1kv Constant Current Soucre using BJT / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

1kv Constant Current Soucre using BJT

1 2 next

Move Thread

LAN_403

borderline128

Mon Jun 08 2009, 04:39PM

Registered Member #1531 Joined: Tue Jun 10 2008, 08:37PM
Location:
Posts: 8

I want to build a linear constant current source that can be used at up to 1kv. I'm using this simple design:

Const Cur Src 111

My problem is I can't find BJT that goes up to 1kv. Is there a different type of transistor I can use or can I use two BJTs set to have a even voltage drop across each?

Edit: My target max current for this source is 1 amp

big5824

Mon Jun 08 2009, 04:47PM

Registered Member #1687 Joined: Tue Sept 09 2008, 08:47PM
Location: UK, Darlington
Posts: 240

you could probably find a 1000v mosfet, would that work for this application?

Proud Mary

Mon Jun 08 2009, 05:14PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

You could try 2SC3998 or BU508A - 1600V horizontal output transistors - but they have very low hfe (as low as 7 or 8 with high current) and gobble up base current like billy-oh, (8A) so you'd need to take account of that with your base feed arrangements.

If it doesn't work - well - BU508 is dirt cheap!

Dalus

Mon Jun 08 2009, 05:22PM

Registered Member #639 Joined: Wed Apr 11 2007, 09:09PM
Location: The Netherlands, Herkenbosch
Posts: 512

I still have a BU209A here. I could send it to you if it's of any real use to you.

Proud Mary

Mon Jun 08 2009, 05:32PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

I have to say that your circuit isn't really suitable for high voltages, but if you'd like to say how much current was needed, and what the application was, we may well be able to come up with something more realistic. High voltage is what we do here!

Sulaiman

Mon Jun 08 2009, 05:44PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

The design will be heavily influenced by the required output current.
At 0 volts out the power transistor must dissipate (constant current) x (supply voltage) Watts
high voltage power bjt as used in TV and monitor would be suitable
as mentioned above, the current gain will be about 10.
A separate low voltage supply would be efficient and easier to work on.
Almost any plug-pack / wall-wart could be used.
(For initial testing use the low voltage for the output also)
(When you do go 'live' put a mains rated filament lamp in series with the ac line,
you should see a brief flash as the hvdc capacitor charges up, then it should extinguish,
if you have a fault the lamp will stay on .. usually preventing a nasty BANG)

A mosfet as used in smps would also be good.
With the added benefit of negligible gate current,
allowing very low current in the high voltage resistor.
(but you need about 5 V for gate-source and about 5 V or more across the current-limit resistor
because at low currents, as a mosfet warms up the required gate-source voltage drops -
since the zener voltage is faily constant this means the output current increases,
which increases dissipation in the transistor, increasing the temperature of the transistor more
, increasing the current more . etc. etc. = Thermal Runaway.)

e.g. 330 Vdc with 330 kOhm gives about 1 mA zener current, low but ok
... 1 mA x 330 V = 330 mW .... use a 0.5W or more resistor

The correct procedure is;
Select power supply voltage and maximum constant current required,
Find a transistor that is easily available and cheap, get the datasheet for it,
Check the SOA (safe operating area) graph to ensure that the transistor is able to handle
the required current from 0V to Supply V.
The rest is straightforward.

borderline128

Mon Jun 08 2009, 07:11PM

Registered Member #1531 Joined: Tue Jun 10 2008, 08:37PM
Location:
Posts: 8

Harry wrote ...

I have to say that your circuit isn't really suitable for high voltages, but if you'd like to say how much current was needed, and what the application was, we may well be able to come up with something more realistic. High voltage is what we do here!

The application is to charge a 1kv capacitor with a constant current in order to evaluate its C versus V characteristics. I'd like to charge it at around 1 Amp. The supply voltage will be variable from 100v to 1kv.

The capacitor we will be testing will be about 1uF. Because we don't have a power supply capable of 1amp @ 1kv, I was going to use our lower current power supply to charge a source cap of 120uF to 1kv then use this to charge the smaller test cap through a linear regulator.

I've attached the LTSpice file if you wanna take a look at what Ive got so far.

borderline128

Mon Jun 08 2009, 07:12PM

Registered Member #1531 Joined: Tue Jun 10 2008, 08:37PM
Location:
Posts: 8

Attachment
]file.zip[/file]

Sulaiman

Mon Jun 08 2009, 07:43PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

An 'easier' way to determine C vs V may be;
Store energy in an inductor and release it into the capacitor
you could switch a large inductor to get an instant single result
(e.g. 1 uF @ 1 kV = 0.5 J = 1H @ 1 A)
or switch a smaller inductor rapidly (e.g. 1000 pulses of 1 mH @ 1 A) for a 'curve'

Benefit is no hv supply required, e.g. 24 Vdc supply,
about 30 ms on-time for 1A in 1H, 30 us for 1 A in 1 mH etc.

You still need a hv transistor (I'd use igbt or mosfet) and diode and a 556 or a pic.

P.S. For accurate results you must be sure that the inductor is linear
so the 'many small pulses' is simplest.

P.S. I think a bjt with the necessary SOA would be more expensive than a mosfet or igbt.

Proud Mary

Mon Jun 08 2009, 07:48PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

As the design stands, the base feed resistor and the zener diode would have to be of prodigious size.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob