My X-Ray Machine

[neukyhm]

Hi everyone, I'm new here. I didn't know this forum existed until the user station240 at EEVblog forum suggested to post my machine here too. I'm glad I found this since I'm in love with everything involving high voltage.

I built an x-ray machine as my final degree project. Quick details:

-Fully 3D printed case
-Modified ZVS Mazzilli driver optimized for high frequency
-CEI OX/70-G4 x-ray tube
-Arduino UNO, so it's remotely controlled
-Custom HV transformer made with an ETD59 core (primary is 3+3 turns, secondary is 232 turns)
-Work frequency: ~110kHz
-Power: ~80W



My college's logo




I uploaded a video showing the construction process. The video is not intended to be a tutorial. I'm planning to release the project paper so anyone can see the details, but it must be after I get my mark.

Here are some tests:

[Mads Barnkob]

Hi neukyhm and welcome to HVF

That is a nice little machine you built, you really went all in on the details in the construction, which leads me to say that its a wonderful build process video. I like the body cam / CAD overlay combination.

Now for the questions

What body cam are you using?

What kind of intensifying screen is it that you are using? I have on/off been looking for phosphorus casettes for future project, but never found any reasonable priced.

What kind of control do you have between the Arduino UNO and the ZVS HV supply? Is it just on/off control of the supply voltage or did you extend the ZVS driver with more active driver circuitry?

[neukyhm]

Hi neukyhm and welcome to HVF

That is a nice little machine you built, you really went all in on the details in the construction, which leads me to say that its a wonderful build process video. I like the body cam / CAD overlay combination.

Now for the questions

What body cam are you using?

What kind of intensifying screen is it that you are using? I have on/off been looking for phosphorus casettes for future project, but never found any reasonable priced.

What kind of control do you have between the Arduino UNO and the ZVS HV supply? Is it just on/off control of the supply voltage or did you extend the ZVS driver with more active driver circuitry?

Hi Mads, thanks for the welcome

I'm using a Gopro Hero 4, wich is not the best option. These cameras are really bad at low light conditions, you can see in the video that it captures very well only the part of the table where the lamp is.

My screen is an old Fuji Speed Green I got at an auction on eBay for 6€, I was very lucky. If you are looking for screens, I have seen some Kodak Lanex on eBay too, they are a bit expensive though

In the video you can see that I'm mounting a relay module with two relays. One of them cuts the x-ray tube's filament and the other one turns on and off the ZVS, no extra circuitry, just this relay between the power supply and the ZVS.

[Mads Barnkob]

Congratulations on getting mentioned first row in hackadays newsletter!

[Max]

Hi neukyhm!


Great to see the developpment documented in your video. Really nice job and fascinating results! Some shots seem to have quite a high resolution.

X-rays are fascinating, and one day, I'll make my own experiments for sure. However, I'm very, very concerned about shielding. In the video description, you write:

it has proper lead shielding. I also used a Geiger counter to make sure there were no radiation leaks.

I cannot see any shielding there. Is the whole room shielded? In any case, I'd alway be concerned about leaks, as your case can never be completely sealed (you need holes for cables, etc).
And, as I've a working x-ray tank (well, the full bridge and control circuit is missing, but transformer, rectifier, voltage divider, etc is all there) capable of a few kW, I consider shielding as crucial. My problem is, if there are any leaks, I and possibly others would be exposed while searching them.

Therefore any details about shielding would be highly appreciated!


Kind regards,
Max

[neukyhm]

Congratulations on getting mentioned first row in hackadays newsletter!

Yes, I'm very happy I wasn't expecting that they published it so quickly.

Hi neukyhm!


Great to see the developpment documented in your video. Really nice job and fascinating results! Some shots seem to have quite a high resolution.

X-rays are fascinating, and one day, I'll make my own experiments for sure. However, I'm very, very concerned about shielding. In the video description, you write:

it has proper lead shielding. I also used a Geiger counter to make sure there were no radiation leaks.

I cannot see any shielding there. Is the whole room shielded? In any case, I'd alway be concerned about leaks, as your case can never be completely sealed (you need holes for cables, etc).
And, as I've a working x-ray tank (well, the full bridge and control circuit is missing, but transformer, rectifier, voltage divider, etc is all there) capable of a few kW, I consider shielding as crucial. My problem is, if there are any leaks, I and possibly others would be exposed while searching them.

Therefore any details about shielding would be highly appreciated!


Kind regards,
Max

Hi Max!

At 4:24 in the video, I'm attaching the lead shielding to the box. The thing is that using a lot of lead all around the box and taking care of the holes for the cables and everything is pretty much pointless because of reflexion. The x-rays will reflect on the room walls and you will have radiation going backwards.

What I did is using 1mm of lead that removes around 97-99% of the photons energy. Taking into account that the machine is activated remotely (with a countdown or via teamviewer to send the command to arduino) I think it's safe to operate. You are just waiting in another room or outside the house.

Using teamviewer you can even activate the machine from another country.

[Agent Smith]

Hello,
Great project, a lot of work I bet.
This very night, actually 5 minutes ago i was able to get my first radiograpf after months of trying using that very same tube, 70-G4.
It seems something is wrong with earth and filament connection.
My does not look as good as yours but hopefully you can point me in right direction, what am I doing wrong?
I am miles from that sharp, crisp image you are getting as you can see. Any suggestions?
Thanks

//OCV

[neukyhm]

Hello,
Great project, a lot of work I bet.
This very night, actually 5 minutes ago i was able to get my first radiograpf after months of trying using that very same tube, 70-G4.
It seems something is wrong with earth and filament connection.
My does not look as good as yours but hopefully you can point me in right direction, what am I doing wrong?
I am miles from that sharp, crisp image you are getting as you can see. Any suggestions?
Thanks

//OCV

Hi Agent Smith, check out my connections:


The two black wires that goes to the multiplier are from the transformer. The small red wire that goes to the tube is the filament and I have the tube's blue wire grounded via the thicker red wire that goes out the box. The tube's black wire is the grid and you have grounded it. It's recommended to not ground the grid, but to make it a bit more negative than the filament, how? with that resistor you see in the picture.

I suggest that you use the red wire to power the filament with +2.50 to +2.75 V, ground the blue wire and connect the black one to the cathode of you HV PSU. Once you've done this, it's time to common reference everything, and you will do this with a 47kOhm resistor placed between the cathode (negative pole) of you HV PSU (where it's also connected the tube's black wire) and the tube's red wire, just like I have in the picture. This way you will have the tube connected as the datasheet says
, although they use different colors.

Also, if your HV PSU is unrectified, it's a very good idea to place one of those big 20kV 10nf capacitor between the tube's black wire and ground, it will stabilize voltage a lot and improve results. This won't help that much if your HV PSU is rectified like a TV flyback.

Note: of course, the cathode of the PSU you are using to power the filament has to be grounded too.


Edit: I hope this helps.



Edit2: typo in the capacitor, it's 20kv 10nF, like these.

The dielectric oil is very recommended.

PS: where did you get the tube?

[Agent Smith]

Hi,

Thank you for helping out, I destroyed a few stepdown converters for fillament heating due to wrong connecion.
I got my tube on E-bay long time ago, like two years back. As you can see it came with a capacitor but I guess I will remove it an install 20kV as you said and resistor as well.
I am usind DC to drive the tube but I guess that HV needs to be incresed ass well so my HV supply will get three more stages of cascades.
Your solution is much neater having everything in one box in oil, tube as well as HV.
As you can see I had tube outside and HV separately connected via HV cables.
Can you please comment on my resistor connected on HV+ it is 2,2MOhm and I put it there to prevent current surge inside the tube?
You have none so I wonder is it needed at all then?
My plan is to get this to work and then build nice looking Box to put everything in. I am thinking controlling it by Futaba RC remot that is left from an old drone I had long time ago.
Cool that are more then one self building this kind of devices.
Thanks again.
//OCV

[neukyhm]

Hi,

Thank you for helping out, I destroyed a few stepdown converters for fillament heating due to wrong connecion.
I got my tube on E-bay long time ago, like two years back. As you can see it came with a capacitor but I guess I will remove it an install 20kV as you said and resistor as well.
I am usind DC to drive the tube but I guess that HV needs to be incresed ass well so my HV supply will get three more stages of cascades.
Your solution is much neater having everything in one box in oil, tube as well as HV.
As you can see I had tube outside and HV separately connected via HV cables.
Can you please comment on my resistor connected on HV+ it is 2,2MOhm and I put it there to prevent current surge inside the tube?
You have none so I wonder is it needed at all then?
My plan is to get this to work and then build nice looking Box to put everything in. I am thinking controlling it by Futaba RC remot that is left from an old drone I had long time ago.
Cool that are more then one self building this kind of devices.
Thanks again.
//OCV

Yes, remove the capacitor the tube has, I did that.

Don't use that 2,2M resistor, the tube's grid is in charge to manage the current through the tube, as long as you configure it with the 47k resistor. Be sure to have a good grounding, this is critical or you may break the stepdown converter.

[Agent Smith]

Great info, thanks.
I am reworking it as suggested when dayjobb allows and will post the results as soon I finish it.
In the meantime take a look (it's OK to laugh) at my previous overambitious (read FAILED) project that awaits correction.
It was 140kV power supply with +70 / -70 Kv terminals.
As you know by now Grounding is not my Forte, haha, and I could not get Foating anode to work, it shorted out over the led casing no matter the amount of insulation.
In the picture you can see oil cooler, oil pump, temp sensor, flow meter, oil level interlock etc. Now I just laugh when I look at it  .
I will get it to work but I need to understand the little one first.

[neukyhm]

140kV could be too much for the mineral oil. If it's still arcing I suggest using a bigger container so that the walls are further from the HV parts.

If your HV source is designed to be middle grounded then you cannot ground one of the terminals but the middle of the transformer so one of the outputs is +70kV and the other is -70kV. However this also prevents you from using a simple step down converter to power the filament. I could use a simple converter because I had my cathode grounded so it's at 0V, but you will have a -70kV cathode, so you must use an isolated transformer for the filament (a step-down converter like those on eBay is NOT isolated).

You could also experiment grounding the cathode so you can use a step-down converter, but I'm not so sure about this.

[Agent Smith]

Hello again,

At last I got my machine to work, thank to your tips.
Pictures obtained are much better then before and I guess they will be even better when folter capacitors arrive.
For some reason I could not get your Mazilli driver to work so I used an old driver and your custom transformer.



Regards,
 O.C.

[klugesmith]

Jut noticed this thread.  One of Agent Smith's pictures from small tube project shows a grey plastic pipe section, perhaps full of oil for insulation and cooling.

I never got as far with DIY radiography as you guys.

But did recognize that PVC plastic is not particularly transparent to x-rays.
At dental tube energies, PVC will attenuate and scatter x-rays more than aluminum metal, or bones, because of all the chlorine atoms.
https://physics.nist.gov/PhysRefData/XrayMassCoef/tab4.html

[Agent Smith]

Hello Klugesmith,

That gray tube was votage multiplier unit for the tube, the tube was connected by HV cables to the multiplier.
But good point in any case. I went with self-contained unit in my last design, just as Neukyhm.

// O.C.

[Davide]

Hello,
Great project, a lot of work I bet.
This very night, actually 5 minutes ago i was able to get my first radiograpf after months of trying using that very same tube, 70-G4.
It seems something is wrong with earth and filament connection.
My does not look as good as yours but hopefully you can point me in right direction, what am I doing wrong?
I am miles from that sharp, crisp image you are getting as you can see. Any suggestions?
Thanks

//OCV

Hi Agent Smith, check out my connections:


The two black wires that goes to the multiplier are from the transformer. The small red wire that goes to the tube is the filament and I have the tube's blue wire grounded via the thicker red wire that goes out the box. The tube's black wire is the grid and you have grounded it. It's recommended to not ground the grid, but to make it a bit more negative than the filament, how? with that resistor you see in the picture.

I suggest that you use the red wire to power the filament with +2.50 to +2.75 V, ground the blue wire and connect the black one to the cathode of you HV PSU. Once you've done this, it's time to common reference everything, and you will do this with a 47kOhm resistor placed between the cathode (negative pole) of you HV PSU (where it's also connected the tube's black wire) and the tube's red wire, just like I have in the picture. This way you will have the tube connected as the datasheet says
, although they use different colors.

Also, if your HV PSU is unrectified, it's a very good idea to place one of those big 20kV 10nf capacitor between the tube's black wire and ground, it will stabilize voltage a lot and improve results. This won't help that much if your HV PSU is rectified like a TV flyback.

Note: of course, the cathode of the PSU you are using to power the filament has to be grounded too.


Edit: I hope this helps.



Edit2: typo in the capacitor, it's 20kv 10nF, like these.

The dielectric oil is very recommended.

PS: where did you get the tube?

Hi! First of all many congratulations for this project, really well done and also clearly explained. I'm trying to build a similar setup and I bought an x-ray tube (model: HBJ11) on Alibaba, it is a 70kv 0.4mm focal spot tube, similar to yours. However I'm struggling to understand how to properly wire it.  I saw that you connect a 47kohm resistor between the grid of the tube (black wire) and one of the two filament wires (blue wire). However if I look at my tube, it seems like the grid and one of the filament legs are already connected with a wire inside the tube (see attached picture). Do you think that I should wire anyway my tube the same as yours?
I checked the specs sheet and it only says to "observe the connection diagram and the grid resistor value", but it doesn't say anything else about this resistor  Sorry about the long question and thank you in advance. 

[paulj]

Hello, for amateur radiography, take the two terminals of the filament (by testing the resistance) and connect one to the ground of your high voltage power supply.

don't forget that even limited in power, radiation is very dangerous.

My 40kv tube produces 13 Sv / h

[Davide]

Hello, for amateur radiography, take the two terminals of the filament (by testing the resistance) and connect one to the ground of your high voltage power supply.

don't forget that even limited in power, radiation is very dangerous.

My 40kv tube produces 13 Sv / h

Hi! Thank you for your answer  I only have one concern about this: if I connect it this way, can I power the filament directly with a DC-DC step down converter (current limited) or will this connection damage the DC-DC converter? I'm asking this because in this way the ground of HV part of the circuit will be directly connected with the DC-DC converter.

Yes I'm aware of the danger involved with radiation, I have many lead panels that I will use as shelding, and a Geiger counter to measure radiation. Thanks for your advice 

[neukyhm]

Hello, for amateur radiography, take the two terminals of the filament (by testing the resistance) and connect one to the ground of your high voltage power supply.

don't forget that even limited in power, radiation is very dangerous.

My 40kv tube produces 13 Sv / h

Hi! Thank you for your answer  I only have one concern about this: if I connect it this way, can I power the filament directly with a DC-DC step down converter (current limited) or will this connection damage the DC-DC converter? I'm asking this because in this way the ground of HV part of the circuit will be directly connected with the DC-DC converter.

Yes I'm aware of the danger involved with radiation, I have many lead panels that I will use as shelding, and a Geiger counter to measure radiation. Thanks for your advice 

I used a stepdown converter from ebay to power the filament, no problem, as long as you care for grounding.