Help with SE 6V6 - excessive dissipation

[martin manning]

It'd be nice to know what the voltages were on the original. Isn't a tweed Princeton OT supposed to have an 8k primary impedance though? That would flatten the load line and make it more compatible with 300-350V B+. That'd make more sense with the 10k screen supply resistor as shown in the schematic. Dropping the B+ down to 250V (which would require about 1.8k) and keeping the 10k dropper would take the screen well below 200, I think. Also, I think new 5Y3's have less voltage drop than the old ones; add the higher line voltages we have now and B+ will be pretty high re the original.

MPM

[MWJB]

Martin wrote: "The way you have it (362V) it won't have much headroom because it is not center biased; my calc indicates only about 2.7W. With 250V B+, and 48mA you can get 5W."

The amp running 362vdc will always have more headroom than the 250vdc amp. most SE guitar amps are biased closer to cut-off than saturation.

With a stock 470ohm cathode resistor the amp will be fine up to & slightly over 400v on the plate. Most 6V6 will be plenty reliable dissipating 20W+ in a SE Champ/Princeton style amp. If you want to drop B+, the most transparent way to do so is with a reverse polarity zener diode installed between B+ centre tap and ground...Bruce Collins has posted about this numerous times, do a forum search. But really, you have no need to drop your B+ as it is.

[RWood]

It'd be nice to know what the voltages were on the original.
MPM

This says "Harvard" but the circuit is 5F2-a, with voltages:

http://www.schematicheaven.com/fende...6g10_schem.pdf

[Groover]

I think I'll drop the screen voltage a bit and see how it sounds.

Now with 4k7 dropping resistor:

B+ = 401v
Va = 391
Vg2= 382
Vk = 24

still getting 39mA Ik and 14 watts dissipation.

[martin manning]

OK, with the voltages from the Harvard, the load lines look like the attached, showing one for 5k and one for 8k Zpri. Thanks, RWood.

Idle current is 38mA for the plate and 4mA for the screen, from the voltages and resistances on the schematic. B+ is 330V, Screen 300V, 12.9W Pa, again matching the schematic voltages.

MWJB, higher B+ usually means more headroom- as long as the load and bias point are reasonable, no?

MPM

[tubeswell]

You can run the 6V6 quite cool with 5k and it sounds good - like when a 6V6 is plugged into a stand alone reverb 5k RT with a 1k cathode resistor. It comes out sweet sounding (and less distorted than if you use 8k and a hotter bias). This is at the expense of output power tho'.

[MWJB]

"MWJB, higher B+ usually means more headroom- as long as the load and bias point are reasonable, no?" Yes, but "reasonable" covers a lot of ground, 35mA+ plate current is usually fine, the only SE 6V6 guitar amps that I have seen that are biased anywhere near mid way between cut off & saturation are SF Champ amps (or other SE amps running SF Champ style voltages) running over 50mA, 400-ish volts at the plate, with a 470ohm cathode resistor. You will see that the original schem (not that I pay any attention to voltages on old Fender schems - they are ALL wrong) does not have the 6V6 biased to mid point, or anything like it.

Also consider that the lower the B+, for a given design, the lower the B+ voltages all the way down the chain, lower voltage on the preamp tube will also cut headroom.

[Steve Conner]

"Saturation" depends on the screen voltage and screen resistor, they act like a stop that sets the maximum current.

When designing a SE amp, you can (and probably should) choose those so that the tube ends up idling at half of the maximum current it can deliver, or halfway between saturation and cutoff as it was put here.

[martin manning]

MWJB, using the voltages on the Harvard schematic, and assuming an 8k impedance, the bias point is almost exactly half-way between maximum current and zero, isn't it? I couldn't find the specs for the transformer listed, but I did find a reference to 8k as the impedance for the Tweed Princeton.

I agree the voltages on the old schematics might not be truly representative. They are only claimed to be +/- 20%, and probably not a real statistical average, but who knows. Edit: The screen node doesn't look right... 40V over 10k = 4mA, and ~2mA is going to the 12AX7. For the load line, it's just B+, the impedance, and a bias of around 100% Pa Max.

MPM

[MWJB]

Voltages are not +/- 20%...it would not be possible to build an amplifier with that kind of variation. The "+/- 20%" quoted pertains to resistor & more importantly cap values.

Champs & similar SE guitar amps were never designed to be paragons of Class A operation, they were simply built to be cheap, entry level amps that would work for the warranty period without failing. All SE Fender champs & Princetons were fitted with the same value cathode resistor, irrespective of B+ voltage (a 100vdc range).

I am impressed that you found a reference to the 8K primary Z for the tweed Princeton, the 5F2A for example was hard on the OT...not many survived to this day, factory advice was to replace with a tolex Princeton P-P OT, use the whole primary & mate to an 8ohm speaker. Even SF Champ OTs can be 7-7.5K, 5.5K was more common in the 50's, but AC voltages were not listed, transformer numbers were not noted for tweed champs, #265 is listed for the 5F2A but no secondary load is specified...in short, the only way to be sure of primary Z is to measure one, or speak to someone who has.

[martin manning]

This schematic note (Princeton Reverb, from '64 based on the K-FD code?) makes it clear as to what the 20% is referring to. It's probably pretty arbitrary- big enough to cover line voltage variation as well as component tolerances.

MPM

[MWJB]

"This schematic note (Princeton Reverb, from '64 based on the K-FD code?) makes it clear as to what the 20% is referring to."

No, that's just the way that you are interpreting it, "voltages read to ground with an electronic voltmeter" and "values (not voltages) + or - 20%"

... then resistor tolerance, & cap voltage ratings. Back then cap tolerances were not as close as today.

If voltages were "+/- 20%" let's look at what that would mean for the AA1164:

6V6 plate voltage:
Quoted = 410v
-20% = 328v
+20% = 492v (...bang go your NOS 6V6s)

...you simply can't play/sell fixed bias amps built to that kind of tolerance variation (164vdc range).

In reality, old Fenders with the same PT & a reasonable current draw tend to run pretty close plate voltages, often +/- 2 or 3% at the most.

[martin manning]

Sorry for the thread drift, but I got to wondering what Fender is putting on modern schematics regarding the voltages specified. Here are two examples, Vibrolux Reverb and Blues DeVille. For the Vibrolux (and other reissues), all voltages +/-20% like on the old schematics. The DeVille is more specific +5/-10% for supply and bias, +/-20% for signal.

[MWJB]

Do the math +/- 20% on dc supply voltages = amps that are impossible to sell. You must see that?

De Ville B+ = 471vdc, +20% = 565vdc...doesn't happen. Vibrolux = 427vdc, +20% 512vdc...doesn't happen.

Vibrolux makes no specific mention of dc supply voltages, just says "voltages". The old Fender schems did not carry AC voltages.

[martin manning]

No argument here, +20% on B+ is nonsense, but the archaeology is kind of interesting.

The way I am reading it, the voltage tolerance on the old schemes was so big as to be useless. I'm betting it was an attempt to convey the message that values measured in the field wouldn't necessarily match those on the drawing. It just wasn't very well thought out.

If the 20% refers to component tolerances the numbers make more sense, but the evolution of the wording suggests to me that it was the voltages.

It's likely that it was just copied from model to model by a draftsman, and its entirely possible that whomever added "voltages" later didn't know the true intent of the note as it was originally written.

In the case of the Blues DeVille, it looks like someone finally gave it some thought and put the +5/-10 tolerance on the supply voltages (105% of 471 is 495). That still seems a little wide given that the line voltage is also specified.

MPM