Plate resistors in series

[ethervalve]

Hi,

I'm trying to quiet down a hissy Garnet Rebel II tube amp and I've already eliminated the tubes as a problem (tho is was much worse before I replaced a microphonic 12ax7).

Next thing I was planning on trying was replacing the old carbon comp plate resistors with metal film. Only problem is that I don't have any 100Kohm metal film resistors on hand. Just wondering if I would be an OK practice to use two 1/2 watt metal film resistors in series to get the right value?

 

[PRR]

It will work.

Is an ugly hack for a classic old amp.

Perhaps I'd do it just to try. But really it might be best to put proper resistors on your shopping list.

100K 1W is *very* common in tube guitar amps, and old ones do go bad, so may as well get a 10-pack.

 

[CJ]

here is a Garnet website, but did not spot any schematics,

http://garnetamps.com/

threw together a PA for Zeplin, full props for that,  and the BTO amp,

carbon comps can absorb moisture and get noisy but using a larger wattage may circumvent this problem,

prices are getting crazy, especially for the 2 watt models,

 

[ethervalve]

Thanks for the info PRR and CJ!
I guess I'll wait until I can get correct discrete resistors.
My Garnet model # is G100G . The closest schematic I can seem to find by Google is the G100 (The Deputy -- supposedly a nod to Marshall). My amp is simpler than the Deputy though (it lacks the presence control) but it seems to share the exact same output section (12AX7 phase inverter; same filter cap values and output transformer taps etc.).
Since you folks seem familiar with Garnets, I was wondering what you think of the variable bias mod suggested on this page?:
http://www.lynx.bc.ca/~jc/garnetBias.html
I have a pretty nice NOS IRC slot adjust 50K pot I could use. 
I already fitted a 1-ohm resistor between pin 8 and ground on one of the 6L6GCs and the fixed resistor sets the bias way off the mark with a set of new Russian power tubes (though it's OK with the slightly tired original USA tubes).

 

[PRR]

> bias way off the mark

What mark? Which way?

"Cold" is not a big problem in these amps. 60mA (60mV across 1 Ohm) is hot but safe.

> what you think of the variable bias

""warning: if you don't have much experience don't do this on your own !!!""

Indeed.

I object to the plan posted, because it allows you to turn the bias to dead-zero.

While I agree the old Garnets *may* not bias-up right with non-1970 tubes, forking with the bias network is one way to kill a lot of tubes fast.

In other words: if you have to ask, you may not be ready.

 

[ethervalve]

Sorry for the slow reply on this.
Thanks for the great info PRR.
I think my initial measurement was off. Seems like the new JJs and Tung Sols bias just a bit cold. Should be ok to keep the bias resistor as is?

Incidentally, is it normal for for the amp to oscillate when measuring plate voltage (the screech goes away when I turn down the treble)?

here are my new measurements:

tube type |mV across 1-ohm|plate voltage|static dissipation watts
stock    |30            |438          |13.14
JJ        |32.5          |442          |14.37
Tung sol  |35            |438          |15.33
Sveltlana |44.5          |427          |19

(The JJ's have a tubestore.com 'perfect pair' (based on plate current, I believe) number #15 and the Tung Sols have #18; the Svetlanas #19)

 

[Gus]

I would keep the plate resistor a CC.  If you search you might find out about voltage ratings of resistors and there might be some truth in the tone of CC resistor when they have a higher voltage drop due to the CC non-linearity(distortion) from the voltage

 

[PRR]

> oscillate when measuring plate voltage

The total amplifier boosts-up signal to like 400V at the 6L6 plates. If the *least* bit of this sneaks back into the stages made for <0.1V signal, it is sure to oscillate.

I assume that whatever you are poking at the plate is radiating plate-signal into earlier stages. So don't do that!

It does shed doubt on your "idle" measurements which should be at NO signal. Screech is signal just as much as your fine fingers.

I think the trick is to measure at the OT CT, *NOT* at the actual plates. The DC will be the same either way (within 5V), unless the OT is fatally sick.

 

[ethervalve]

Ah thanks, that's a great tip. My voltage readings from the 6l6GC pin three were indeed a few volts lower than the OT CT. And the amp didn't squeal.
I found some 100K 1 watt carbon film (not CC, tho I would like to try A-B them sometime Gus) resistors today, so I'll see if they quiet down the hiss.

 

[ethervalve]

Hi again,
someone kindly sent me the proper schematics and it seems like there were maybe some mods done to the amp.
Does anyone have an opinion on the discrepancies (in red)?
I found that subbing in a 12au7 as the phase inverter cleaned up the hissy noise, but now the amp seems excessively bright.

 

[ethervalve]

And here's the pre-amp section. The bright channel is unusably treble-y right now.

 

[andyfromdenver]

Sounds like you should put it back to stock and go from there with a decent speaker and good tubes.
Then if it's too something or not enough something, post and people can recommend a tweak or two.
Their mods (in red) seem very trebl-y.

So, one vote for put it stock

And it looks like they removed the (-) feedback

Do you want a big clean machine? Or a crunchy amp?
I'll bet the original design is pretty clean, and they were trying to get some crunch?

 

[CJ]

yes that cathode bypass cap getting changed to 0.15 uf will cause degeneration on the bass frequencies, and the 0.002 uf will limit the transfer of bass frequencies to the next stage,

cathode resistor value getting changes to 2.7 k will also add a bit more bass degeneration, as the neg feedback voltage on bass freqs will be higher,

so i would change those 3 things back to stock one at a time and see where you like it,

those changes were made to try and make that brite channel into a Marshall (check Marshall schemo), but the 6L6 tubes are brighter than EL 34's and i bet the speakers are brighter than a celestion 4 x 4 so the mod is too brite,

and the Marshall is even too bright with those preamp values, so i bet the amp is super brite,

 

[ethervalve]

Ah I see! Thanks CJ and Andy. The Marshall theory makes sense.

I'm going to try changing those things back.

On the power side of things, just wondering what those phase inverter changes are supposed to accomplish?

In terms of intentions: I wouldn't mind if there were slightly more gain but I mostly just want to tame the brightness.

 

[PRR]

> what those phase inverter changes are supposed to accomplish?

Not much changed in driver/inverter.

What changed is that negative feedback from the speaker is eliminated.

Older amps often had no NFB. This means little damping on the speaker. The speaker bass resonance will be strong. In a small open-back cabinet this may balance for low levels. On high level the undamped speaker may slap. Also some companies moved to separate heads and closed-back speaker cabinets, where the bass resonance may be huge. NFB helps to tame things.

OTOH some of the Ampeg VT monsters ran no NFB.

This is for-sure a Tuning Point, and one that has to be tweaked for the specific speaker and playing style. I'd put the NFB back but add a switch (it is sufficient to cut the connection to the output, no need to rip all the parts out). To be fancy, add a pot.

Your lower channel currently has NO bass so these changes may not appear until you get some bass back.

NFB also affects the speaker's rise above 1KHz; with-NFB is less rise. So you want to try that. Depends some on the speaker.

 

[PRR]

I missed the 12AU7 12AX7 change.

The schematic says "components may be changed" and Gar was known to fine-tune each amp to the customer. You could ask him, but he passed away.

Guitar amps would "usually" use the AX7 here. The AU7 will be more mellow (lower gain and lower THD below overdrive).

The cathode bias resistor for 12AX7 with 100K plate load would normally be 1K-1.5K (Fenders all used 1.5K).

The cathode bias resistor for 12AU7 with 100K plate load would normally be around 5K (the 5.6K on the drawing).

Either tube will "work" with either resistor. (You see Gar used 1.5K with 12AU7 and total 200K loading on the cathodyne splitter.)

My first-thought is to run 12AX7 and 1.5K. A billion Fenders and Fender-clones work this way.

Bypassing this resistor changes gain and, with NFB, the amount of effective NFB. My first-thought is to use >10uFd (such as the 22uFd shown).

The other 5.6K and the 270r are the NFB divider. Change the 5.6k to change NFB. A value of 1K or less will not only be low-gain and too-clean to rock, it may excite supersonic squeals. I'd be thinking a 1K fixed resistor in series with a 50K Audio pot wired to give 51K at one end, ~~6k in the middle, and 1K at the other end of the rotation.

 

[ethervalve]

Thanks for sharing that excellent info PRR. I'd never considered variable negative feedback outside of a presence control.

My goal is to have it a little less bright and, if possible, to have a 'tighter' bass and a maybe bit more gain if possible.
(I'm used to playing through a 2xEL34 50 watt head with a closed back cab with 2 12 altec 417 8-H alnicos; I know the laws of physics prevent an open back 1x12 from sounding the same... but still).

Just to clarify, are the changes in red in line with what you're saying?

 

[walter]

The changes in red do look right. You might also like to change the .002 to .1uF going to point Y, and change the two .047s to .02uF between PI and output tubes to tighten up the low end.

 

[CJ]

full props on the monster 417-8H speaks, i had a pair of those and they were loud,

dropped the cab one night and the heavy magnet slipped a micron and we had some voice coil rubbing going on, darn it!

look at the back of the Live at Filmore East album,

you can do the pot on the feedback loop as PRR said, leave it, or turn it til you get your crunch dialed in and then sub in a fixed resistor,

you can even put a cap in the feedback line if you want to have fun with frequency response,

 

[ethervalve]

Thanks guys. I tried PRR's arrangement for the variable negative feedback and it's really huge improvement! As he predicted, on the low range of the pot, the amp is unstable. There seems to a real sweet spot so I'll probably take CJ's advice and fit a fixed resistor once I settle on a speaker.

It's still a tad bright tho. I haven't tackled the preamp de-modding or tried Walter's suggestion on the cap changes yet. Should the film caps be 600v?

PS nice catch on Fillmore East LP cover--never noticed that! I found another 417 locally for cheap that I hoped to put in the Garnet but it ended up having coil rub and a weird thing (cone cry?) where it almost gives a delay effect and cartoonish long sustain.