Simplistic MosFET HV Shunt Regs

[Salas]

Thanks, Salas.
Modified schematic attached.
Any comments are welcomed.

Would it be safe to adjust the output voltage without any load?
Or is it better to use a dummy load, say 220V 100W light bulb?

I prefer a dummy. It will be hard on the shunt element without. See to use an Ohmic load that will consume about 45mA (analogous to 2x6V6 pre channels). That is 15W dissipation. Maybe 2x 15k 20W resistors in parallel. What kind of Ohmic value that bulb gives when hot?

P.S. Having 175k, set the trimmer around 200R before installation, should get you to your target when going lower. Don't start high trimmer, so the Vdif will not tax the components until you fix it.

 

[analog_sa]

Did you do the Jfet Vref mod BTW?

No, just learned from Goshu_G's circuit above about it. Busy with another project at the moment but i can see more of your regs in my system soon

 

[Salas]

If under 300V read some pages before, its with MJE350 and tested. Joshua is going to use a 400V smaller dissipation driver transistor and simulated with some differences in values, so its to be seen if it will work OK or it will need mods.

 

[Joshua_G]

I prefer a dummy. It will be hard on the shunt element without. See to use an Ohmic load that will consume about 45mA (analogous to 2x6V6 pre channels). That is 15W dissipation. Maybe 2x 15k 20W resistors in parallel. What kind of Ohmic value that bulb gives when hot?

P.S. Having 175k, set the trimmer around 200R before installation, should get you to your target when going lower. Don't start high trimmer, so the Vdif will not tax the components until you fix it.

100W light bulb is way too much, 15W is closer to 45mA current draw. I'd rather use resistors. Anyhow 340V supply may well burn 220V light bulbs.

 

[Joshua_G]

If under 300V read some pages before, its with MJE350 and tested. Joshua is going to use a 400V smaller dissipation driver transistor and simulated with some differences in values, so its to be seen if it will work OK or it will need mods.

Indeed, there is nothing to learn from my suggested version of the circuit, since I didn't build it yet.

 

[Salas]

It will probably do well. But we wait for your feedback.

 

[Joshua_G]

Feedback will be posted once the regulator and the amp will work. Right now I'm ordering parts.

 

[quanghao]

THe simplistic design by SALAS

Hi Salas, Hi all
With the acceptance from Salas, I have made some testing boards for me and gift for interested members. Please log in your name and address for sending. Thank you Salas, thank all

1. Nicola =1
2.Jean-Charles=1
3.DUC985=1
4. Mathias=1
5. Joshua_G=1

Thank you very much!

 

[Joshua_G]

Salas and all,
Which 1u 400V capacitors do you use, or recommend?
I believe that the selection of this capacitor should make a difference, not less than the capacitors used in the rectifier's filter.

 

[Salas]

Wima, Rifa, Vishay, MKP if possible.

 

[Joshua_G]

Wima, Rifa, Vishay, MKP if possible.

Not any one of those?:
AEROVOX AFPS 1uF 400VDC +/-10% MKP
CORNELL DUBILIER 935C 1uF 400V
AUDIOPHILER 1uF 400VDC +/-3% MKP

 

[Salas]

Don't know, haven't used. They look like good components. Of course they will work. Now on how much they can change something subjectively, I don't have a clue.

 

[Joshua_G]

Thanks.

 

[Joshua_G]

Salas,
Any recommendations for the type, model and make of the various resistors used?

 

[Jarno]

1% metalfilm? MOX for the higher wattage stuff?

 

[Salas]

Carbon for the gate stoppers and base stopper, if easy to get, it will still work with metal film, keep legs short and direct to the gates & base, film for the rest.

 

[Joshua_G]

Thanks, Salas,
You mean Carbon Composite, or Carbon Film, or both?

 

[Salas]

Carbon film is better for stability in the long run.

 

[Joshua_G]

While parts are being ordered, I thought of a modification, separating L and R channels supply, for perhaps better sound image.
Any comments?

 

[Salas]

It negates the low output impedance goal, doesn't it? I would find the inclusion of remote sensing nodes more preferable to test VS just a twisted pair. (Post 831).