I have tried it on my shunt reg experiments, not exactly this schematic, but similar. The first obvious thing you notice is that the capacitance of the paralleled mosfets is higher. On account of this you lose all that good performance in the high frequency range. So you may as well go for the simplistic version. Also, you can chose much beefier mosfets.
Thanks Salas.
Though I abandoned this idea, I cannot see why it would accomplish a disaster.
Anyways a good PSU is needed whatever it is. "It is most important to realise that an amplifier is merely a modulator and controls the flow of energy from the power supply to the load. If the power supply is poor and has insufficient energy to meet the amplifier's peak demands, then the most beautifully designed amplifier will be junk."
MORGAN JONES
VALVE AMPLIFIERS 4TH EDITION page 334.
MORGAN JONES
VALVE AMPLIFIERS 4TH EDITION page 334.
I've been running my SSHV1 at pretty stressful conditions ( I thought that the more extra current the better.) It supplies the tube 20mA @180V and dumps an extra 40 mA, V-in is about 230V with a big heatsink. Its been stable for a year (red PCBs.)
So I don't think supplying an output of 65ma +20ma @370 is too much of a stretch for the SSHV2. Especially with the new IXTY-TA-TP08N50D2 depletion mosfet at Q1(which can be mounted to the off-board sink).
I think a TO-247 like IRFP450 , would sub for Q3 ?
As I mentioned a few days ago, I have completed a couple of SSHV2s on matrix board and installed them on the chassis of my RTP3. These are working almost fine at the required current and voltage.
I have one question: the heatsinks for the current regulators (Q1) get quite hot, even though they are passing only 70mA with an input of 345V and a regulated output of 300V, which I calculate should only dissipate just over 3W. The sinks on the CCS device are rated at 3.9C/W, so should be only 12 degrees or so above ambient, but are uncomfortably hot to the touch, and are definitely hotter than the ones on the shunt MOSFETs, which I reckon are dissipating 4.5W at the shunt current of 15 mA. Is this indicative of some FR instability?
Alex
I have one question: the heatsinks for the current regulators (Q1) get quite hot, even though they are passing only 70mA with an input of 345V and a regulated output of 300V, which I calculate should only dissipate just over 3W. The sinks on the CCS device are rated at 3.9C/W, so should be only 12 degrees or so above ambient, but are uncomfortably hot to the touch, and are definitely hotter than the ones on the shunt MOSFETs, which I reckon are dissipating 4.5W at the shunt current of 15 mA. Is this indicative of some FR instability?
Alex
I had another look at your construction instructions and realised I had missed the vital phrase "Keep last raw DC filter capacitor near the regulator's input"...
I have put a 0.22uF cap (what I had at hand) right at the regulator input and all seems rather more stable.
Many thanks!
Alex
Another question: is the function of the network R6/C1 as a low-pass filter, or is it an impedance correction for stability? How critical are the component values here?
I wonder whether a 1/4W resistor for R6 will have a high enough voltage specification in this position in some applications - the 0.6W MF parts I use have a rating of 250V, which doesn't seem enough for a 300V output, and I'm tempted to upgrade this to a 2W resistor with a commensurately higher voltage rating. I have no real feeling for what the maximum voltage will be across this component - I guess it will depend on the rise time of the raw DC input.
Alex
I wonder whether a 1/4W resistor for R6 will have a high enough voltage specification in this position in some applications - the 0.6W MF parts I use have a rating of 250V, which doesn't seem enough for a 300V output, and I'm tempted to upgrade this to a 2W resistor with a commensurately higher voltage rating. I have no real feeling for what the maximum voltage will be across this component - I guess it will depend on the rise time of the raw DC input.
Alex
Last edited:
Its the latter. A "snubber" or Zobel network more precisely. Voltage drop across that resistor will be insignificant because it just poses "ESR" there. That capacitor has all of it across in essence. The values are critical enough and better not be altered unless someone optimizes damping for a specific load evaluating with lab gear. In general its a little wonder how this design interfaces with so many people's stuff without a multitude of major drama stories. One thing is I don't push the spec too far on purpose, it could be impressive to have 1/3 the now Zo and 3x the now bandwidth, but it would not be generally easy to get widely applied with simple means and no deeper understanding effort. There is a point of diminishing returns subjectively also, so its a juggling act to make a useful general purpose solid state HV shunt reg with some field success. Until now it seems serving the goals rather satisfactorily is my impression. The main purpose is to bring good tonality regulation in compact space with better damping for little money and weight VS iron clad traditional passive filtering cells. A rail hum killer with added value and good sonics in other words.
Thick aluoxide pads I guess is minimum at those voltages. And even that doesnt live up to electrical norms and standard which I think say like tou have to keep some 1,5 mm distance between hot and ground. And that is hard with a thru hole metal screw. Clamps would be the most appropriate I guess.