DCAudioDIY.com

Rock4016,
I'd like to supply the voltage from a 150-Volt VR tube through a CCS that is pretty stable with 20V across it, hence the 130V plate voltage. To get the plate current up, I chose a forward-biased silicon carbide diode (see Dave Berning's response) to push the grid voltage down to 0.8 - 0.9 V. This should give me an operating current of 30 mA at 130V, so the tube dissipates almost 4W, which is half the maximum dissipation. That's conservative.

I think I did that because I'd like to use VR tubes, and the 0D3 has a max current of 40mA. I'm concerned that if the 6C45P dies, all the current will be sinked by the 0D3 - like 20mA for the VR and 30mA for the tube. My memory is foggy, and this may be why I don't have a power-supply schematic posted: maybe I dropped the triode current to 30mA and planned to drop the VR current to 10mA in hopes of preventing a blowout. Anyone want to help with my VR situation?

Looking back at the plate curve, 130V Ep and -0.8V Eg should produce 40 mA through the tube, so maybe I wrote it down wrong, or didn't re-evaluate with the lower current. The CCS is adjustable, so if the current is dropped, the reduced plate voltage is absorbed by the CCS (which gets hotter).

_________________
- Guy

David B,

That is not something I considered. My hope was that the very-small input signal wouldn't vary the plate current enough to cause much change in the cathode voltage, but any small variation on the cathode could be on the order of the input signal in this case. Let me look and maybe experiment with this and see what I can measure. I have one channel of the first stage breadboarded, so I should try it - maybe feeding it with a VR tube that is in turn fed with a CCS. I hope to keep the Magic Smoke inside where it belongs.

_________________
- Guy

Okay: Hoots & Grins, here is what I was thinking about the pre-amp power supply:



Run everything from a combination of solid-state Constant-Current Sources (CCSs) and cold-discharge voltage-regulator (VR) tubes. The tubes look cool, and are apparently quiet if driven by a CCS. The High-Voltage (HV) shunt regulator is an 0A3 in series with an 0D3 to regulator to 225V. The two first-stage regulators are a single 0D3 per channel. I'm undecided about the two CCSs.

My problem is I have no experience with VRs. I don't know if I should split up VRs for the later stages like I did for the first stage because of the current draw: If I run the VRs at about 20mA, I can probably loose up to two of the four modules (two second stages, two drivers) without overloading the 40mA VRs. That's not true in the first stage - if I loose a triode, the VR will overload with 55mA of current. I need enough voltage to strike the series VR string, but I'm worried about power dissipation of the HV CCS (maybe one of Charlie's C5Ss). If it drops 105V, it has to dissipate over 13W.

Just for information, here is a block diagram of my heater arrangement. No cascodes or wickedly-powered cathode followers so I don't have to worry about multiple heater supplies, though I should bias the whole thing up to +40V to turn off the heater-cathode diode. I ended up with an odd number of tubes, so instead of loading one side of a split 12.6V supply with the driver heater, I'll use a 12SN7 across both 6.3V supplies instead of a 6SN7 across one leg. Looks like I need an Amp of heater current - can't do a series string because of all the different currents.



More Comments?

_________________
- Guy

Guy,

You could run a series string but I doubt that you'd want to. All you have do is shunt the lower current filaments with resistors so that they end up using the same current as the higher current filaments. I wouldn't bother however. For low voltage supplies I tried fancy things like "super regs" and ended up liking the sound of well filtered 317s best. YMMV!

charlie
charliewphelps@gmail.com

The diode is non-linear no matter what the current. See for example the following link.

http://pveducation.org/pvcdrom/pn-junct ... e-equation

The idea I was trying to get across is that if you can swamp the modulating signal current from the tube with a fixed constant current through the diode, you can make the diode provide a quiet constant voltage source, which is what you want to bias the tube.

Guy,

Note that these gas reference tubes (0A3 etc. ) are very noisy. Their output should be well filtered by an RC network. Do not connect a capacitor directly to the tube though, as it will motorboat, and perhaps destroy itself. If you want a regulated tube supply, I would look on the web for a series-pass tube voltage regulator. Then I would decouple that from your sensitive preamp with an RC network so that you don't introduce the sound of the regulator into your preamp.

See fig. 4 in following link as an example. You can substitute a gas reference in place of the zener reference if you like. As to gas reference diodes, the 5651 is a much quieter tube and is recommended over the 0A3 family. See also the 2nd link for tube data and a better regulator.

http://www.blackdahlia.com/html/tip_77.html
http://frank.pocnet.net/sheets/049/5/5651.pdf

David

I forgot to mention that any of these series-pass regulators require that the heater of the series-pass tube is elevated in dc to a level that exceeds the safe heater-to-cathode breakdown voltage if the heater is at ground potential. Therefore, you would need an extra floating heater supply to heat that one tube. For this reason, some folks use a shunt regulator rather than a series-pass solution. But that is practical only for a circuit that draws very little current. Since you are planning such high load current, that is not a very good solution, since the shunt current should be nearly equal to the load current. Unless you are trying to make a room heater.

I'm using gas tubes for the 450v regulated supply in my phono stage. With and LCLC filter after the regulators, I've noticed no noise problems. Gas tubes can handle more than rated current for a fairly long time, unless you're planning to leave the phono stage running 24/7/365, I'm sure you'd notice the failed 6C45P long before any damage to the gas tube...

Roscoe

_________________
I can explain it to you, but I can’t understand it for you.

Yes, Roscoe,

The real important thing is to use a filter after the regulator. More important than having a regulator in the first place, at least when you are not trying to do a dc-coupled amp.

David

I see a major problem with your first-stage PS design. No two gas tubes are identical. Once one of them strikes, the other one never will. Use separate CCSs for each channel, or just go with separate 1.8k/5W dropping resistors. Either way, you're going to need an LC (minimum) filter after the gas tubes, which will drop some voltage on you anyway. You may want to think about a stacked pair of 90V tubes to make up the drop in the dcr of the L stages(s)...

Roscoe

_________________
I can explain it to you, but I can’t understand it for you.