Successful TPU-ECD replication !

[BEP]

@All

FYI:

Tubes don't always require higher voltages to switch on. All they require is a cathode or heater with extra particles (heater is on or made to work cold) and a difference in potential between the cathode and plate. The 12AT7 and its cousins will work well enough with 12VDC.

I'm not saying high voltage isn't required to make a TPU. I don't know that at this point.

[Jeff B]

Hello all,

@All

You all have to understand 1 thing: with SS - MOSFETs - its almost NOT possible to have a good working TPU!! Yes, I had 3 runaways with my SS driven TPUs but I was working like a idiot to achieve this.

With TUBES its not sooo a hard work. Of course you all have to understand how a tube works, how to connect it....

NO, dont ask me how to connect it. Im a newbi with tubes but a lot of you missed how the output signal of a TPU looks like: SINE WAVES AND ON TOP OF THIS SINES ARE KICKS.

Now imagine the plate voltage of a tube is 200VDC!!!! WOOOOW!!!! And now imagine the heating voltage of this tube is 12,6VAC. Again: WOOOOW!!!!!

END RESULT WHEN THIS 2 VOLTAGES COMES TOGETHER: SINE WAVES WITH KICKS ON TOP.

Otto

Hmmm....
Interesting statement on the tube drive.

Yes you have 200V on the plate, but this still has to be connected to the TPU as you mentioned, and (I imagine) that only a portion of this will be coupled to the TPU as an AC drive signal (for a linear output) - not that I'm knocking the amplitude you could achieve compared to the 12V PSU.
Also - most valves are 6.3Vac on the heater. Which ones are you using ?
And what specific relevance is the heater voltage - which should have no bearing on outcome.
Another thing is that the 2 voltages won't give you "Kicks" on the output - it is an amplifier, and so should only give you a larger version of what you feed in on the gate.

Coupling:
This perhaps isn't quite as easy or clear as some might think at first. (though I haven't tried it, so I say this with some reservation).
You have to have a method to transfer the AC portion to the coil, without the tube becoming part of the TPU and upsetting its tuning/balance.
By this I mean that with the FETs, when they switch on they are just putting a kick of power in, then when they switch off, they are effectively Disconnected from the TPU and play no part in its running or self oscillation (ringing).

With a tube, you have no way of seperating it from the TPU.
When it's injecting a signal into the TPU - great.
When (IF) it isn't, it's still connected to the TPU.
Will it disturb the tuning or Dampen the TPUs operation ?

Did you plan on having the tubes constantly injecting signal into the TPU, or just use them in burst mode ?

Obviously the easiest way is to use a series resistor from HT supply to plate, and coupling to the TPU through a capacitor.
Or maybe you try could winding some kind of matching transformer that would then connect to the TPU drive coils.
I imaging a small matching tfmr with a series limiting resistor on the HT side.
(You need the resistor to limit tube current, and give an impedance to create the voltage swing across.)

If you find the valves' basic noise is too much, try a split heater supply (6.3V centre tapped tfmr, with CT to ground).
But I doubt this will be a significant issue.

If you wanted to use it in burst mode, perhaps you could try a (biased) diode switching array to gate the signal.
Sounds like it's starting to get more & more komplickated all the time.

I'm probably telling you things you already know.
In fact I'm sure I am....

Jeff.

PS: I have an expensive book on valve amps, so if you have ideas, or wanted something clarified, I could have another read of it for you.
      (I haven't got my multifunction printer scanning to my Mac yet via my airport, but I might be able to do it some other way if I needed to.)

[Jeff B]

Just had another thought -

If you used a different valve - like a pentode for example, (which has more than one gate, though I'm a bit rusty), you could inject more than one frequency, and have the valve mix them for you.
With the smoothing qualities of a valve, it should also eliminate the really harsh spikes.

Jeff.

[otto]

Hello all,

@Jeff

Im using a ECC82 = 12AU7 tubes.Double triodes. 3 of them for 3 oscillators. Heating voltage is 12,6VAC.

Of course, youre right. Output should be sine waves. Yes, the tubes will be a part of the TPU.

The point is that I now "convert" my kicks into sine waves directly on my bulb. The input signals are square waves (kicks) and at my bulb I have connected a oridinary cap that converts my signals. I have arround 200VAC on my bulb. OK, input is 12V or 24VDC from the power supply and the current is the same 4-5A. And I saw that I can mix my square waves in the coils and have a fine output on my bulb.
With a voltage of say 250V from the tube it would be interesting to see whats going on in my coils. Yes, I know that the plate voltage is only in miliamperes but there is a heater with his current. So, when I connect the plate VOLTAGE and the heater CURRENT.....so they work together....50Hz + a high frequency...I dont know what will happen. For sure I know that there will be a vibration because yesterday I feeded a unknown  AC voltage (12V??) 50Hz to my pulsed coils and I had a vibration.

Everything Im wrighting today are guessings (I hate to guess!!) or to say it in another way, that are my feelings!!! Now Im the PC hero!!! 

Otto

[JESTES]

I am new here. Is the well-read one several documents on these topics and me gustaria that you were recommending to me some circuit to begin. I know one that otto and roberto designed (TPU_ECD_V1_0.pdf):

Do you recommend to me to do this setup? Will I be able to see in my oscilloscope the beginning of the conversion really?

Do you recommend to me better other one?

That I must use, mosfet or tubes?

Best Regars,
J.