Thanks for the drawing. It makes it easier to follow your thinking.
First new question, in the signal circuit drawing, one end of the cathode on each of P1 and P2 (I'm numbering left to right) are grounded and in the battery drawing it's just P1. Whichever is the case you are not using a negative bias on the control grids. Is running P1 at zero bias and P2 at Positive Bias intended?
First new question, in the signal circuit drawing, one end of the cathode on each of P1 and P2 (I'm numbering left to right) are grounded and in the battery drawing it's just P1. Whichever is the case you are not using a negative bias on the control grids. Is running P1 at zero bias and P2 at Positive Bias intended?
I managed to reduce a lot of the oscillation by converting the PI to triode mode. Now it even plays clean at lower volume settings without the tremolo.
There is still some motorboating at the last 10% of the volume pot.
I haven't tried the alternative heater setting yet, maybe tomorrow. For now I'm quite happy with the result.
There is still some motorboating at the last 10% of the volume pot.
I haven't tried the alternative heater setting yet, maybe tomorrow. For now I'm quite happy with the result.
Yeah, it sounded pretty good to my ears in the video.
I still have this notion that the motorboating comes from feedback through the common cathode string. . . . . but that's just a wild guess based on some limited experience with another circuit . . . . . long ago . . . . . . . in a place far, far away . . . . . . . . . : )
I still have this notion that the motorboating comes from feedback through the common cathode string. . . . . but that's just a wild guess based on some limited experience with another circuit . . . . . long ago . . . . . . . in a place far, far away . . . . . . . . . : )
The heating current is perhaps 100mA. Each cathode current is under 1mA. Since all tubework is +/-20%, it is reasonable to neglect the small cathode current adding to the large heating current.
Or: the heating resistance is about 12 Ohms. The cathode internal impedance (1/Gm) is over 1K Ohms. Only 1% of what happens at the internal cathode interface appears at the heater.
The mutual coupling is small for most situations. In a high-gain chain (such as a guitar amp!!) you may have to deal with it. Capacitors may be wanted. Note that you want to well-bypass the *heating* resistance so this may be a far larger cap than we could get in 1920, even large today (thousands of uFd for full bass forward response).
Small filament tubes were usually designed to have grid returned to the more negative end of the filament. One end of the filament gives Vgk of zero, the other end Vgk is -1.2V, average -0.6V. The grid is negative of most of the filament. Yes, the near-zero end does pass some grid current. In a HOT tube like 12AX7, zero Vgk can be significant grid current. The filament tubes were not so tightly wound, or highly heated, grid current was not a big deal.
The Russian Rod Tubes use side-rods, not a 'fence' IN the electron stream. Grid current is apparently quite small even for Vgk positive.
Thanks for the input PRR, this stuff interests me and I was hoping you might chime in.
Not to argue, but with his stated 31mA heating current, I get 387Ω heating resistance. Does that make it more influential?
Another thing I noticed is that in the signal circuit schematic it shows neither output tube has a cathode connection to ground. Probably just a mis-print but that's all I have to work with . . . . . . . So stuff like that is bothering me.
A full schematic of both heating and signal together would likely have made me pack up and be quiet some time ago.
Not to argue, but with his stated 31mA heating current, I get 387Ω heating resistance. Does that make it more influential?
Another thing I noticed is that in the signal circuit schematic it shows neither output tube has a cathode connection to ground. Probably just a mis-print but that's all I have to work with . . . . . . . So stuff like that is bothering me.
A full schematic of both heating and signal together would likely have made me pack up and be quiet some time ago.
Not sure what plan you cite. This one has a full path.
Attachments
I think what might be messing up with the signal has something to do with the bias of the output tubes. One is at a higher bias, in the beginning of the series string with the leak resistor connected to another filament level to compensate this. While the second one is at the filament voltage with a grounded grid leak resistor. They have quite different voltage drops at the filament, with the second one presents a higher voltage drop (hotter bias?)
Another update!
I added a tone control and some minor changes.
There still some noise, but it only is present for the last 10% of the volume pot. It seems that the signal at the PI is been received by the preamp tube. By grounding the output of the first stage all the noise is gone. I blame the layout, which is really bad with all the cables around.
So now I will work on a proper PCB layout and just keep some spare grid stoppers just in case the noise is still there.
Something that kept me busy was the PI. I was measuring the signal at the grids of the power tubes and I can't find a way to balance the PI. If I choose the pair of resistors for a given volume setting it works, but if I change the volume the differene between signals increase.
For example, I adjusted the voltage for a 17V p-p at both grids, that resulted in the 150k and 30k resistors as in the schematic. When I increase the volume say to the max. swing that I could get, 32v p-p, the second side, after the voltage divider stays at 23v p-p. It could always be that I'm hitting the power tubes too hard. But the same occurs when I reduce the volume...
My current bias is -3.5v, so 32v is almos 10 times this value. The problem is that I kind of like the distortion when at full swing. It is really full sounding.
Well I tried the feedback idea presented by Merlin, but it had the same behaviour. I guess I need to try a fixed grid leak resistor instead of the volume, and add a cap between both so that the grid bias stays constant right?
I guess my approach is wrong since the tubes are at fixed bias, and not cathode bias as in other circuits.
I added a tone control and some minor changes.
There still some noise, but it only is present for the last 10% of the volume pot. It seems that the signal at the PI is been received by the preamp tube. By grounding the output of the first stage all the noise is gone. I blame the layout, which is really bad with all the cables around.
So now I will work on a proper PCB layout and just keep some spare grid stoppers just in case the noise is still there.
Something that kept me busy was the PI. I was measuring the signal at the grids of the power tubes and I can't find a way to balance the PI. If I choose the pair of resistors for a given volume setting it works, but if I change the volume the differene between signals increase.
For example, I adjusted the voltage for a 17V p-p at both grids, that resulted in the 150k and 30k resistors as in the schematic. When I increase the volume say to the max. swing that I could get, 32v p-p, the second side, after the voltage divider stays at 23v p-p. It could always be that I'm hitting the power tubes too hard. But the same occurs when I reduce the volume...
My current bias is -3.5v, so 32v is almos 10 times this value. The problem is that I kind of like the distortion when at full swing. It is really full sounding.
Well I tried the feedback idea presented by Merlin, but it had the same behaviour. I guess I need to try a fixed grid leak resistor instead of the volume, and add a cap between both so that the grid bias stays constant right?
I guess my approach is wrong since the tubes are at fixed bias, and not cathode bias as in other circuits.
Yes, I tested it right after my post and with another cap and resistor I could get the same output and use the whole volume pot.
The grid leak resistor is only 30k, so I guess it was a little over the top. Now the grids are both swinging at 17Vpp and this gives me 177mW at the output.
Again, I'm testing this with a crappy 4" speaker, the same that will go in the small combo, but when using a proper speaker the amp is loud enough. The tone is much better too. Much cleaner than with the 4" speaker.
The grid leak resistor is only 30k, so I guess it was a little over the top. Now the grids are both swinging at 17Vpp and this gives me 177mW at the output.
Again, I'm testing this with a crappy 4" speaker, the same that will go in the small combo, but when using a proper speaker the amp is loud enough. The tone is much better too. Much cleaner than with the 4" speaker.
Good stuff! When you do the board post a finished diagram would you? It would be nice to see the end solution.
Yes, that's right! According to general understanding, in directly heated tubes alle electrodes' voltages are referred to the filament's negative end. So the upper final tube is biased just by the voltage drop over the lower final's filament. I don't know the first four tubes' designation, nor their specs, but if we assume their filaments' voltage being half of the finals', then the lower final is biased by two times the voltage of a small signal tube, which per assumption is the same as one final's heater voltage, plus the voltage drop over that 22R resistor at the filament string's grounded end. So, for equal biasing the lower final's grid leak resistor should be referred to the upper end of the 22R.
But perhaps disbalance, and hence more even order distortion, is exacly what the OP wants?
Best regards!
After a bit of research, I did some more reflections on the heater circuit. As I've found out, heater data are 1,2 V @ 13 mA for the 1P24-B and 2.4 V @ 30 mA for the 1P29-B tubes, respectively. Thus, the difference of 4 mA has to be bypassed by resistors in parallel with each pair of 1P24-B's. The nominal value of each resistor would be 1.2V/0.004A = 300 ohms. In addition, the plate current of the finals, which assumptedly is not negligible at all, also flows through the heater chain. Hence the lower value of 200 ohms. The difference between 300 and 200 ohms corresponds to an additional current of 2 mA, i.e. 1 mA per final tube. Is that enough? Otherwise the value of the bypass resistors needs to be decreased correspondingly.
Best regards!
Best regards!
You are right!
the 200R resistor is there for the extra current.
The datasheet says 0.95 mA +- 0.45 mA for the plate and 0.1 mA for the screen.
Since I'm running the tubes between 60v and 70v, the 51k resistor should give a max. current of 1.0 mA, when the tube reaches +1v at the grid and cuts off.
With 100k this should drop to 0.55 mA.
On the other hand, the PI tubes have a higher voltage, almost 70v, and lover plate resistors, which should require more current when the grid gets positive, but I'm running them on triode mode, where I expect the plate current to drop.
At Radiomuseum there are some curves for the 1J24-B in triode mode, I tried drawing a load line and it gave me roughly 1 mA.
the 200R resistor is there for the extra current.
The datasheet says 0.95 mA +- 0.45 mA for the plate and 0.1 mA for the screen.
Since I'm running the tubes between 60v and 70v, the 51k resistor should give a max. current of 1.0 mA, when the tube reaches +1v at the grid and cuts off.
With 100k this should drop to 0.55 mA.
On the other hand, the PI tubes have a higher voltage, almost 70v, and lover plate resistors, which should require more current when the grid gets positive, but I'm running them on triode mode, where I expect the plate current to drop.
At Radiomuseum there are some curves for the 1J24-B in triode mode, I tried drawing a load line and it gave me roughly 1 mA.
Last edited:
No, I measured it with a multimeter and the current is 4.5 mA for the plate.
According to datasheet I can go up to 5.3mA +- 1.7mA at the plates (at 45v I guess) and 0.5mA at the screen. Adding everything I should expect a max. current of 37.5mA with heaters.
If I consider that the preamp tubes each require 14.4mA I could reduce the 200 resistors to something like 140 ohms.
What is weird is that the voltage drop that I'm measuring is 1.5v and 2.2v at the output tubes, where it should be 2.4v. I'm not sure why it is so low at one of the tubes.
According to datasheet I can go up to 5.3mA +- 1.7mA at the plates (at 45v I guess) and 0.5mA at the screen. Adding everything I should expect a max. current of 37.5mA with heaters.
If I consider that the preamp tubes each require 14.4mA I could reduce the 200 resistors to something like 140 ohms.
What is weird is that the voltage drop that I'm measuring is 1.5v and 2.2v at the output tubes, where it should be 2.4v. I'm not sure why it is so low at one of the tubes.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.