In the case of the attached triode graph, you have to use an error amplifier or large amounts of cathode degeneration to get a CCS characteristic. Say you want it for 100mA, you can go down to 75V (the lowest voltage possible at Vg=0V as indicated by the curve marked as such). If plate voltage rises with grid voltage constant (i.e., remaining at 0V), plate current increases according to plate resistance (about 368 ohms, is that an EL509/519 in triode mode?). This needs to be canceled with a change in grid voltage, or more typically, a change in cathode voltage, keeping grid constant. You could put a CCS below the cathode then, using a transistor, and it only has to be rated for the maximum (er, minimum) grid voltage - maybe 300V (cathode rising means grid falling), if the tube is rated for so much bias that is. (Doubtful, as no screen grid this size goes to 3kV, nor any grid more than 100 or 200V negative. I would suspect 900V absolute maximum.)
If you use a transmitter tube with sufficient perveance that it does not require grid current at low plate voltages, while having enough gain that it doesn't require immense bias voltages at high plate voltage, this will work much better.
Or use a pentode.
Tim
If you use a transmitter tube with sufficient perveance that it does not require grid current at low plate voltages, while having enough gain that it doesn't require immense bias voltages at high plate voltage, this will work much better.
Or use a pentode.
Tim
Umm, it's not a CCS, it's a voltage (or current) controlled CS, otherwise how can I get audio? 
I'm looking at closer to 200 mA, and I find up to 500 V drop in the tube acceptable. Now, say with 180 mA as the bias, the audio modulation needed will not be more than 20 mA peak to peak. Essentially, I only care about linearity in that small range around the operating point. Does this not improve the situation?
The curves are for a 300 W triode. I thought triodes were the most linear devices.
I'm looking at closer to 200 mA, and I find up to 500 V drop in the tube acceptable. Now, say with 180 mA as the bias, the audio modulation needed will not be more than 20 mA peak to peak. Essentially, I only care about linearity in that small range around the operating point. Does this not improve the situation?
I don't follow that at all. If grid voltage is constant, what will make plate voltage rise? Everything is constant, with the only thing changing being the grid voltage, in a grid driven setup. 😕
The curves are for a 300 W triode. I thought triodes were the most linear devices.
Okay. Set up a triode with cathode grounded. Bias it. Add a voltage source to the grid, an AC current or voltage meter to the plate, and an AC voltage source. AC voltage source in the plate? Why? To demonstrate the output impedance of the plate. It needs to be high for a CCS output, if it is low, it will have a constant voltage characteristic which is probably no good here. That the point, if the load tries decreasing the voltage across its terminals, what does the tube do?
I still don't see why you want such a small current variation, and more than likely, you'll find yourself reaching for the volume control... but whatever... Current can be easily throttled by varying grid voltage, depending on what's under the cathode.
The best suited tube is still a pentode, or if you are certain of the voltage range, a transistor might work...
Tim
I still don't see why you want such a small current variation, and more than likely, you'll find yourself reaching for the volume control... but whatever... Current can be easily throttled by varying grid voltage, depending on what's under the cathode.
The best suited tube is still a pentode, or if you are certain of the voltage range, a transistor might work...
Tim
Most probably the MHCD sustained discharge configuration behaves resistively rather than constant voltage. Now, since I need output POWER to be proportional to the input sound voltage for the driver, should plate voltage also vary or not with varying current?
In any case, can you recommend specific tubes? What tubes would sound best, up to a couple hundred bucks/tube (preferably half that)?
That's easy. The plasma is not linear unless the perturbations are very small compared to the bias (Hill's patent says temperature fluctuations do not approach 1% of the operating point temperature, so 10% swing ability of input power should be sufficient). As for if it's loud enough, consider a 100 W per channel amplifier feeding a dynamic driver. Most of these, in their speaker enclosure, are several percent efficient, which means several watts of acoustic power is generated. Now, I have 300-400 W average plasma power per channel. Small variations are sufficient to generate comparable acoustic power for me to mach to the low frequency amplifier/driver.
In any case, can you recommend specific tubes? What tubes would sound best, up to a couple hundred bucks/tube (preferably half that)?
That's easy. The plasma is not linear unless the perturbations are very small compared to the bias (Hill's patent says temperature fluctuations do not approach 1% of the operating point temperature, so 10% swing ability of input power should be sufficient). As for if it's loud enough, consider a 100 W per channel amplifier feeding a dynamic driver. Most of these, in their speaker enclosure, are several percent efficient, which means several watts of acoustic power is generated. Now, I have 300-400 W average plasma power per channel. Small variations are sufficient to generate comparable acoustic power for me to mach to the low frequency amplifier/driver.
Also, someone from rec.audio.amateur.homebrew told me that I could use a 304TL or 304TH, as it has very high perveance and won't be overvoltaged if there's a short from the HV to the plate. I don't think I need that much wattage, so what about the half-sized version, 152TL? Here's the datasheet with constant current characteristics:
http://www.mif.pg.gda.pl/homepages/frank/sheets/090/1/152TL.pdf
http://www.mif.pg.gda.pl/homepages/frank/sheets/090/1/152TL.pdf
Well.... if you consider 600V saturation at 0Vg, 250mA perveance...
Why do you insist on a triode?
Tim
Why do you insist on a triode?
Tim
Well now I've got the output devices.
I've also got the air system sockets for these, SK-630, which are actually bigger than the tubes themselves. According to the datasheet, The screen grid connection is bypassed by a 1100 pF capacitor to ground (socket case). It looks like the capacitor is integrated in such a way into the socket that I will not be able to disconnect it (do I need to?).
An externally hosted image should be here but it was not working when we last tested it.
I've also got the air system sockets for these, SK-630, which are actually bigger than the tubes themselves. According to the datasheet, The screen grid connection is bypassed by a 1100 pF capacitor to ground (socket case). It looks like the capacitor is integrated in such a way into the socket that I will not be able to disconnect it (do I need to?).
The SK-630 sockets, amusingly larger than the tubes themselves:
The sockets' inner tongues touching the screen contact ring on the tube connect to a ring around the inside top of the socket separated from the case by a dielectric, forming a screen bypass capcitor. When I was cleaning these silver-plated sockets (they were quite tarnished when I got them) using the hot-baking-soda-solution-in-aluminum method, I broke the capacitor seal on one and baking soda water got inside the capacitor. Now it reads some resistance on the DMM... Maybe if I soak it for a while to clean out the baking soda and when it dries it will be OK. That's assuming I don't need to disable the capacitors, which I'm not sure of at this point.
I need to find a way to build chimneys that seal the gap between the socket and anode heatsink so air flows between the fins.
An externally hosted image should be here but it was not working when we last tested it.
The sockets' inner tongues touching the screen contact ring on the tube connect to a ring around the inside top of the socket separated from the case by a dielectric, forming a screen bypass capcitor. When I was cleaning these silver-plated sockets (they were quite tarnished when I got them) using the hot-baking-soda-solution-in-aluminum method, I broke the capacitor seal on one and baking soda water got inside the capacitor. Now it reads some resistance on the DMM... Maybe if I soak it for a while to clean out the baking soda and when it dries it will be OK. That's assuming I don't need to disable the capacitors, which I'm not sure of at this point.
I need to find a way to build chimneys that seal the gap between the socket and anode heatsink so air flows between the fins.
- Status
- Not open for further replies.