CCS transistor choices

[hirafu_boarder]

I'm pretty new to transistors but would like to experiment with CCS as a anode load (eg B+ up to 500V and Ip up to 50mA).

I've read Morgan Jones' intro to CCS and believe for anode load CCS I need PNP transistors.
His table on p139 lists out the following parameters Vce(max), Ic(max), Pmax,ft, hFE (min) and 1/hoe(typ).

Unfortunately being in Japan I cannot readily acquire the recommended MJE350 nor MPSA92 so I would like to source some local alternatives.

Are these the only parameters I need to watch out for in selecting the power transistor?

eg would the 2SA1486 be a good choice?
http://www.classiccmp.org/rtellason/...ta/2sa1486.pdf

Many thanks!
Jeff

[SY]

That looks like a superb choice. Besides voltage rating, the hfe is reasonably high and the speed (gain-bandwidth) is reasonable.

[Eli Duttman]

Yes, "P" type semiconductors are better for constant current sources, as the high impedance electrode (collector or drain) feeds the tube. However, GOOD results have been achieved with "N" type stuff. Check the work of Gary Pimm out.

FWIW, I'm a MOSFET, as opposed to BJT, fan. If Fairchild semiconductors are available to you check the FQP3P50 out.

[SY]

FWIW, I'm a MOSFET, as opposed to BJT, fan.

So am I. I've gotten excellent results from bipolar cascodes, but MOSFET cascodes (using n-channel) are even better and simpler.

[Geek]

I have a nice MOSFET CCS design (source and sink) that has proven itself quite well

Cheers!

[Eli Duttman]

Originally Posted by SY

So am I. I've gotten excellent results from bipolar cascodes, but MOSFET cascodes (using n-channel) are even better and simpler.

I wonder if something could be made of a DN2540 "upstairs" and a FQP3P50 "downstairs". A 100 Ω resitor between the 2 source electrodes would provide a test point and reverse bias for the depletion mode FET. A 1 MOhm pot. sitting between B+ and ground would have its wiper connected to the enhancement mode "P" channel FET's gate.

[jcx]

there are depletion mode mosfets that would work as two terminal ccs without the extra pass Q

IXYS
IXTP01N100D
VDSS, max, (V) 1000
ID(cont), TC=25°C, (A) 0.1
RDS(on), max, VGS = 0 V, (Ω) 110
VGS(off), max, (V) -5
Ciss, typ, (pF) 120
Crss, typ, (pF) 3
Qg, typ, (nC) 5
PD, (W) 25


http://ixdev.ixys.com/DataSheet/98809.pdf

I've purchased them from Mouser @ US$0.80 ea

[Eli Duttman]

Originally Posted by jcx

there are depletion mode mosfets that would work as two terminal ccs without the extra pass Q

IXYS
IXTP01N100D
VDSS, max, (V) 1000
ID(cont), TC=25°C, (A) 0.1
RDS(on), max, VGS = 0 V, (Ω) 110
VGS(off), max, (V) -5
Ciss, typ, (pF) 120
Crss, typ, (pF) 3
Qg, typ, (nC) 5
PD, (W) 25


http://ixdev.ixys.com/DataSheet/98809.pdf

I've purchased them from Mouser @ US$0.80 ea

IXYS' 10M45S has acquired considerable notoriety. The problem with single device setups is a comparatively low dynamic AC impedance. Cascodes do much better in that important dept. One device = good. Two devices = better.

[jcx]

depletion mode enables two terminal floating ccs with more complex current regualtion if needed

this simulated circuit is limited by Crss (~5 pF for IXTP01n100) |Z|> 1 MOhm @ 20 KHz



MOSFET model uses fit to measured gm points, datasheet C

[hirafu_boarder]

Thank you guys.
For the first attempt, I'll go with the 'P' BJT in a cascode config.
I need a bit of time to get my head around any transistor design that gets more complicated than that!

For my next trick, perhaps MOSFETs.
Neither the IXYS nor the FQP3P50 are available locally, so for MOSFET based CC Source I would need to find alternatives.
Would the key parameters to meet be VDSS, ID, PD?
If so how does the 2SJ117 stack up (for 400V)?
http://www.classiccmp.org/rtellason/...ata/2sj117.pdf

Thanks again,
Jeff