At only 600V and 50uF of capacitance after the regulator, the chances of cathode stripping are almost certainly nil.
Eh!
The cap across car ignition points is there to create a CL circuit that deliberately oscillates to maintain a LONGER spark duration in the cylinder.
It certainly does not prevent or reduce the spark across the points. It probably makes the sparking worse. Look at the burning of the points. Look at the metal migration across the points. Look at the material (Tungsten/Wolfram) they used to make the points (nearly) survive 20000miles before they had become so bad they had to be replaced.
This was one of the main problems that had to be overcome to enable passing that 50000mile test that the US invented. It virtually demanded electronic ignition with SS switching to avoid points degradation.
close rated T (fast blow) fuses after the smoothing caps will blow very quickly in severe abuse conditions. They should be selected to just survive normal operating conditions. Nuisance blowing annually can probably be tolerated.
But are 600Vdc fuses available? 250Vac fuses use the zero volts crossover to extinguish the arc after rupturing.
But are 600Vdc fuses available? 250Vac fuses use the zero volts crossover to extinguish the arc after rupturing.
Last edited:
Fuses are hard to beat, but need to be rated for the voltage and fault current - particularly on DC. At 600V it's going rather bigger than a mains one. I think the ones used in modern DVMs are 600V rated. I have one directly on the PT secondary to protect against rectifier failure.
Yes a fuse is a piece of wire, but a very thin one. People worry that this means that it is a high temperature coeficient (as it heats and cools) low value resistor. Relevant if it's in series with a speaker, but I wouldn't think it makes any difference in an HT supply... any more than the mains fuse.
One thing I forgot to mention, I use a series chain of 4 x 25R Inrush current limiters between the rectifier and the first reservoir cap. This limits my cold switch on surge to less than 4A - gentler on the PT, relay, rectifier and caps. For 600V you could add a few more. It also limits the surge on the fuse, allowing you to use a lower rating.
Yes a fuse is a piece of wire, but a very thin one. People worry that this means that it is a high temperature coeficient (as it heats and cools) low value resistor. Relevant if it's in series with a speaker, but I wouldn't think it makes any difference in an HT supply... any more than the mains fuse.
One thing I forgot to mention, I use a series chain of 4 x 25R Inrush current limiters between the rectifier and the first reservoir cap. This limits my cold switch on surge to less than 4A - gentler on the PT, relay, rectifier and caps. For 600V you could add a few more. It also limits the surge on the fuse, allowing you to use a lower rating.
Even simpler: Try a thermistor in series with the B+ These if sized correctly will act like a 200R resistor at turn on and then over a period of a few seconds drop to a value of about 2R or 1R. They look like an oversized ceramic disk cap and cost about $2. They come with all kinds of different thermal time constants and resistance values.
The thermistor will slow the raise of the B+ voltage but of course is not as good as microprocessor controlled relays. And yes, I've seen that done in a amp. The uP monitored the voltages and sequenced the power up process by closing relays then stopped it's clock so as not to generate any noise.
hidnplayr, what is the max B+ voltage for which your
circuit will work as shown? Can I use it for 420V B+,200ma?
Any alternatives to the BUZ91A?
What is ramp up time with the shown components?
Thanks in advance
Prasad
The fashion to delay B+ in so called High-end amps come from fields like broadcast transmitters and radar stations. In audio amps it does nothing except creating abruptly control grid currents of already ready to conduct tubes suddenly charging coupling capacitors through them.,
So we are trading one problem for another? The problem that I've been trying to deal with is the power supply offload voltage going to 1.41 x transformer voltage while the filaments are warming up. If the caps can handle the voltage, should I not worry about it and just fire up mains power? I'm using CT SS FW rectification with 360V secondary transformer voltage and 500V rated caps in a PP EL34 triode design (Poinz' EL34 machine). The electrolytics are Panasonic TS-UP rated at 500V with 550V surge rating (whatever that means?) My first cap is a 630V film.
They are trading a problem that exist in another gear for problems in their own gear.
I would rather build a very good solid state regulator (or at least a good filter with a source follower) to stiffen B+ and bring it up slowly, than abruptly switched on unregulated high voltage when tubes have already hot cathodes.
Here it is again, what I did to bring both B+ and filament voltages up slowly:
Do you mean a voltage drop on MOSFET?
It does not need to be 75V, it may be as small as max ripples under the load + DC source sag under the load + 5V for MOSFET
I was looking at the B+ regulator and there is a 73VDC drop shown. I was wondering how the circuit is functioning. IF 5V across Q1 and the rest across 100R that would make load draw at 680mA. I simply don't know how it works and was trying to figure it out.
You may see 2 regulators powered from the single source in series.
A first one, is a shunt-regulator made of Q3, D3, R3, C3. Qr has big enough capacitance not only for filter purpose, but for purpose of delay. Time to charge it depends on a current that tubes draw through it. When tubes are cold rise of a voltage is very slow. But as soon as tubes are hot and start drawing the current they charge C3 Faster, until the voltage stabilizes on the level defined by D3 (plus Vbe of Q3).
The second regulator, on top of the picture, uses that bias shunt regulator for a reference voltage, so it's output voltage goes up according to increase of a voltage on C3 (look at the voltage divider R5R6R7).
Plate voltage goes up slowly, and depends on current that tubes draw. That ensures delay, that depends on tubes that consume the current, and brings up voltages slowly as soon as they are ready to conduct. Since coupling capacitors between stages with respective resistors have much shorter time constants, no significant current through grids goes when coupling caps are being charges. Everything comes up gently and intelligently.
This SS regulator is superior to kenotrone with CLC filter, since it's output resistance is much lower on wide frequency range, no resonant peaks on impedance on lower frequencies, like with chokes. Also, steepness of grow of a voltage does not depend on a kenotrone, it is ore intelligent: it depends on tubes themselves that are powered from it, and output voltages (both B+ and bias) are regulated.
However, I would rather use it for bias and screen grid power supply, while plate voltage may be still unregulated.
You may make it safer on a bench, or in a prototype, shifting current limiting resistor R4 to the left, between Zener and source. Such a way in case of an accidental shortage maximal current will be limited like 8V divided by value of the resistor.
Speaking of 5V for MOSFET, it is minimum, after sag and ripples, with lowest possible voltage in the outlet. If you think of boundary conditions, you can get reliable well working regulator, in normal conditions, between boundaries (like lightest load and highest voltage in outlet, VS lowest voltage in outlet and heaviest load)
Last edited:
Hi,
I am looking for a replacement in-line circuit for NTC resistor/thermistor. The voltage used is approx. 520V for the +B on the plates of KT88 .
Original a unknown thermistor was used in line and before the capasitor. Can I use your circuit and what is the approx. delay( have not read the whole thread).
Regards
Kim
The time constant for the circuit to reach is maximum can be calculated as 5*R*C Thus in the given example 5*75 000*(0,000 010+0,000 047) should be approx 20 seconds.
It should work, but i give you no guarantee whatsoever 😉
Thanks - hidnplayr ; I have burned my fingers more than ones - so one time more or less does not matter, and of course you can't give any guarantee!
Have you tried the circuit up to 500V? I will try to find a fet for 800V and make the circuit on a woodboard. The amp draws approx. 250mA - I only need the circuit for soft start, and I think the circuit is perfect for that.
May I send and show the circuit to other hi-fi people?
Rgds.
Kim
I have tested the circuit at 510V, make sure your mosfet can handle the dissipation on startup, a large heatsink will be required.
Sure, do with it whatever you like.
- Status
- Not open for further replies.