts to see. I want to give mucho thanks to Sluckey and 2deaf for their input. As of today I have not built this amp due to some health issues, but soon plan to restart my effort. Octal, you will be wise to take the advise of Sluckey and 2deaf, and several others here. I hope this helps, and if I can provide any further info just let me know.

[2deaf]

Hello HotBluesPlates, and thanks for the mention. Octal, I am attaching the preamp stage to my amplifier, "The Tremor-Verb".

I added a 220K pull-down resistor to the 12DW7 driver so that the Reverb Send doesn't charge up when no tank is connected. 

The driver was designed for the input inductance of a type F tank.  I'd have to look into it to see about a type E tank.





   

[2deaf]

Here is a redesigned driver for a type E reverb tank.
 

[kagliostro]

I was (wrongly) thinking that the reverb circuitation used in some old FBT italian amps

will be of interest

(they use the pentode of a triode/pentode tube to drive the reverb tank without a transformer)

but the triode was used as gain tube that feeds the pentode

so not useful for you

https://el34world.com/charts/Schematics/files/Fbt/Fbt_500_r.pdf

https://el34world.com/charts/Schematics/files/Fbt/Fbt_500_r2.pdf

https://el34world.com/charts/Schematics/files/Fbt/Fbt_1200_r.pdf

Franco

[octal]

I decided to try to find out how important the 'constant current pentode driver' is to the reverb sound. I used the 6gh8A pentode driver and then fed the output of the reverb tank into an separate guitar amp (just to see how it would sound.) I then switched to using a 12dw7 driver to compare. I swear the 12dw7 sounds a bit cleaner and has a slightly darker sound. I didn't find the difference compelling either way. I then wired up the 12ax7 side of the 12dw7 for reverb recovery. It sounds good, I'm pretty happy with it. I also found a bag of LND150s, so I may go for a 'beefed up" driver stage with both sections of a 12au7 in parallel and then SS recovery. We'll see, I'm playing around with it.  As it stands, it's plenty of reverb for any genre but surf/rockabilly or dream pop.

[2deaf]

...so I may go for a 'beefed up" driver stage with both sections of a 12au7 in parallel...

I'm curious about this beefed up driver.  If this is a transformerless driver, parallel 12AU7's give very little extra drive mA's while chewing up way more power supply mA's.

[octal]

Is the limiting factor the max voltage swing of the stage vs running out of drive current? I would think 2X the triode sections would idle at twice the current, the plate resistor would be halved and therefore the stage could source twice the current?

I was also considering switching to a 6cg7 and cascading two stages (like some Ampeg designs) to provide plenty of signal to drive the triode section connected to the reverb tank.  I'll take some measurements with a scope and see what my signal levels are, how hard the driver stage can be pushed w/o clipping, etc. These questions could be answered with theory I'm sure,  but sometimes it's fun to just experiment.

[DummyLoad]

6DX8 , 6EB8 , 6GN8, 6JA8 , 6JW8 may be candidates.

a 6JA8 is the one spin up first. socket pinning, pentode gm, pentode plate diss., etc., triode mu and gm - just seems like they'd work well together.

--pete

[2deaf]

Is the limiting factor the max voltage swing of the stage vs running out of drive current?

Yes.  That and the average plate dissipation.

I can't directly measure the AC voltage across the input coil, but I can measure the AC voltage across the series combination of the coupling capacitor, the resistor, and the input coil.  I can then calculate the impedance of the series combination and derive the current.  The current is the same through all three components, so I now know the current through the input coil. 

No matter what else happens, it is the voltage swing at the plate that determines how much current will be going through the input coil.

You're not going to get very much out of your driver unless you exceed the maximum plate dissipation during part of the cycle.  This doesn't scare me as long as the idle condition is below the maximum dissipation and the average dissipation is also at or below the maximum dissipation.  The average dissipation is a function of volts, amps, and time.  My calculus days are far behind me, so I just draw it out and look at it with my thumb at arm's length.

   

[2deaf]

I would think 2X the triode sections would idle at twice the current, the plate resistor would be halved and therefore the stage could source twice the current?

Yes, but unfortunately most of that extra current would be going through the plate resistor that has been halved.  There can be a slightly larger voltage swing at the plate because the pitch of the AC load line is not as steep, but this only results in a slight increase in current through the input coil.

[octal]

6DX8 , 6EB8 , 6GN8, 6JA8 , 6JW8 may be candidates.

a 6JA8 is the one spin up first. socket pinning, pentode gm, pentode plate diss., etc., triode mu and gm - just seems like they'd work well together.

--pete

Just in case anyone is looking at this post down the road, I tried two NOS GE 6gn8 (tubes marked 6gn8 [etched in glass] and 6gn8/6eb8 silkscreened on it) and the triode section had decent gain but was kind of microphonic in both tubes. At the end of the day, these are TV tubes not intended for audio, so ultra low microphonics probably were never part of the design brief. I tried a 12dw7 cap coupled driver (OK) but finally ended up going a different direction and used a 6bm8 and transformer driven reverb. I think the pentode + cap coupled/transformerless route could work well, but I was having a hard time getting enough level in my build to drive the reverb stage. The pentode+ transformer version seemed to have the best input sensitivity. Thanks to everyone who offered advice and suggestions!

[AmberB]

What reverb transformer did you use with the 6BM8?

[octal]

I had a small hammond universal output transformer...it seemed to work well over a pretty wide range of impedances, but I settled on 5.6K to 8ohms. I'm running the 6BM8 biased very cold and my B+ in this amp is under 250V, so you might want to try a different setup if you're running a higher B+ or run the 6BM8 off a separate power supply node with an additional dropping resistor /bypass cap to keep the voltage to the 6BM8 reasonable.  The standard 22K Fender reverb transformer makes sense (sort of) for doubled up 12AT7s, but doesn't seem like a great load for a proper output tube.

[jabguit]

I don't know if Swampertech ever built his amp or if 2deaf tried building the recovery stage shown.  100kΩ and 1.5kΩ for the 6GH8 triode side will probably work about as well as having a half-12AU7 using those values.

I never bothered to try the triode for a recovery stage.  It's just like you say, it wouldn't have anywhere near enough gain.

.

I built this and for me it has all the gain needed. This is the best sounding reverb I've ever heard - the notes are not swamped by reverberation, rather the reverb follows the decay of the notes. No it won't get Dick Dale type sounds, but then I never cared for that sound anyway....


cheers

[HotBluePlates]

I built this and for me it has all the gain needed. This is the best sounding reverb I've ever heard - the notes are not swamped by reverberation, rather the reverb follows the decay of the notes. No it won't get Dick Dale type sounds, but then I never cared for that sound anyway....

Glad it's working out for you!

The cause of the over-intense Fender reverb is the linear taper of the 100kΩ Reverb control.  Switch to an audio taper, and the reverb will be much reduced.

Having a 12AX7 in the Fender circuit, and then using an audio taper pot to get less reverb results in a better signal-to-noise ratio, as the high stage gain is amplifying a very weak reverb signal.