Simplistic MosFET HV Shunt Regs

[vgeorge]

Depending on voltage output a nos az1 can supply about 100 mA at 300 V (philips datasheet)

So only two channels with one shunt possible.

 

[10y]

So only two channels with one shunt possible.

Yes VGeorge,

at this point I will prefer to try a single dc raw with tx-rectifier-c-l-c line and after two shunt regs . I will see if my old az1 mesh will be sufficient strong on emission.

Cheers,
Anto

 

[albertli]

Reminder to all forum members

I placed my order to 'quanghao' for two boards and paid him through paypal over 2 weeks. I sent him couple email in between and tried to know when probably I'd have my two boards, but with no response. Please be warned not to send him any money before he clears this transaction.

 

[roger57]

He seems a bit slow in responding when I sent him emails. 2 weeks is a little soon, so be patient, our mail system, specifically Canada Customs, can be a little slow!
Let us know how the boards and operation turn out.

Gary

 

[kacernator]

Shunt problem

Hi,

yesterday I had problem with Salas shunt reg. When I powered off the amp, something blew and smoked from regulator. It was working for two months without any problem so far.
This is my second experience with Salas reg that something suddenly blew off.
Version is "simpler simplistic Nov 09". Vin is 360V Vout is 290V.

It sounds superb, but i am now concerned with its reliability. I dont have gate protection zeners installed.
So I could install them and probably raise the gate stoppers to 330 Ohm?

Or does anyone know about other very good shunt reg design?

Thanks for advice.

 

[Salas]

Use the small Zeners on the gates anyway. Is there the MJE350 working near its Collector Emitter Vmax spec in your application?

 

[kacernator]

Use the small Zeners on the gates anyway. Is there the MJE350 working near its Collector Emitter Vmax spec in your application?

Yes, the max spec of MJE350 is 300V and I am running it at 290V.

 

[Salas]

Don't know, that prototype of Nov2009 is still in service 315Vin,285Vout,40mA. Never blew. Has the gate Zeners non the less. HV, semis, and heat, are always potentially troublesome stuff on the other hand. Anyway, since higher Vmax PNP BJTs are hard to source, I will show you a bit later a schematic of how to cascode MJE350s for doubling their voltage spec and halving their dissipation. In general, keep it coolest temp possible and lower the Vin-Vo if you can. Sink the Jfet somehow with a little metal TO-92 winged cap or some improvised thing if its receiving some over 40C ambient heat from the sinks, its sensitive also.

 

[kacernator]

I cant find any substitutes for MJE350 for higher voltage in my country. What can I do?

 

[Salas]

You can cascode them, with this mod. That is one potential issue solving with this mod. It may well had been a transient that killed it or something.

 

[massimo]

MJE340/350 aren't very robust devices IMO. They were used in the Mission Cyrus One and Two amps as drivers and you don't have idea how many blew!

 

[Salas]

Well maybe, but you never know the main culprit of a failure in every application and diy build unless there is full disaster data collected. Cascoding 2 MJEs will certainly take off much burden when working near single MJE's Vmax.

 

[Knorfski]

Dear Salas,

The regulator looks beautiful. What is the maximum capacitance after the regulator? I am thinking of integrating the regulator in my pre-amp. The pre-amp will perform better with some lower voltage. The setup is C-L-C-L-C, I will change it then C-L-REG-L-C.

 

[Joshua_G]

It seems to me that there is no need for LC, nor any other filtering, following the Salas shunt regulator, since it's quiet enough as it is, even for a phono stage. CLC preceding the regulator may be a good idea.

 

[Salas]

L right after the reg is dangerous for oscillating.

 

[kacernator]

MJE350 substitutes

Hi,

looks like I have found substitutes for MJE350 in my country.
I have available KSP94, specified for 400V max.

Can I use it? Does it need some circuitry changes?

Thanks.

 

[Salas]

Its a much faster unit (<<Cob than MJE350) and TO-92. That means sink it somehow, and lower the output resistor in the Zobel from 3R3 to 2R7. That is the way I would try it, but I haven't had it ever to say for sure its gonna be issues free. Also more linear in low Ic and Hfe, possibly better if successful. Are you using a PCB?

 

[kacernator]

No PCB, I do it P2P on board.

Hm, so it is safer to stay with cascoded MJE350.

Ok, thanks.

 

[hesener]

Hi Salas,
maybe a stupid question, and it may have been answered in one of the 175 previous pages, but did you observe a small drift in output voltage as the MJE350s are warming up?

 

[Salas]

Yes its normal. Its Vbe comes to warm up, that is the Jfet's VDS bracket. It has an IDSS in turn which is both VDS and ambient temp sensitive. To add insult to injury there is a Norton resistive reference to directly I/V that, and much heat around because its a high burning shunt usually near to hot tubes! Don't expect any serious drift stability performance from such a design. It listens to the Drifters! OTOH most tube circuits sneeze at few Volt bias change amongst hundreds they need and all that stuff floats and filters elegantly. I.e. Fet CCS, Johnson noise only passive resistive reference, small MKP caps only, few components, parallel current damping, 10mOhm flat Zo exteneded in BW. Matters most hand in hand with tubes IMHO. Tune DC voltage output after well warm and forget. Else, you stick a Zener string instead of the Norton ref and you listen to... Art Of Noise.