Page 1 of 2

Power Supply from an "interesting" XRT

Posted: Tue Dec 31, 2019 2:41 am
by Pablo Llaguno

I just bought an interesting 65kV transformer from Joshua Guertler viewtopic.php?f=20&t=13104&p=85386#p85386. The plan is to build a 0-30kV power supply that can produce some neutrons in the fusor. Here are some images from the transformer that Joshua took (I haven't got it in my hands yet, I need to pick it up at the US):
XRT specs.png
XRT specs.png (160.61 KiB) Viewed 2141 times
XRT.png (272.67 KiB) Viewed 2141 times
So based on my experience with neon sign transformers, and by reading the HV FAQs, I made this diagram of the electrical system of the power supply:
Power Supply Diagram.png
Power Supply Diagram.png (52.95 KiB) Viewed 2141 times
One main problem of this diagram is that I don't know yet if the transformer is centre tapped so that's something I'll have to check once I've got the xfmr in my hands for testing. The electrical components will consist of a 20A variac that I already have, some diodes, a ballast resistor and some resistors for measuring current and voltage.

By the way, I found it hard to locate some of these components, specially the ballast resistor. I read that some here use a 100W 60kohm wire wound resistor and found this supplier:, while it is a bit expensive I don't have any hamfests where I live so for anyone looking for a 100W ballast resistor this one will do for $25. Also the 300m ohm resistors can be found here (thanks Rex Allers): ... 2749.l2649, these are 5W and the specs say it holds up to 20kV.

The only part I am missing are the HV diodes, but I will wait to see if the XRT is centre tapped or not, because that would change the design of the rectifier. I was thinking of ordering some 20kV 100mA diodes of ebay and build 2 strings of 3 diodes each. At 65kV, if its centre tapped, each coil would produce 32.5kV, a weird value that makes me think this might not be centre tapped.

One more thing. Joshua suggested to put the thing under oil and using a pump to cool it. Since the XRT is rated at 10mA for 2sec in air, it may be able to give 10-15mA for 5min under oil. The only thing I am worried about is if it will produce enough current, since I have a rather large chamber:
Demo System.jpg
Demo System.jpg (104.61 KiB) Viewed 2141 times
I will update you when I have the transformer in my hands, meanwhile any input would be greatly appreciated.

Re: Power Supply from an "interesting" XRT

Posted: Tue Dec 31, 2019 5:06 pm
by Rich Feldman
Good luck, Pablo. Lots to learn from experimenting with any transformer. I have an dental XRT about that size that's still on the workbench, not yet used by me for high voltage.
Electrical characterization reported here:

Here are a few comments regarding things people have said in this thread.

The 2-second limit on x ray head is from the little Coolidge tube, not the XRT itself. Tube anode, compared to transformer, receives a 10 times bigger fraction of the input power, and weighs 100 times less. As Richard said, a normal session is a few shots a minute apart, each lasting a fraction of a second. Tube and transformer were designed accordingly.

As for duty cycle of the transformer itself, it's misleading to think in terms of a watt-seconds limit.
Power loss in core is a nonlinear function of voltage & has nothing to do with secondary current.
Power loss in windings is proportional to square of current, and has nothing to do with the voltages. Easy to figure out from winding resistance.

The oil is for high voltage insulation, not cooling. Search the forum for stories about using vacuum to eliminate air bubbles in secondary windings. It's best to apply the vacuum before admitting oil to tank. That's more trouble, but is how professional HV transformer makers/installers and X-ray repair techs always do it.

I have never heard of anyone actively pumping oil in an XRT tank, but suppose it could be done. Might be easier to arrange tank for good natural convection in the oil, and blow air over the outside if you ever run long enough for whole tank to get hot.

Re: Power Supply from an "interesting" XRT

Posted: Tue Dec 31, 2019 10:31 pm
by Rich Feldman
Forgot to mention some details peculiar to XRT nameplates.
  • Primary voltage value is RMS.
  • Secondary voltage value is peak (indicated by the symbol kVp), because that's what matters to radiologists. It's common, but sloppy, practice to use kVp and VRMS and VAC as if they represent units of measurement, which they don't.
  • Secondary current is average DC value in a half-wave rectified (self rectified) HV circuit, in this case. For practical convenience, heat capacity of X-ray tubes is often given in heat units (peak kilovolts * average mA * seconds) instead of watt-seconds or joules. Value is significantly different for operation on DC (from high frequency power converter) compared to half-wave rectified 60 Hz sinusoids.
I bet your transformer secondary is center-tapped, actually two separate windings. Usually the start of one winding is tied to the grounded core, and the start of other winding is brought out for current monitoring. An obvious modification for your application is to tie both "start of winding" wires together, as center tap, and put current sense resistor between that and the grounded core. Could all be done inside the oil tank.

65 kVp is an extremely common maximum voltage in dental x-ray generators. 32.5 kV peak is plenty for fusion.

Re: Power Supply from an "interesting" XRT

Posted: Wed Jan 01, 2020 12:30 am
by Pablo Llaguno
Thank you for your input Rich, I always appreciate it.

I think you are right about using oil for insulation rather than cooling. Also, I think that with a big enough case, natural convection and a fan (as you said) would allow for "long" runs. About eliminating air bubbles inside the transformer, I once saw this video of a guy building a 50kV supply (impressive work) and what he did is put a flare fitting in the case and applied the vacuum with the case half filled with oil. ... dex=2&t=0s @min 37. I'll search through the forums to see if there is an easier way to do this.

If as you say the transformer secondary is center-tapped, then I will tie these two "start of winding" wires and use them as the center tap, which will then be used as ground and for the current measuring resistor. By the way I am searching for some HV wire, since the supply will provide 32.5kV peak I was thinking 18AWG 30kV wire might suffice? Walmart is selling 5 meters of this wire at $9 so I am tempted to buy it. ... /244434583

Lastly, if each winding gives 32.5kV then could the diode string have in total a series PIV of 40kV? Or should I go to the safe side and aim for 65kV?

Re: Power Supply from an "interesting" XRT

Posted: Wed Jan 01, 2020 12:51 am
by Rich Feldman
The latter.
When one secondary terminal and the rectifier output are at -32 kV, the other secondary terminal is at +32 kV.
So diode needs to withstand the end-to-end secondary voltage.

Re: Power Supply from an "interesting" XRT

Posted: Wed Jan 01, 2020 9:28 am
by Rex Allers

The voltages across the diodes in this center-tapped full wave circuit is something you pointed out to me in an earlier thread.

The full end-to-end voltage requirement is something I missed after dozens of years of looking at that circuit configuration. It is important and many, like me, might not see it and think only the output voltage (half transformer secondary) is needed.

Thanks, Rich, for putting me right and for sharing again this information that I missed and others might too.

Re: Power Supply from an "interesting" XRT

Posted: Wed Jan 15, 2020 6:32 pm
by Pablo Llaguno
The transformer arrived and I did a couple of tests yesterday. First I measured resistance on the wires to find the primary and the secondary outputs. It was quite easy because the primary read less than 5 ohms and each secondary read 90kohm. After that I proceeded to test the output of the transformer in air with a small input voltage. The way I did it is with a method that I read in another post (by Rex Allers if I recall correctly). It goes something like this: Plug a 12V transformer to a Variac and with that you have a precise way of controlling the input voltage to your XRT. I did it slightly differently using a power supply set to 2V and powering it with the Variac, that way I had full control from 0 to about 3V input for the XRT.
IMG_1492.jpg (118.74 KiB) Viewed 1748 times
As you can see in the picture I used two DMM, one for measuring input voltage and the other coupled with a high voltage probe (basically a voltage divider) for the output of the transformer. Nevermind the high voltage probe, it was very inaccurate and since my output was acceptable for the DMM I just used the banana wires for the tests.

I'll summarize the tests with these pictures:
- Input voltage was 1V, measuring voltage between ground (core) and each output:
Screen Shot 2020-01-15 at 12.22.02 PM.png
Screen Shot 2020-01-15 at 12.22.02 PM.png (121.61 KiB) Viewed 1748 times
- Input voltage was 1V, measuring voltage between the two output wires of the secondary:
Screen Shot 2020-01-15 at 12.22.13 PM.png
Screen Shot 2020-01-15 at 12.22.13 PM.png (121.96 KiB) Viewed 1748 times
With this I now know the transformer is in working order and the next thing to do is build its case and add the rectifier and measuring systems. Rich was right in that the secondary is center-tapped, and as he said one "start of winding" is connected to the core. However the other winding is odd because it has its "start of winding" connected to the core, however it has two HV outputs. My plan would be to just leave one of this outputs isolated and solder a thick cable to the core as my center-tap for the current measuring resistor and ground to the fusor.
Screen Shot 2020-01-15 at 12.22.27 PM.png
Screen Shot 2020-01-15 at 12.22.27 PM.png (269.54 KiB) Viewed 1748 times

Re: Power Supply from an "interesting" XRT

Posted: Wed Jan 15, 2020 7:08 pm
by Mark Rowley
I have one of those transformers but elected not to use it due to its smaller size and potentially limited duty-cycle. My Fusor overtaxed a couple that we’re about double the size so I didn’t want to risk damaging it. I hope it works out and am interested in how you ballast and power it.

Mark Rowley

Re: Power Supply from an "interesting" XRT

Posted: Wed Jan 15, 2020 8:35 pm
by Rich Feldman
Good work so far, Pablo! Thank you for sharing.

>>one "start of winding" is connected to the core. However the other winding is odd because it has its "start of winding" connected to the core, however it has two HV outputs.

That detail is like my similar-size XRT, except mine has three HV outputs. Either way, the secondary with more than one connection (not counting "start of winding") is for the cathode end of the x-ray tube. Your two HV outputs are probably connected internally to a Filament winding designed for a few volts and a few amperes. The x-ray tube self-rectifies the HV current, by not conducting when cold electrode is negative.

When a third HV output is present, it's internally connected to the others with a resistor of some tens of kΩ. X-ray tube would have a third terminal at cathode end, connected to the "focus cup" metal behind and around the cathode. Arranged so the HV current, passing through the resistor, develops a bias of a couple hundred volts between focus cup and cathode. Designed to reduce the e-beam spot size on anode, and also to somewhat regulate the HV current.

Re: Power Supply from an "interesting" XRT

Posted: Wed Feb 19, 2020 10:55 pm
by Pablo Llaguno
Got the supply all wired up and ready to be filled with oil:
IMG_0055.jpg (58.06 KiB) Viewed 1493 times
Thankfully I have a friend who will help me with filling the tank with oil. He has years of experience in the utility transformer business and has equipment to correctly fill the tank. His idea is to bake the supply (without the tank) for 48hrs at 100ºC and the apply the high voltage mineral oil. I am kind of worried about the diodes or the ballast breaking up with that amount of heat. Does anyone have experience with this to know if the components of the supply could handle the heat?

Alternatively, we could apply a vacuum to the tank and then apply the oil, however I am not sure if this would be enough as I don't know how much time this transformer has been outside its oil tank.