I was hoping I could just put a load on it, watch the voltage drop and get a current when the voltage hits such 'n such value. like -5% or something. I warned an ebay seller to think twice about selling an amp with the 6.3V winding buck wired in. He told me it's not that way, but I think so;
While I'm normally a Safety Sally, interpreted as "anything that can be done for safety must be done", in the case of 6.3VAC filament windings being connected into the primary, I'm much less strident.
Two possible insulation failure cases: shorting to other windings and shorting to core/case/chassis. The first is likely to cause some damage, but the transformer failure is already catastrophic, so a bad day but nobody hurt.
The second case is potentially more dangerous, and the degree of danger varies individually. I personally will not restore any vintage electronics without bringing it up to Class 1 standards, meaning proper 3-wire primary circuit - line thru fuse and switch, neutral floating, PE bolted to chassis. (I live in America.)
The question then becomes: does a 6.3VAC winding in the primary circuit add any (additional) element of potential danger to a Class 1 primary? Or, any additional danger element larger than would be added by an additional transformer? My judgement call would be no, but opinions will differ.
All good fortune,
Chris
Two possible insulation failure cases: shorting to other windings and shorting to core/case/chassis. The first is likely to cause some damage, but the transformer failure is already catastrophic, so a bad day but nobody hurt.
The second case is potentially more dangerous, and the degree of danger varies individually. I personally will not restore any vintage electronics without bringing it up to Class 1 standards, meaning proper 3-wire primary circuit - line thru fuse and switch, neutral floating, PE bolted to chassis. (I live in America.)
The question then becomes: does a 6.3VAC winding in the primary circuit add any (additional) element of potential danger to a Class 1 primary? Or, any additional danger element larger than would be added by an additional transformer? My judgement call would be no, but opinions will differ.
All good fortune,
Chris
Thinking about the 24V filament drop idea, it fails under transient conditions.
When I was a kid, I found the 12AX7 heater flare delightful to watch in a small AC/DC series heater stereo amp my father brought home. I've seen such flare in a 12AX7 as recently as this month. The phenom tells me the heater resistance isnt 12V/0.15A until up to temperature...
This amp has SS rectifiers for the B+ of ~400V, which means that voltage is up in a couple hundred milliseconds. The 12AX7 heaters would be 0V at this time. The 6BQ5 heaters are cold too - and when they heat up, the fun begins. As electrons come ripping off their cathodes, there's no grid bias to restrict them. So B+ goes ungh, 12AX7 heaters goes ungh, 6BQ5s go POW!
Glad I didnt try it. 6BQ5s be expensive... Perhaps the extra 6.3V winding can come to the rescue; voltage tripler to kick start the 12AX7 heaters, whose thermal time constant is faster than that of the big 6BQ5 cathodes, then as I get current flowing through the 6BQ5s, that drags the "pre-heated" 12AX7 heaters up to 24, reverse biasing the diodes in the voltage tripler - it only supplies current at startup.
Sigh. Maybe I should just use a voltage tripler to generate the bias; it'll be up with the 400V B+ - the output tubes will come on under a fully set biased condition. Put some 10Rs in each cathode to ground for "watchful waiting" and be done with it.
When I was a kid, I found the 12AX7 heater flare delightful to watch in a small AC/DC series heater stereo amp my father brought home. I've seen such flare in a 12AX7 as recently as this month. The phenom tells me the heater resistance isnt 12V/0.15A until up to temperature...
This amp has SS rectifiers for the B+ of ~400V, which means that voltage is up in a couple hundred milliseconds. The 12AX7 heaters would be 0V at this time. The 6BQ5 heaters are cold too - and when they heat up, the fun begins. As electrons come ripping off their cathodes, there's no grid bias to restrict them. So B+ goes ungh, 12AX7 heaters goes ungh, 6BQ5s go POW!
Glad I didnt try it. 6BQ5s be expensive... Perhaps the extra 6.3V winding can come to the rescue; voltage tripler to kick start the 12AX7 heaters, whose thermal time constant is faster than that of the big 6BQ5 cathodes, then as I get current flowing through the 6BQ5s, that drags the "pre-heated" 12AX7 heaters up to 24, reverse biasing the diodes in the voltage tripler - it only supplies current at startup.
Sigh. Maybe I should just use a voltage tripler to generate the bias; it'll be up with the 400V B+ - the output tubes will come on under a fully set biased condition. Put some 10Rs in each cathode to ground for "watchful waiting" and be done with it.
Last edited:
I have a Belcor FAX-200 that uses the 12AX7s in the preamp section to bias the 6BM8 output tubes... Uses the heaters as cathode resistors AFAIK. It's worked for years like this with instant B+ on the plates of all tubes...
Well, maybe the 6BQ5s wouldn't go POW! then. Maybe they'd just laugh it off and say; "do that again - it tickles!" That's an area of much dont know for me...
This thread was very timely for me since I have a similar issue.
I am building an original ECL86 Baby Huey, using the schematic from Marc (bandol), very close to the original from Yves ...
Dissident audio PP ECL86
... and I need the following voltages: 250Vdc (B+), -20Vdc (to CCS for driver/splitter). My transformer has a secondary 108-0-108, and I am using it without the CT and with a bridge rectifier.
I looked at the various suggestions for creating the negative rail, and adapted one of those, simulating it in LT Spice. The output looks OK, but I'd appreciate a sanity check on the component selection in case I've misunderstood something critical.
The Zener is 20V, but I will use what I have at hand which is 8V2 1W and 12V 1.3W in series. The load on the -12V rail is very low - just around 3mA x 2, plus whatever is consumed in the CCS, so no more than 20mA (hence R2 = 20k for the load on the -20V rail). I used R7 = 1.6k for the B+ load, approx 2 x 80mA. Looking through my box of spares, C1 is rated 450Vdc, C2 and C3 350 Vdc, C4 25Vdc, D1 - D5 will be 1N4007 and R3 will be 0.6W.
R1 is a bit more of an issue - 0.16A flows through it so it drops 24V, so 3.84W dissipation. is 10W enough for this one?
Any comments appreciated!
I am building an original ECL86 Baby Huey, using the schematic from Marc (bandol), very close to the original from Yves ...
Dissident audio PP ECL86
... and I need the following voltages: 250Vdc (B+), -20Vdc (to CCS for driver/splitter). My transformer has a secondary 108-0-108, and I am using it without the CT and with a bridge rectifier.
I looked at the various suggestions for creating the negative rail, and adapted one of those, simulating it in LT Spice. The output looks OK, but I'd appreciate a sanity check on the component selection in case I've misunderstood something critical.
The Zener is 20V, but I will use what I have at hand which is 8V2 1W and 12V 1.3W in series. The load on the -12V rail is very low - just around 3mA x 2, plus whatever is consumed in the CCS, so no more than 20mA (hence R2 = 20k for the load on the -20V rail). I used R7 = 1.6k for the B+ load, approx 2 x 80mA. Looking through my box of spares, C1 is rated 450Vdc, C2 and C3 350 Vdc, C4 25Vdc, D1 - D5 will be 1N4007 and R3 will be 0.6W.
R1 is a bit more of an issue - 0.16A flows through it so it drops 24V, so 3.84W dissipation. is 10W enough for this one?
Any comments appreciated!
Looks like you did a nice job here with your investigation of the above circuit. Yeah - 10W should do.
If you need to lose 24V off the top of the B+ as I do; but my problem is that I dont have the liberty to wire components any way I'd like, I'm stuck with the original cans connected to chassis ground and couldnt interrupt the B+ return current with the 150 easily, i.e. I'd have to physically add insulators.
I'll work with my fortune of having the extra heater winding. I stumbled on John Broskie's "Octupler" voltage supply design; I should certainly be able to do 3X...
If you need to lose 24V off the top of the B+ as I do; but my problem is that I dont have the liberty to wire components any way I'd like, I'm stuck with the original cans connected to chassis ground and couldnt interrupt the B+ return current with the 150 easily, i.e. I'd have to physically add insulators.
I'll work with my fortune of having the extra heater winding. I stumbled on John Broskie's "Octupler" voltage supply design; I should certainly be able to do 3X...
Well I did 3X with no issue. Generates a nice stable negative bias that I can adjust with the trim pot. For my 6BQ5 tubes, they wanted about 33mA a piece to eliminate the crossover distortion, apparent on an oscilloscope trace. Just turn the dial and no problem, adjust to 1.32V across a 10 Ohm resistor connecting all 4 cathodes to ground, All 4 cathodes were connected together in the original design and the 4 6BQ5s were purchased new, matched. With 33mA, I'm getting about 1% THD at 8W with ~14W maximum. B+ is ~400V, screens ~350.
Having adjusted all the internal voltages, I though I might play it loud, cause...that's what you do with an amplifier. I happened to be monitoring the cathode current. I noticed it sounding a little brittle; glancing at the DMM I see it rise from the quiescent setting, which I believe is normal when the amp is putting out power. Much to my chagrin, I then noticed the current going up, up, up so I stopped the music source. Still going up, up and I noticed one of my brand new tubes red plating, so I pulled the AC cord immediately
Letting it cool, I plugged it back in. 400V, 350V, 133 ma again, all stable sitting there. Played music again at the same level for 5 minutes. Cathode current going up as expected with music playing. Track ends, current settles back down to 133 mA again. Amplifier sits there looking at me; "What? You expect me to behave consistently? I've got about 5 different ways..."
How can this be? This isnt something fixed bias amps just do, is it? How does one tube just decide to take off into self-destruct mode one time and then a second time, no problem?
I thought I measured 370 at some point recent to the above event on the screen voltage "tap", but never again and dont know how that could be; it's a 2.7k resistor drop, so any more current should pull it lower. All the above with -12, -13V volts on the grids, consistent across all 4 tubes. No measured evidence of leakage across the output tubes coupling caps.
What things will I need to do to make this amp behave the same way, every time I turn it on?
Is it the original capacitors in the SS voltage doubler B+ supply; one time it's this capacitance, the next time it's that; sometimes it changes value on the fly? I settled on a 10 Ohm in series with one of the B+ tranny windings, to drop 420 to ~400, with 120VAC in. There's no hum and I believe the voltage rails are clean, even with the original caps still in place.
Anyone have a similar "sometimes, sometimes not" voltage stability with old tube amps? It's not enough to just set the voltages correctly according to the original schematic, you also need to exorcise the ghost in the thing. If you happen to buy one that haunted.
Any inputs very much appreciated!
Having adjusted all the internal voltages, I though I might play it loud, cause...that's what you do with an amplifier. I happened to be monitoring the cathode current. I noticed it sounding a little brittle; glancing at the DMM I see it rise from the quiescent setting, which I believe is normal when the amp is putting out power. Much to my chagrin, I then noticed the current going up, up, up so I stopped the music source. Still going up, up and I noticed one of my brand new tubes red plating, so I pulled the AC cord immediately
Letting it cool, I plugged it back in. 400V, 350V, 133 ma again, all stable sitting there. Played music again at the same level for 5 minutes. Cathode current going up as expected with music playing. Track ends, current settles back down to 133 mA again. Amplifier sits there looking at me; "What? You expect me to behave consistently? I've got about 5 different ways..."
How can this be? This isnt something fixed bias amps just do, is it? How does one tube just decide to take off into self-destruct mode one time and then a second time, no problem?
I thought I measured 370 at some point recent to the above event on the screen voltage "tap", but never again and dont know how that could be; it's a 2.7k resistor drop, so any more current should pull it lower. All the above with -12, -13V volts on the grids, consistent across all 4 tubes. No measured evidence of leakage across the output tubes coupling caps.
What things will I need to do to make this amp behave the same way, every time I turn it on?
Is it the original capacitors in the SS voltage doubler B+ supply; one time it's this capacitance, the next time it's that; sometimes it changes value on the fly? I settled on a 10 Ohm in series with one of the B+ tranny windings, to drop 420 to ~400, with 120VAC in. There's no hum and I believe the voltage rails are clean, even with the original caps still in place.
Anyone have a similar "sometimes, sometimes not" voltage stability with old tube amps? It's not enough to just set the voltages correctly according to the original schematic, you also need to exorcise the ghost in the thing. If you happen to buy one that haunted.
Any inputs very much appreciated!
Tubes do that LOL. That's why I tend to use automated grid bias... I used to have current meters on the amp and use 5 turn WW pots to set bias on every usage...
How large are your grid leak resistors?
How large are your grid leak resistors?
2.7k. I thought I measured a slight voltage drop across them as I was touring the amp with my DMM and wondered how that could be? Thanks!
2k7 from grid to bias voltage? Or are you talking grid stopper between grid and input? In that case what's the value of resistor between C- and grid?
So they're the grid "stoppers". 330K to the bias voltage. No sign of leakage through the .047s, near as I can tell. The "outside foil" marking is installed backward on the two caps of the failing channel, fwiw. The other channel has one correct, cant ID the direction of the other. Also, Fisher used two different cap brands on the cathode and plate of the phase splitter; a kind of "black drop" for one phase, a "Ero-foil" waxy job on the other. The Ero-foil connects to the tube that failed...
Those old "Ero"s are now all pretty much leaky. Staunch the flow and replace them. Fisher's and others are full of them, but entropy happens.
All good fortune,
Chris
All good fortune,
Chris
Last edited:
Someone should write a book about power supplies for tube amps...
I did that; replaced all the OEM coupling caps...and got me a nice red plate AGAIN on the same tube, 1/2 hour into operation at a background volume level.
So I tore out all the voltage tripler, used the spare 6.3V winding to buck the primary, got rid of the 15 Ohm dropping resistor I put in the voltage doubler, changed the 1st dropping resistor in the B+ string back to the OEM part (1.8k); ended up with 400V, 350V, 300V on the B+, Screen and 12AX7 supply. Dropping 13.7V across a 100 Ohm resistor, that is connecting all 4 cathodes together to ground as original.
Not a lot different than the voltages it was operating at with the negative supply at about -13V and cathodes grounded. It's been on for over an hour without that tube red-plating. Distortion however went up from 1 to 2% for 1kHz at 8W out. It doesnt make 14W at 4% anymore either; less power at more like 9% - with the typical crossover distortion you can see on the scope trace.
Murphy's law for tubes; buy a matched quad and one wont work quite the same as the other three, taking your project from not bad to unusable.
I hope that buck trick is actually a safe thing to do. Somehow I doubt UL would approve...
Whether or not the buck trick is safe depends how well insulated that “spare 6.3 V winding” is from all the other secondary windings. With proper insulation - the same as used between primary and secondary - it will be fine. Just wound in there on top of the others without so much as a layer of Mylar tape, well, that’s another story.
Your ghost in the machine might just be a tube that doesn’t like being run at 400 volts. Does “new” mean new production or old but unused?
Your ghost in the machine might just be a tube that doesn’t like being run at 400 volts. Does “new” mean new production or old but unused?
I'm pretty sure NOS Sovtek. I got a matched quad because I knew they'd all have the same bias, done one way or another. In order to get a little bit better distortion performance, I need to turn the tube current up. But I'm stuck with the voltage being 400, or 400+...