Converting 50W Plexi to 20W 6V6

[shocki]

big_teee tell me how you like it !

The 250K grid resistors/PPIMV will work fine with 6V6s, use the 8ohm tap with a 16ohm cabinet.

Could you elaborate? Are you saying actually 100K would be correct but 250K will work fine? Or that 250K is correct?

[MWJB]

250K is "correct"...unless you prefer the sound of 100K, then that would be "correct" too. It's a question of taste with no bearing on reliability.

I don't recall seeing a 6V6 powered amp with 100K grid loads, but I'm not saying they aren't out there.

[shocki]

250K is "correct"...unless you prefer the sound of 100K, then that would be "correct" too. It's a question of taste with no bearing on reliability.

I don't recall seeing a 6V6 powered amp with 100K grid loads, but I'm not saying they aren't out there.

Here you see it
http://mhuss.com/Plexi6V6/Plexi6V6c.gif

And also like a said it is in the 6V6 spec sheet.

[big_teee]

Here is a TrainWreck with either or 6V6, EL34s.
It has the 220Ks.
http://www.ceriatone.com/images/layo...nCeriatone.jpg
So it sounds like the 220Ks will do either.
The Fender BFDR uses the 220Ks also.
T

[big_teee]

Also if you need 6V6s, tubes, they are on sale at TubeDepot.
SALE on Guitar Amp Kits, Select Tubes and Tools + NOS RCA 2A3 & 211
I may see if I can blowup a set of the Tung-Sol, with the 2204 Plate Voltage.

[bob p]

I don't recall seeing a 6V6 powered amp with 100K grid loads, but I'm not saying they aren't out there.

Yes, I think it's safe to say that Leo Fender never went there with the Western Electric circuits. Of course that's not to say that you won't find schematics on the web where people will do all sorts of interesting things.

When I see circuits that use novel values in modified circuits, I'm inclined to ask myself whether the person who drew the amp had enough knowledge to properly re-design the circuit to accommodate all of the changes he was making, whether he just substituted a few parts without realizing all of the other changes that might need to be made, or whether he had a stroke of genuis that nobody else had ever thought about.

As a reader I think it's important to sort through all this before thinking about building an amp from a schematic that I find posted on the web. It's important to be able to look at "novel circuits" and come to your own conclusion about them. There are plenty of good schematics out there. But with the proliferation of amateur amp builders there are plenty of bad ones too. It's important to be able to review a schematic with a critical eye. Sometimes you need to plot loadlines to really be sure whether novel values make sense or not.

[Gingertube]

This IS a reliability issue.

Gingertube's Essay - all you ever wanted to know about maximum grid one resistance but were afraid to ask.

Maximum Grid 1 Circuit Resistance for a 6V6 is quoted as:
For Fixed Bias = 100K
For Cathode Bias = 500K

Why the difference?
The difference is due to grid current.
There are 2 types of grid current:

1) the current which flows into the grid to cancel the negative charge which accumulates from some of the electron cloud "space charge" from the cathode gathering at the grid. This is a low level effect and is the basis for grid leak bias in small signal tubes, it ADDS to the bias and is NOT the sort of grid current which is of concern here.

2) the current which flows out of the grid to cancel the positive charge build up at the grid from residual gas ions. As electrons accelerate from cathode up to the screen and anode they collide with residual gas atoms. This collision is energetic enough that it can strip an outer orbit electron turning the gas atom into a positive charged ion, which accelerates back "down" the tube toward the grid and cathode. Some of these end up at the grid. The grid becomes positively charged. Current flows out of the grid to cancel this charge. This current flows through the grid 1 circuit resistance and thus develops a voltage across that grid 1 to 0V resistance.

This makes the grid more positive which SUBTRACTS from the bias (so you get more tube current, more collisions with residual gas atoms, more positive ions, more grid current, less bias , more tube current .... BOOM, thermal run away).

The subtraction in bias depends upon the amount of this grid current and the grid 1 to 0V circuit resistance.

With Cathode bias there is a mechanism which is opposing that reduction in bias (more tube current gives more bias) hence the larger value for max grid 1 resistance for cathode bias.

So lets look at that maximum grid1 resistance of 100K for fixed bias. That is the maximum SAFE grid 1 resistance when the anode dissipation is at the 14 Watt maximum and grid 2 (screen) dissipation is at the 2.2 Watt maximum.

Have a look at this datasheet for a KT88
http://www.tubezone.net/pdf/kt88new.pdf
Fixed bias maximum grid 1 resistance for fixed bias is stated as :
- 100K for Pa+Pg2 (anode + screen power dissipation) greater than 35 Watts
- 220 K for Pa +Pg2 less than or equal to 35 Watts
Note that for that tube Maximum Pa + Pg2 is 46 Watts Absolute maximum with 40 Watt design maximum.

The reason for this (2 values for max Rg1 in fixed bias) is that at lower idle current setting (less electrons zipping up tube) you get less of this gas ionization effect and less grid current because of it, so you can safely use a higher grid 1 resistance.

Hope this makes sense to you.

The consequence of all of this for a 6V6 is that 220K is perfectly OK as long as you bias the tube for a Pa + Pg2 of not more than 75% of maximum.

So when you do the cathode current times anode to cathode voltage calculation to arrive at Pa + Pg2 you want that to be no more than 12 Watts if using 220K (75% of 14 + 2 = 16 watts is 12 watts).

That is where that "popular wisdom" of biasing your output tubes at the 70% of max dissipation point comes from. It is valid because the "popular wisdom" Rg1 for a 6V6 is 220K.

If using cathode bias or fixed bias with 100K Rg1 you can safely bias the tube hotter.

Is this sort of theoretical post useful / valuable or should I just "pick up my marbles" (if I can find them) and go home"?

Cheers,
Ian

[Leo_Gnardo]

I may see if I can blowup a set of the Tung-Sol, with the 2204 Plate Voltage.

492V on the plate of an EH/TungSol - I can just about gar-on-tee they won't survive long.

I find the EH/TungSol handle a B+ less than 400 OK. 492V is a stretch even for JJ 6V6 but you'll have a much better chance they'll work without turning into wreckage.

[big_teee]

492V on the plate of an EH/TungSol - I can just about gar-on-tee they won't survive long.

I find the EH/TungSol handle a B+ less than 400 OK. 492V is a stretch even for JJ 6V6 but you'll have a much better chance they'll work without turning into wreckage.

I think the EHs are good to 475, and I think the Tung-Sols say 500.
I've played a bunch of 6L6s, and Sovtek 5881 WXTs, with no problems.
I did have a JJ KT77, short out and blow the B+ fuse 2 years ago.
T

[Leo_Gnardo]

Is this sort of theoretical post useful / valuable or should I just "pick up my marbles" (if I can find them) and go home"?

Useful/valuable? yes it's terrific! Just slapped a "like" on it. Certainly explains what I've been noticing as output tubes, especially large ones 6550 & KT88 go to their demise.

I used to swap out 220K bias/G1 resistors for 100K as a matter of course in many amps that crossed my repair bench. Think I'll go back to doing that.

Thank you Gingertube!

[Leo_Gnardo]

I think the EHs are good to 475, and I think the Tung-Sols say 500.
I've played a bunch of 6L6s, and Sovtek 5881 WXTs, with no problems.
I did have a JJ KT77, short out and blow the B+ fuse 2 years ago.
T

You can believe the bunkum about max plate voltage from New Sensor or wherever if you wish. If you have some used ones you don't mind melting down (pulls), be my guest. There are some who claim the TS are "better than" EH. Look past the paint & into the tubes. They look and act identical every time I encounter them. TS "better", yes they make a better profit for New Sensor. And neither one IME holds up to voltage in the high 400's. Just trying to save you some lunch money Terry. Maybe you'll be lucky and they'll work for you.

As far as JJ KT77, I tried 'em in 2006 with only one failure in a dozen tubes. They worked just fine as long as signal level below clipping - very hi fi. Clipping tone wasn't ear-friendly in guitar amps. Ran a quad for 5 years in Dyna MkII amps where they gave very satisfactory service, with plate voltage @ 460V.

[Steve Conner]

What Gingertube said. Pretty much every classic amp design ignored the maximum grid circuit resistance for fixed bias. IMO the result is that the tubes don't last as long as they ought to. As their positive grid current increases with age, they start to go into thermal runaway, but if you put the same tubes in an amp that had the correct grid circuit resistance, they might work again.

I always use 100k in my own builds. It does change the sound: the gain is reduced slightly because the PI plates are more heavily loaded. But a regular LTPI can still drive a pair of EL34s or 6L6s into clipping with the 100k.

The fun really starts with high-powered amps that use several power tubes in parallel. The required grid circuit resistance is per tube, so 3 tubes requires 33k. Pretty soon the impedance gets too low for an ordinary PI to drive, and you need to add cathode followers. Or use an interstage transformer, which presents a very low DC resistance to the grids. Or drive the cathodes instead ala Music Man.

[MWJB]

Leo Gnardo wrote: "You can believe the bunkum about max plate voltage from New Sensor or wherever if you wish. If you have some used ones you don't mind melting down (pulls), be my guest. There are some who claim the TS are "better than" EH. Look past the paint & into the tubes. They look and act identical every time I encounter them. TS "better", yes they make a better profit for New Sensor. And neither one IME holds up to voltage in the high 400's. Just trying to save you some lunch money Terry. Maybe you'll be lucky and they'll work for you."

EH 6V6 are stock fitment to 65 Princeton Reverbs and have to deal with 470vdc on the plates, which they do ...all day long, as long as you keep plate current <23mA (they don't like a whole lot more current than this in fixed bias at any voltage). JJs are good to over 500vdc in my experience (and will take curents comparable to NOS at lower voltages), but I'd play safe and stick to manufacturer's recommended max.

It would serve a modern manufacturer no purpose whatsoever to make 400v 6V6s or 460v 25W+ octals, simply because they would have a very limited market with respect to amps that could use such tubes.

[Gingertube]

Yes,
In 95+% of cases it is not the voltage which causes problems but over dissipation from too much current at that higher voltage.

If you go back and look at Black and White Television use of the 6V6 (Vertical Deflection Amplifier) they ran peak positive pulse of up to 1200V (15% duty cycle).
http://www.triodeel.com/6v6_p3.gif

Cheers,
Ian

[Gingertube]

I'm being a bad lad and infesting the forum with Solid State today. Feel free to tell me to stop it.
For your possible interest - this is how I drive the output tubes in my own design HiFi Amp. I can't help thinking that a Guitar Amp would benefit from a similar treatment.
Current Source loaded Mosfet source follower for each output tube.
Output Tubes biased at 90% of max. dissipation without any max Rg1 issues. The low impedance drive is also incredibly quiet. That grid current has random fluctuations in it which develops significant noise signal across any Rg1, the bigger the Rg1 the more noise (hiss). Also the high AC Impedance of the current source load isolates any residual noise on the -ve raw bias supply.

Cheers,
Ian