I have two DC coupled cathode followers; one with 100k and the other with 47k off the cathode.
Both resistors are hot to the touch even when idling. Measurements show they are drawing less than half a watt. I changed one of them to 2W but it is too hot as well.
There is also fairly nasty hiss at high gain settings. The heaters are DC and regulated. Because of the poor power supply regulation, the voltage going in had to be raised to avoid dropout when the amp is pushing full power.
As a consequence, the rectifier and the regulator are running so hot, eggs could be fried on the heatsinks. In fact, the diode pack didn't even have a heatsink when I got it. I wonder if thermal noise is being generated there.
Otherwise, the amp works and sounds perfect. I know the regulator situation has to be remedied, and it will. But why are those resistors so hot?
Any ideas much appreciated.
Both resistors are hot to the touch even when idling. Measurements show they are drawing less than half a watt. I changed one of them to 2W but it is too hot as well.
There is also fairly nasty hiss at high gain settings. The heaters are DC and regulated. Because of the poor power supply regulation, the voltage going in had to be raised to avoid dropout when the amp is pushing full power.
As a consequence, the rectifier and the regulator are running so hot, eggs could be fried on the heatsinks. In fact, the diode pack didn't even have a heatsink when I got it. I wonder if thermal noise is being generated there.
Otherwise, the amp works and sounds perfect. I know the regulator situation has to be remedied, and it will. But why are those resistors so hot?
Any ideas much appreciated.
How did you measure the dissipation? Not clear to me which resistors you talk about; when you say 'off the cathode' do you mean the cathode resistor or some load resistor?
I see in one schematic a 100k cathode resistor with 212V across it - that would be less than 0.5W.
A 47k with 350V across it would dissipate some 2.6W and that one would definitely be hot.
What is the total current the amp draws from the supply? Did you measure that?
In the upper schematic, V3a output seems heavily attenuated and that could account for the high noise. But the circuit doesn't seem to be right - why the high attenuation?
jan
I see in one schematic a 100k cathode resistor with 212V across it - that would be less than 0.5W.
A 47k with 350V across it would dissipate some 2.6W and that one would definitely be hot.
What is the total current the amp draws from the supply? Did you measure that?
In the upper schematic, V3a output seems heavily attenuated and that could account for the high noise. But the circuit doesn't seem to be right - why the high attenuation?
jan
Last edited:
Thanks for answering. This is very bewildering.
C = 340
PS DC Amps = 8mA Two other 12AX7 dual triode gain stages are also being powered.
V across R25 100k on V3A is 167V = 0.28W That one is not noticably warm with 2W metal film resistor.
V across the 15k on V4B is 43v = 0.12W Also not warm with 3W resistor.
V across the 47k on V4B is 133V = 0.37W Too hot with the same 2W metal film resistor as R25.
The attenuation on V3 output is to bring the output to pro equipment level which I understand is about 2V. Tapping further down, or reducing the 2.2k to ~ 200R will bring it closer to instrument level (~100mV) at high gains and medium inputs.
That resistor is currently a 0.5W carbon pot.
C = 340
PS DC Amps = 8mA Two other 12AX7 dual triode gain stages are also being powered.
V across R25 100k on V3A is 167V = 0.28W That one is not noticably warm with 2W metal film resistor.
V across the 15k on V4B is 43v = 0.12W Also not warm with 3W resistor.
V across the 47k on V4B is 133V = 0.37W Too hot with the same 2W metal film resistor as R25.
The attenuation on V3 output is to bring the output to pro equipment level which I understand is about 2V. Tapping further down, or reducing the 2.2k to ~ 200R will bring it closer to instrument level (~100mV) at high gains and medium inputs.
That resistor is currently a 0.5W carbon pot.
Maybe your idea of 'too hot' is too restrictive. If you can barely keep your finger on it, that will be about 45 degrees Celcius which is not hot at all for a resistor.
Are you sure the resistor value is correct? It's easy to make a mistake with one of the color rings and be a factor of 10 off.
Still puzzled about that 1st stage. The attenuation on V3a appears some 45 times, that's a waste and may be the cause of the noise.
You should chuck the gain stage and use just the cathode follower, much better!
Edit: taking away C11 on V3b will give better performance and lower gain, that will help already a lot!
jan
It has no bands, but it's the size of a 1/4W. I think Mouser made a mistake. One time I got a bag with no resistors at all!
The level has to come down somehow. I don't know how else to do it. Perhaps the total resistance could come down, then readjust the divider. If 100k is faithful, and 47k clips half the waveform, then I guess ~85k would do no violence
Can't do that, that stage colors the sound and would upset the chain. Better to leave well enough alone.
I will try that.
Thank you so much. Back to rubbernecking....
It says 47k and measures the same. I looked up the code in Vishay Dale and it says it's 1W.
Ok so it's not 2W, but it's still more than twice the rating necessary, yet too hot.
Ok so it's not 2W, but it's still more than twice the rating necessary, yet too hot.
Either don't amplify (as Jan suggested, remove the gain element) or put a volume control at the input. Your scheme increases noise and output impedance.
Other thing to note is that cathode followers can oscillate easily- they're operating under 100% feedback. This is a common source of noise; you can't hear the RF, but you hear the effects. You want local bypassing of the plate, and grid stoppers placed physically as close to the grid pins as possible.
Thanks Sy for piping in.
There is a divider of sorts before the V3A and B. It's a 330k Resistor from the incoming signal to ground. Seems doable to some extent I would imagine.
I upped the suggested 100R to 4.7k in the second diagram, which includes the startup protection.
I see there is no gridstopper going into V4B. I had some oscillation previously due to a ground loop which has since been corrected (after one month of aggravation). Placing a 1M in series with the grid --at the pin--had staved off the oscillation to some degree, but since the oscillation is gone, I took it off.
What do you mean by local bypassing? Bypass the plate resistor, plate to gate, or plate to cathode?
Thanks again.
There is a divider of sorts before the V3A and B. It's a 330k Resistor from the incoming signal to ground. Seems doable to some extent I would imagine.
I upped the suggested 100R to 4.7k in the second diagram, which includes the startup protection.
I see there is no gridstopper going into V4B. I had some oscillation previously due to a ground loop which has since been corrected (after one month of aggravation). Placing a 1M in series with the grid --at the pin--had staved off the oscillation to some degree, but since the oscillation is gone, I took it off.
What do you mean by local bypassing? Bypass the plate resistor, plate to gate, or plate to cathode?
Thanks again.
CF plate to a local ground (e.g., the terminating end of that stage's grid leak- if there is one- and the bottom of that stage's cathode resistor).
I still think a quick 'free lunch' fix would be to remove the cathode decoupling cap on the first tube.
jan
jan
Question would be the new harmonic content since this is essentially an effects box application. Maybe better, maybe worse, surely lower gain.
Local bypass of the supply for stability ought to be large-ish, say 1u.
You can play with different cap values across the cathode resistor as a sort of tone control. Or even use a series RC for frequency response shaping.
You can play with different cap values across the cathode resistor as a sort of tone control. Or even use a series RC for frequency response shaping.
Now I see. I have never seen this technique anywhere in the literature (or didn't catch it).
This is starting to make sense, even though it's my first foray into tubes. In high gain tube amps, really nothing can be compromised. This was learned through hunting down the oscillation I had, especially the localization consideration. So the same could be applied to HT power. Interesting. This makes even more sense since C node is the first out of the supply, and feeds everything.
Also cheap sockets, dirty pins, lousy solder joints, longer than necesary leads, sloppy lead dress, and even one substandard component; these will not cut the mustard! If you already spent $800 for a tinkerers amp, may as well go all the way.
Yeah, if you're going to do it, do it right. 😀
If you don't have them already, I'd grab copies of the Morgan Jones books- this kind of thing is his bread and butter, and there's lots of gems like this explained in a really clear way.
If you don't have them already, I'd grab copies of the Morgan Jones books- this kind of thing is his bread and butter, and there's lots of gems like this explained in a really clear way.
This looks like a guitar amp loop send and return for effects, as such this really belongs in Instruments & Amplifiers. You may also get some insights from the amp gurus who hang out over there.
You may also get some insights from the amp gurus who hang out over there.
It is in fact an FX loop. I felt the original question was fundamental to tube amps.
I will go to Instr. & Amps from here.
Thanks.
🙂
- Status
- Not open for further replies.