120kv 10-20mA SMPS / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

120kv 10-20mA SMPS

first previous 4 5 6 7 next last

Move Thread

LAN_403

Adam Munich

Tue Aug 23 2011, 02:36AM

Registered Member #2893 Joined: Tue Jun 01 2010, 09:25PM
Location: Cali-forn. i. a.
Posts: 2242

Maybe you could go the DC flyback route and half wave rectify it? That would reduce the capacitive + inductive losses, and make insulation a lot easier...

Linas

Tue Aug 23 2011, 07:10AM

Registered Member #1143 Joined: Sun Nov 25 2007, 04:55PM
Location: Vilnius, Lithuania
Posts: 721

jpsmith123 wrote ...

I think it will be very difficult if not impossible to get 65 kv (peak) out of a "typical" ferrite transformer. You will run into problems with parasitic capacitance and/or capacitive coupling to the core, and insulation problems.

I think you'll probably have to go with a significantly lower transformer voltage and a correspondingly bigger multiplier.

thanks for simulation. you get same results as i with miltisim.
I have ferrite core from original high power 140kv CT

now just question, should i wind turns like in photo, or should i use 2 U in parallel, so window will be 6x6cm (not 12x6cm as in photo) but that will double core area to 11cm^2 and will reduce turns by two in secondary as in primary.
right not thinking about 6.5k turns in 12x6cm core, or 3.25K turns for doubled core
and how to calculate voltage degradation caused by parasites ?

jpsmith123

Tue Aug 23 2011, 03:39PM

Registered Member #1321 Joined: Sat Feb 16 2008, 03:22AM
Location:
Posts: 843

Wow I've never seen a ferrite core like that before!

I wonder, how was it used in its previous life, i.e., what was the secondary voltage, drive frequency and winding geometry, do you know?

I think if it was me, my first inclination would be to try to run the secondary at no more than 30 to 35 kv and then quadruple it.

What's the core cross sectional area? Can you get 30 volts(peak)/turn at 25 or 30 kHz without heating up the core too much?
If so you'd need somewhere around 1000 to 1200 turns of 30 gauge (or maybe 28 gauge) wire.

Adam Munich

Tue Aug 23 2011, 03:46PM

Registered Member #2893 Joined: Tue Jun 01 2010, 09:25PM
Location: Cali-forn. i. a.
Posts: 2242

I still think a CW is the way to go. Make it a 4 stage center tapped one, so you get the voltage increase of 4 stages with the problems of only 2.

That core looks much like the one I used for my fryback, and I was only able to get a maximum of 30kV out of it before the core saturated.

Linas

Tue Aug 23 2011, 04:13PM

Registered Member #1143 Joined: Sun Nov 25 2007, 04:55PM
Location: Vilnius, Lithuania
Posts: 721

wow, nice flyback
But my core is at least 2x bigger and has 2-3x cross section.
its 28x20mm ( at least what i am seeing from photo, tomorrow i will have it near me, so i will do measure it)
This core was used in 100KW CT machine, and biggest part of power was used by tube, and tube was powered by flyback.
Flyback was huge, with oil impregnated paper as insulator, and had 20 layers. each layer had it's one 17xBYV26E diode string to reduce capacity.
i have 2 from 3 U shape cores, so i should easy get around 20KW output from this core.
i already have transformer for filament, and made driver for it on my SMPS pcb, so i will keap filament power supply as it.
and most important part:
in my CT scanner box i have 19*19x11cm space for voltage step up part, so i don't want again remade all existing parts

so 65KV and 2x sounds best for me.
( one more question, if i increase V/turn, i will get less parasites, and therefore more voltage in output, right ?, because if i stack U shapes, i just need around 5-10 (15-30V/turn) turns for half bridge, but if i use it in series i need (12x6 window)10-20T (7,5-15V/turn) )

Fabi

Tue Aug 23 2011, 04:55PM

Registered Member #3510 Joined: Mon Dec 13 2010, 08:31AM
Location: western germany
Posts: 19

Hi

Intersting project :)
I'm working parallel on pretty the same project (CT) and I'm designing a SMPS and x-ray sensor

(I need 110 kV at 10-20 mA.) btw.

I think that a CW multiplier is the most-effective solution. (max. 4 stages)
But I have to think once about it... maby there is a better one...

How about a marx-generator? You can trigger it by a trigatron)

The next thing is:
Do you want the tube to produce x-rays continously or do you want to trigger it to the camera?

Anyway: Most likely we have to use a combination of caps and diodes... because it's nearly impossiblt to handle such a voltage with a "stans-alone" transformer with an acceptable budget

I hope, I didn't skip anything...

Dr. ISOTOP

Wed Aug 24 2011, 12:51AM

Registered Member #2919 Joined: Fri Jun 11 2010, 06:30PM
Location: Cambridge, MA
Posts: 652

Fabi wrote ...

Anyway: Most likely we have to use a combination of caps and diodes... because it's nearly impossiblt to handle such a voltage with a "stans-alone" transformer with an acceptable budget

This this this.
There is nothing wrong with a CW, especially at these low currents.
All of the world's high-quality HV supplies use a multiplier stack.

Linas

Wed Aug 24 2011, 03:32PM

Registered Member #1143 Joined: Sun Nov 25 2007, 04:55PM
Location: Vilnius, Lithuania
Posts: 721

i finally get cores, but noticed that if i use 2x multiplier, i have no place for caps and diodes.
So i have to make single 120KV transformer, with 15V/turn and use full 4 diode rectifier, that sounds like breakdown for me

Fabi

Wed Aug 24 2011, 06:22PM

Registered Member #3510 Joined: Mon Dec 13 2010, 08:31AM
Location: western germany
Posts: 19

ok...

I'm gonna do it with a multiplier...

anyhow

Please note that 120kV eff is about 170 kV peak btw...

Linas

Wed Aug 24 2011, 07:05PM

Registered Member #1143 Joined: Sun Nov 25 2007, 04:55PM
Location: Vilnius, Lithuania
Posts: 721

Fabi wrote ...

Please note that 120kV eff is about 170 kV peak btw...

no, because i have square wave from half bridge ( IR2153 controlled ) (150Vac Half-bridge with 300V supply)
check your pm

first previous 4 5 6 7 next last

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