Substituting 12AX7 for 6SF5

[Apexelectric]

 I’m in the middle of modifying an old 1940s radio into a small guitar amp. I’m reusing the original chassis and would like to use all octal tubes like the original chassis had. Id like to substitute a 6SF5 for 1/2 of a 12AX7. No need to use dual triodes. I have a lot of holes in the original chassis that I’d like to plug up. Are there any adjustments for values of cathode resistors and plate resistors that need to be considered with the different tubes?

The amp is a single ended 6V6 type.

Thanks in advance.

[shooter]

from the data sheet looks "close enough" for DC operation, how it sounds

[PRR]

Just do it.

The published data is slightly different, but within the spread of 12AX7 production slop.

[Apexelectric]

Thanks guys!

[Apexelectric]

Should I just ground the metal shells, pin 1, to the PT ground on these tubes?

[DummyLoad]

yes, ground pin 1. ground to the signal ground or a ground lug on the socket.


--pete

[Apexelectric]

Any problem if they are strung together and grounded at one point collectively?

[DummyLoad]

that works.


--pete

[Apexelectric]

  Thanks

[Apexelectric]

Got it all up and running. Sounds pretty good. I decided to use a Supro Tremo-Verb SE 6V6 for the platform for this project minus the reverb section. My voltages are a tad high with the standard dropping resistor values. Right around 400 after the rectifier and dropping to 395 and about 325 at the pre amp node. These should be more like 360, 355 and dropping down to about 240 at the preamp.

The power transformer is the right one for this build. 325-0-325 70ma with a 5y3 rectifier.

Would the removal of the reverb load create a fairly substaintial increase in the preamp node voltage?

Reverb transformer and output tube were on node 2 and reverb return and preamp and preamp on node 3

I’m assuming my best option would be to just increase the value of the first dropping resistor to drop values across the board then adjust the other two accordingly. I have increased the last dropping resistor already from 15K to 30K but it’s still not enough to get my preamp voltages down to the original spec.

What’s the recommended process in a situation like this to get the voltages in line? Trial and error works but doesn’t seem like an effective way to go about it.

[PRR]

> Supro Tremo-Verb

Many questions, no schematic link. Who has all the Supros memorized?

> What’s the recommended process in a situation like this to get the voltages in line? Trial and error works but doesn’t seem like an effective way to go about it.

I'm leveling my chicken-shed. Trial and error is an effective way to do it. Put a brick under the low corner and re-check. Try a half-brick, a few shingles. Remember the chickens (or tubes) don't have Laser Levels.

I hear there is "math". You need numbers to start, and enough fingers to work them out. I don't know how much weight sucks each shed corner (or B+ node), and math is not glamorous action work. Bang 3 or 4 bricks in there and call it square.

[Apexelectric]

Same schematic as a Gretsch 6152. Not showing up under the Supro schematics in Hoffman’s library. Built by Valco.

[mresistor]

Well some well informed people here would say that grabbing a handfull of resistors and trying them out would be a good way to do it.

[Apexelectric]

Not trying to say that it’s a bad way to do it. You guys always seem so knowledgeable about doing things mathematically I figured this would be one of those type of processes. My bad for assuming. Didn’t mean to offend.

[PRR]

photo feb 03, 4 44 40 pm.jpg shows two values for V3 cathode voltage, and both look wrong. 12AT7 with 2k2 and 350V will be near 3mA (so 5V-6V at cathode). However a similar plan showed 12AX7 which would show nearer 2V (but not slam the spring so well).

V4a may be taking 1.5mA (1.2V/820r).

An added/lost 4.5mA in the two 500r (=1K) drop resistors makes like 5V difference. "350V" becomes 355V. Not much difference.

The effect at the "360V"(400V) node may be hard to measure.

The 400V suggests your PT or 5Y3 are working much better than Valco's. Is this a Valco PT or some "similar"? Is the 5Y3 healthy vintage or new-made?

That drawing's 17V for 660r at 6L6 cathode seems awful skimpy. It runs a 6L6 (and the 40-18019!) at half the power it could take. You pay for 6L6, you want to be up to 50mA-60mA to get all you paid for. 6V6 similar if not so extreme, say 40mA. This is also a far better match for a 5K load. I think the cathode resistor could be 250r for 6L6, 270-330r for 6V6. Another 25mA+ will pull-down the 400V, much more than +/- the reverb section.

[Apexelectric]

I have it built with just the 6V6 tube and not the 6L6 option. The transformers are the Classictone indicated on the schematic and are good for up to 15W output. I changed to a 1K, not a 750R, cathode resistor for the 6V6. The 750 had the output at around 13W. Plate to cathode voltage was around 365 with 27V across the cathode resistor.  Wound up with a 56K dropping resistor instead of the 15K for the last node and left the first two dropping resistors at 470R.

All the tubes are vintage. I used the old 5Y3G that was originally in the radio and bought four NOS 6SF5 tubes and an old 6V6 metal shell that I had from another project.

17v across the 6V6 cathode on the schematic seems low based on the readings I was getting on the bench.

[Apexelectric]

Did I get my output dissipation calculation correct? Or do I need to change my cathode resistor again?  It seems like if I go smaller than 1K that I will exceed the 12W capacity for the 6V6.

[shooter]

Did I get my output dissipation calculation correct

there's a bunch off #'s in this thread so I'm not sure, here's the math I learnt;

cathode volts (Vk) /cathode R (Rk) = tube current (I)
(take plate volts - cathode volts) * tube current = Pdis W

you will notice every time you change Rk, everything else changes and you have to repeat.  You will also notice there are diminishing returns.  Once you throw your hands up and say "I just want to play", then you can kinda sorta get a feel for how each of those changes change sound, if any.  ALSO, your bias, basically Vk will have an impact on how much more/less drive you need, again, how it sounds.

class A self biased typically wants to be abused to a point, I've set amps free into the world that were running from 82% low side to 105% high side max Pdis.  The amp on the bench, will probably dial in around 95-110% idle plate dis

[Apexelectric]

Numbers were 27V across the cathode resistor at 365V plate to cathode at the tube pins with a 1K cathode resistor.  So I did it the way you just described.

Thanks for the confirmation. PRR ball parked it but I think he was either thinking 6L6 or PP instead of SE on the cathode value.

Thanks as always

[sluckey]

Numbers were 27V across the cathode resistor at 365V plate to cathode at the tube pins with a 1K cathode resistor.

Those numbers give 9.855 watts.

[Apexelectric]

Correction.  Cathode voltage was 27 with the 750R resistor. Jumped up to around 30 when I changed it to the 1K. Dissipation was just under 12W.

[sluckey]

Jumped up to around 30 when I changed it to the 1K. Dissipation was just under 12W.

I get 10.95W (assuming plate voltage is still 365v). Every time you change cathode resistor values you must recheck plate voltage. Could you show us exactly how you calculate those numbers.

[shooter]

Jumped up to

Every time you change

that is why SE is so frustratingly FUN

once you get the 1st tube, start adding more, 4 I find is a practical limit
also once you got your calculator all warm, you can say, use the 1k, then calc || R values to go with the 1K to quickly dial in, for less than say $5 you can probably find a 827ohm 3W R 

[Apexelectric]

I did this all yesterday so I’m kind of going on memory on the exact numbers. I was just confirming that my formulation method was correct.

I checked my notes and it was 369 plate volts and 30v across the 1k cathode resistor. Right around 11W dissipation.

[sluckey]

I checked my notes and it was 369 plate volts and 30v across the 1k cathode resistor. Right around 11W dissipation.

Makes sense to me.