High voltage current and voltage sources

I failed to mention some of the specs: DC, continuous operation, maximum 1000V, maximum 10mA-1A. At this point I'm only worried about the low end current range of about 10mA but in the future I do believe I might need possibly a 1000W supply

The short answer is that it is possible, but it would take a complicated, bulky and expensive circuit. Anybody who took the job on would want to know a lot more about what you were actually trying to do, and why.

If you don't want to publish the kind of detail in a public forum, you are welcome to e-mail me directly (the adress I post under here - bill.sloman at ieee.org - is real and works). It's unlikely that I would be able to do the job for you but I've got a couple of private e- mail adresses that might be able to help.

-- Bill Sloman, Nijmegen

Why do you want it to be transformerless?

How big is compact?

At 1000W it likely has to be mains powered. What is the mains voltage?

What does the load look like?

Can the output terminals look capacitive?

On Sun, 12 Apr 2009 02:06:35 -0700 (PDT), James Rollins wrote:

--- Why not try this:

and configure it like this: (view in Courier)

.+V>-----------------------+ F10 SUPPLY . | NPN +----------+ .+15V>--+--------------+ +--C E-----|+IN +HV|--------->HVOUT . | | B | | . [R1] | | +--|-IN -HV|--+ +-->HVRET . | | | | +----------+ | | . +------+ | | | | | . | | | | | | | . [REF] [POT]------+ | | | . | | +--|-/ | | | . | | | | | | | . | | +---|-------------|----------------|---+ . | | | | | | . | | | | | [R5] . | | | | | | .GND>---+------+-------+-------------+----------------+---+

Looks like it might even work! :)

Version 4 SHEET 1 1444 680 WIRE 704 -16 592 -16 WIRE 592 32 592 -16 WIRE 704 32 704 -16 WIRE 32 48 -80 48 WIRE 192 48 32 48 WIRE 256 48 192 48 WIRE 416 48 352 48 WIRE 544 48 416 48 WIRE 544 96 512 96 WIRE 32 112 32 48 WIRE 704 160 704 112 WIRE 704 160 128 160 WIRE 192 208 192 48 WIRE 128 224 128 160 WIRE 160 224 128 224 WIRE 304 240 304 112 WIRE 304 240 224 240 WIRE 416 240 416 48 WIRE 704 240 704 160 WIRE 32 256 32 192 WIRE 160 256 32 256 WIRE -80 304 -80 48 WIRE 32 304 32 256 WIRE -80 416 -80 384 WIRE 32 416 32 368 WIRE 32 416 -80 416 WIRE 192 416 192 272 WIRE 192 416 32 416 WIRE 416 416 416 320 WIRE 416 416 192 416 WIRE 512 416 512 96 WIRE 512 416 416 416 WIRE 592 416 592 112 WIRE 592 416 512 416 WIRE 704 416 704 320 WIRE 704 416 592 416 WIRE -80 480 -80 416 FLAG -80 480 0 SYMBOL npn 256 112 R270 SYMATTR InstName Q1 SYMATTR Value 2N3904 SYMBOL res 16 96 R0 SYMATTR InstName R4 SYMATTR Value 4.7k SYMBOL References\\\\LT1009 32 336 R0 SYMATTR InstName U2 SYMBOL voltage -80 288 R0 WINDOW 0 18 7 Left 0 WINDOW 3 19 105 Left 0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V1 SYMATTR Value 15 SYMBOL res 688 224 R0 SYMATTR InstName R2 SYMATTR Value 249 SYMBOL e 592 16 R0 WINDOW 3 10 120 Left 0 SYMATTR InstName E1 SYMATTR Value 83.3 SYMBOL res 688 16 R0 SYMATTR InstName R3 SYMATTR Value 100k SYMBOL res 400 224 R0 SYMATTR InstName R1 SYMATTR Value 1000 SYMBOL Opamps\\\\LT1007 192 176 R0 SYMATTR InstName U1 TEXT -66 442 Left 0 !.tran 1 TEXT 536 136 Left 0 ;Av =

JF

Fairly large, if it includes surface and airflow for extracting

1kW worth of heat.

A transformer would be smaller.

Cheers, James Arthur

Define 'transformerless'; a Model T spark coil generates high voltage, it's CALLED a coil but it's also an autotransformer.

An LM317 requires several milliamps output current for normal bias, so it won't do the low current range. As for high voltage, you can get amplifiers (grounded-base transistor) that take current at low voltage as input (emitter current) and generate current at high voltage as output (collector current). That means you can potentially separate the 'current regulate' and 'high voltage' problems.

There are (currently on the market) no good references for high voltage; the Zener diode, or bandgap reference, or integrated circuit regulator using them, will be a part of your low voltage section. There ARE some high voltage ICs, but unless you're manufacturing motor controls it's unlikely you can find ones that do what you want. Integration doesn't, in and of itself, constitute a virtue.

I've used 4.5 GV systems, which qualify as 'exotic'. Most of the

1kV-and-under electronic field seems ... domestic... to me.

On Apr 12, 11:56=A0am, whit3rd wrote: [.. good stuff ..]

MOSFETs as common gate stages work too.

--
Oops...
Appearances can be deceiving.

The circuit needs work, sorry. :-(

JF

Should work, with maybe some loop compensation added.

You can replace the NPN with an LM317 or an LM1117 and get current/thermal limiting for free.

And bypass the inverter input.

John

Sorry for not being clear and thanks for the replies.

I want to create several power supply from mains(110V or 220V AC) to ~0-1000V DC @ 10-100mA. I need several *variable/programmable* voltage and current sources in the same housing. Weight is a big issue but temperature/heat is not if within reason. These sources need to operate continuously.

The initial supply can actually be from DC by using a large transformer rather from mains if need be. Ripple is not a huge issue <

1-3%. Size is important but secondary to weight. The loads is mainly resistive but with with some capacitance probably less all less than 1nf.

I believe that any standard topology used for for standard microelectronics projects, if adaptable to the higher voltage, would work fine. I am drifting towards using a simple switching topology.

Seems like it would work fine if using igbts and give me the variability/probability I desire and somewhat lower in weight than an equivalent transformer based system.

Also, of course is the standard

.

Again, regulation is not necessarily an issue as I can do it before the power supply so a simple linear attenuation could work, at least for prototyping. Switchers seem to be the way to go though. My question is, would it be difficult to adapt such topologies to the voltages I require? As far as I can tell I would just have to find the rated components? Is this correct?

Thanks, James

--
This is better...

Version 4
SHEET 1 1444 680
WIRE 704 -64 592 -64
WIRE 592 -16 592 -64
WIRE 704 -16 704 -64
WIRE 256 0 -224 0
WIRE 416 0 352 0
WIRE 544 0 416 0
WIRE 32 48 -80 48
WIRE 192 48 32 48
WIRE 544 48 512 48
WIRE 32 112 32 48
WIRE 704 160 704 64
WIRE 704 160 128 160
WIRE 192 208 192 48
WIRE 128 224 128 160
WIRE 160 224 128 224
WIRE 304 240 304 64
WIRE 304 240 224 240
WIRE 416 240 416 0
WIRE 704 240 704 160
WIRE 32 256 32 192
WIRE 160 256 32 256
WIRE -224 304 -224 0
WIRE -80 304 -80 48
WIRE 32 304 32 256
WIRE -224 416 -224 384
WIRE -80 416 -80 384
WIRE -80 416 -224 416
WIRE 32 416 32 368
WIRE 32 416 -80 416
WIRE 192 416 192 272
WIRE 192 416 32 416
WIRE 416 416 416 320
WIRE 416 416 192 416
WIRE 512 416 512 48
WIRE 512 416 416 416
WIRE 592 416 592 64
WIRE 592 416 512 416
WIRE 704 416 704 320
WIRE 704 416 592 416
WIRE -224 480 -224 416
FLAG -224 480 0
SYMBOL npn 256 64 R270
SYMATTR InstName Q1
SYMATTR Value 2N3904
SYMBOL res 16 96 R0
SYMATTR InstName R4
SYMATTR Value 4.7k
SYMBOL References\\\\LT1009 32 336 R0
SYMATTR InstName U2
SYMBOL voltage -80 288 R0
WINDOW 0 18 7 Left 0
WINDOW 3 19 105 Left 0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V1
SYMATTR Value 15
SYMBOL res 688 224 R0
SYMATTR InstName R2
SYMATTR Value 249
SYMBOL e 592 -32 R0
WINDOW 3 10 120 Left 0
SYMATTR InstName E1
SYMATTR Value 83.3
SYMBOL res 688 -32 R0
SYMATTR InstName R3
SYMATTR Value 100k
SYMBOL res 400 224 R0
SYMATTR InstName R1
SYMATTR Value 1000
SYMBOL voltage -224 288 R0
WINDOW 0 18 7 Left 0
WINDOW 3 19 105 Invisible 0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V2
SYMATTR Value PULSE(0 50 .1 1)
SYMBOL Opamps\\\\LT1218 192 176 R0
SYMATTR InstName U3
TEXT -202 442 Left 0 !.tran 1
TEXT 536 136 Left 0 ;Av =

There\'s still some stuff up in the air, like the choice of the pass
transistor, and that\'s going to depend on your supply voltage and the
range of resistance the load is expected to be.

Can you elaborate on that?

JF

--
I Agree.

What do you think, some brute force C across R5?

Something like this ought to work...

Where R3 may not be needed if R5 is fairly big.

Maybe a diode or zener across R5 too, in case the load arcs or something, to protect the opamp.

John

--
That\'s all well and good, but you need to specify precisely what you
want instead of just spouting generalities.

4.5GV ?

With spice I can deal with that everyday, but is that really GV? I can think of MV, but GV... What was that?

-- Thanks, Fred.

"James Rollins" a écrit dans le message de news: snipped-for-privacy@w40g2000yqd.googlegroups.com...

Uh, how can this have been granted a patent?

Is the next step patenting a two stages resistor divider? Well, no it can't now. I've just made it public domain...

-- Thanks, Fred.

On Sun, 12 Apr 2009 16:51:38 -0700, John Larkin wrote:

--- The LM1117 won't work in there because it can't supply the 1.5A the supply needs under full load (1kV, 10mA).

The value of R5 is fixed by the drop needed at 10mA through the load and the reference voltage needed to drive the output to 10mA. I chose a 10V reference, so R5 is 1000 ohms.

Other than C1, it doesn't look like any of the other stuff is needed since a cap from the opamp's inverting input to GND cleans everything up.

According to LTspice:

Version 4 SHEET 1 1444 680 WIRE 416 -64 304 -64 WIRE 304 -16 304 -64 WIRE 416 -16 416 -64 WIRE -112 0 -432 0 WIRE -32 0 -112 0 WIRE 128 0 64 0 WIRE 256 0 128 0 WIRE 256 48 224 48 WIRE 416 192 416 64 WIRE 416 192 -192 192 WIRE -112 240 -112 0 WIRE 128 240 128 0 WIRE -192 256 -192 192 WIRE -144 256 -192 256 WIRE 16 272 16 64 WIRE 16 272 -80 272 WIRE -144 288 -320 288 WIRE 416 400 416 192 WIRE -432 416 -432 0 WIRE -320 416 -320 288 WIRE -192 416 -192 256 WIRE -432 544 -432 496 WIRE -320 544 -320 496 WIRE -320 544 -432 544 WIRE -192 544 -192 480 WIRE -192 544 -320 544 WIRE -112 544 -112 304 WIRE -112 544 -192 544 WIRE 128 544 128 320 WIRE 128 544 -112 544 WIRE 224 544 224 48 WIRE 224 544 128 544 WIRE 304 544 304 64 WIRE 304 544 224 544 WIRE 416 544 416 480 WIRE 416 544 304 544 WIRE -432 624 -432 544 FLAG -432 624 0 SYMBOL npn -32 64 R270 SYMATTR InstName Q1 SYMATTR Value 2N3055 SYMBOL voltage -432 400 R0 WINDOW 0 18 7 Left 0 WINDOW 3 19 105 Left 0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V1 SYMATTR Value 20 SYMBOL res 400 384 R0 SYMATTR InstName R2 SYMATTR Value 1K SYMBOL e 304 -32 R0 WINDOW 3 10 120 Left 0 SYMATTR Value 83.3 SYMATTR InstName E1 SYMBOL res 400 -32 R0 SYMATTR InstName R3 SYMATTR Value 10K SYMBOL res 112 224 R0 SYMATTR InstName R1 SYMATTR Value 10k SYMBOL Opamps\\\\LT1218 -112 208 R0 SYMATTR InstName U3 SYMBOL cap -208 416 R0 SYMATTR InstName C1 SYMATTR Value 1e-7 SYMBOL voltage -320 400 R0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V2 SYMATTR Value 10 TEXT -424 584 Left 0 !.tran 1 TEXT 248 88 Left 0 ;Av = With a few tweaks, this works:

. EMCO F10 . 2N3055 +----------+ .+20V>--+--------------+------C E--+--|+IN +HV|--------->HVOUT . | | B C1 | | . [5k1] | | +--|-IN -HV|--+ +-->HVRET . | | | | +----------+ | | . +------+ | | | | | . | | | | | | | . | [10k]------+ | | | . [LM4040-10] | +--|-/LT1218 | | | . | | | | | | | . | | +---|-------------|----------------|---+ . | | | | | | | . | |[100nF | | | [1k] . | | | | | | | .GND>---+------+---+---+-------------+----------------+---+

JF

Works in real life, or in simulation? Your model ignores any dynamics inside the F10. And that opamp may not be able to drive the base of the NPN... 1.5 amps, beta 20, needs 75 mA of base current. The LT is only good for a few mA.

John

snipped-for-privacy@a7g2000yqk.googlegroups.com...

It was an ultrarelativistic synchrotron, SPEAR by name. Basically a vacuum tube with a bunch of trapped electrons (positive electrons, usually) going in circles.

Individual kinetic energy of the positrons was routinely set at 4.5 GeV, there weren't actually any static fields that were creating such a large potential difference.

The power supply was a two-mile-long linear accelerator... strung out collections of megawatt power klystrons... like I say, exotic. The best part was the safety orientation, all about the various safety features (4 foot concrete for radiation shielding; alarms for: hydrogen leaks, earthquake, and radiation; interlocks that could shutdown the whole two-mile linear accelerator if your doors weren't shut).