hv psu load drop. any ideas

[yiancar]

hello dear eev comunity. i have build 2 of the following psus http://www.cavalliaudio.com/diy/bijou/main.php?page=psschematic and in one of them i experience an unusual load drop. target voltage is 250v which is achived without load. when i load them, one of the psus gets stuck at 244v max even if i increase the trim pot. The other psu runs fine. i have swapped transformers and rectifiers without any results. Any ideas?

[SeanB]

I would change R1 and possibly Q1. Likely the resistor has gone high resistance. Measure the voltage across C3, if it is not 6V2 and is low then either R6, C3 or Z1 is faulty. If it is high then definitely Z1 is faulty. Also check C6 is not open circuit.

[yiancar]

I would change R1 and possibly Q1. Likely the resistor has gone high resistance. Measure the voltage across C3, if it is not 6V2 and is low then either R6, C3 or Z1 is faulty. If it is high then definitely Z1 is faulty. Also check C6 is not open circuit.

r1 is correct, z1 is 6.275 is that ok or is it considered high?

[SeanB]

That is fine, within 5% of the voltage which is fine. What is the voltage across the C4, C5 and C6 when it is delivering this low voltage. That will tell if it is in the rectifier and RC filtering or in the regulator. There should be around 300-350V across C7, and over 260V across C4, basically a drop of around 20V across each resistor.

[yiancar]

c6 ~294 c5 ~275 c4 ~256

[SeanB]

C6 is either open circuit or very low in value is then the best bet. Check by changing it with another, or the one from the good channel. The rectifier does not have a drop of 60V at the current drawn. Typically it would be under 30V, and the HT should be around 330V. Compare with the good rail and you will see.

[yiancar]

chenged c6. worked for about a minute then the same. this is confusing.... also measured its capacitance, 47.6uf

[SeanB]

Then either a track to one of the anodes is open circuit or the valve is faulty. Check for continuity from the pins  to the transformer, and that the windings are not open circuit, or the socket had a broken or damaged pin socket for the one anode.

[Paul Price]

If there is voltage tolerance of the supply is rated to +- 5%, then 244V output with a load is in spec.
Otherwise the 6.1V zener diode on the source of the bottom MOSFET must be  the problem.
I would toss the 6V4 tube into the bit bin and replace it with two 1N4009 rectifiers each bypassed with a .001/1KV cap and save a lot of global warming.

[Andy Watson]

I would toss the 6V4 tube into the bit bin ...

+1
The fact that it is there in the first place suggests that this power supply and amp is more about looks than functionality. I'd toss the whole lot in the bin.

[yiancar]

I would toss the 6V4 tube into the bit bin ...

+1
The fact that it is there in the first place suggests that this power supply and amp is more about looks than functionality. I'd toss the whole lot in the bin.

it is an amp psu . the 6v4 tube is there so there will be no need for a B+ delay circuit.....

[Andy Watson]

I would toss the 6V4 tube into the bit bin ...

+1
The fact that it is there in the first place suggests that this power supply and amp is more about looks than functionality. I'd toss the whole lot in the bin.

it is an amp psu . the 6v4 tube is there so there will be no need for a B+ delay circuit.....

Care to elaborate? At the moment it sounds very much like regurgitated HiFi lore. There is already a (silicon) regulated control element. If you were serious about needing a B+ delay would it not be better to design-in a well defined delay rather that rely on the thermal characteristics of the rectifier valve?

Edit: I'd still toss the whole lot in the bin, although I'm willing to entertain any half-decent reason that might redeem this design.

[yiancar]

I would toss the 6V4 tube into the bit bin ...

+1
The fact that it is there in the first place suggests that this power supply and amp is more about looks than functionality. I'd toss the whole lot in the bin.

it is an amp psu . the 6v4 tube is there so there will be no need for a B+ delay circuit.....

Care to elaborate? At the moment it sounds very much like regurgitated HiFi lore. There is already a (silicon) regulated control element. If you were serious about needing a B+ delay would it not be better to design-in a well defined delay rather that rely on the thermal characteristics of the rectifier valve?

Edit: I'd still toss the whole lot in the bin, although I'm willing to entertain any half-decent reason that might redeem this design.

to be truthful i was more interested in the amp part rather than the power supply, it just came as a kit.I know that a silicon rectifier would be more efficient but being an amateur i didnt want to risk it as i think i need t also tweak the rc ladder. Can you elaborate on that (how to use a silicon rectifier)?

[SeanB]

Take tube out and place in box again. Leave socket in the chassis as a convenient mount. Take two 1N4007 diodes or 2 1N5407 diodes and solder into the circuit as  follows- anodes to pins 1 and 7, join the 2 cathodes together and connect to pin 3. Leave the heater pins alone. Then apply power and you will find the regulator will be able to supply the required 250v easily.

[Andy Watson]

Can you elaborate on that (how to use a silicon rectifier)?

SeanB seems to have covered it. As suggested further up, I would add a capacitor across each diode to prevent them turning off too abruptly.

I assume the whole B+ delay thing is to reduce the possibility of cathode stripping. Millions of radios and TVs have been manufactured with little regard to this problem - so it's probably not a problem! Ironically, the valves most affected by cathode stripping are usually power devices - like the rectifier.