EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: Neukyhm on September 08, 2018, 12:11:15 pm
-
Hi there, I'm building an x-ray machine, I have built the main parts like the HV power supply and the box containing the voltage multiplier and the tube, covered with oil.
Everything works and I can make radiographs, however the machine does not work as expected. It should have around 100W but it only draws ~40W from the PSU. At first I thought it had something to do with the grid resistor of the tube. This is the datasheet of the tube (http://www.cei-xray.it/content/uploads/prodotto_image/OX70_G4.pdf), I was adviced by the manufacturer to use 47K as grid resistor, OK. If you look at the filament current and anode current vs filament current, you see that if I apply 2.5V to the filament I have around 2A at the filament and let's say almost 3mA at anode when voltage is 65kV, so power is big.
To my surprise, today I measured the current at the filament, at the moment I apply 2.5 to it, current rises to >2A as expected by the datasheet but then it decreases instantly to 1.5A, I now think that's what makes the machine work at 40W and not at 100W as I expected.
Do you guys have any idea why this happens? If the datasheet doesn't lie, current at filament should be 2A and not 1.5A. You may suggest to increase the filament voltage but again I was adviced by the manufacturer to not apply more that 2.5V.
-
Modern X-rax heads provide limits for filament current/voltage and control the filament current in order to stabilize the anode current in a control loop. Have you got provisions in your design to measure total tube voltage and current?
For an example how a (fairly) modern X-ray had is constructed, you may want to have a look at this thread: https://www.eevblog.com/forum/projects/x-ray-generation-and-detection/msg1507648/#msg1507648 (https://www.eevblog.com/forum/projects/x-ray-generation-and-detection/msg1507648/#msg1507648)
-
To my surprise, today I measured the current at the filament, at the moment I apply 2.5 to it, current rises to >2A as expected by the datasheet but then it decreases instantly to 1.5A, I now think that's what makes the machine work at 40W and not at 100W as I expected.
Do you guys have any idea why this happens? If the datasheet doesn't lie, current at filament should be 2A and not 1.5A. You may suggest to increase the filament voltage but again I was adviced by the manufacturer to not apply more that 2.5V.
Cold filaments have lower resistance, so resistance increases, and current decreases as it warms up.
-
Do you guys have any idea why this happens? If the datasheet doesn't lie, current at filament should be 2A and not 1.5A. You may suggest to increase the filament voltage but again I was adviced by the manufacturer to not apply more that 2.5V.
The emission characteristic curves on the datasheet don't start until 2.75v and the permissible voltage range is shown as 2.5-5v. I think you are under driving the filament.
-
Do you guys have any idea why this happens? If the datasheet doesn't lie, current at filament should be 2A and not 1.5A. You may suggest to increase the filament voltage but again I was adviced by the manufacturer to not apply more that 2.5V.
The emission characteristic curves on the datasheet don't start until 2.75v and the permissible voltage range is shown as 2.5-5v. I think you are under driving the filament.
I know, but then why did the manufacturer say that I should not apply more that 2.5V? I think I'm going to contact them again.
Also if you take a look to emission vs filament voltage you get an idea of the anode current when filament voltage is 2.5V, much higher than my 0.6mA :-//
-
Do you guys have any idea why this happens? If the datasheet doesn't lie, current at filament should be 2A and not 1.5A. You may suggest to increase the filament voltage but again I was adviced by the manufacturer to not apply more that 2.5V.
The emission characteristic curves on the datasheet don't start until 2.75v and the permissible voltage range is shown as 2.5-5v. I think you are under driving the filament.
I know, but then why did the manufacturer say that I should not apply more that 2.5V? I think I'm going to contact them again.
Also if you take a look to emission vs filament voltage you get an idea of the anode current when filament voltage is 2.5V, much higher than my 0.6mA :-//
Maybe they are a bit old-school and actually mean a nominally-2.5v-under-max-load secondary winding on a transformer, which would put out more under the reduced loading of a hot filament.
They probably ought to be specifying the max filament drive as a current rather than a voltage.
-
Do you measure the anode voltage, and are you absolutely certain that it is 65 kV? For a given filament current, the emission current will be lower with a lower anode voltage, mostly due to space charge effects. The datasheet you linked only gives emission curves at 65 kV, here's a datasheet for a similar Toshiba tube with more comprehensive information: https://etd.canon/eng/product/read_binary.php?cid=100200100100&pid=20140730_1058&fn=PE-D-045_2017-06.pdf
-
Do you measure the anode voltage, and are you absolutely certain that it is 65 kV? For a given filament current, the emission current will be lower with a lower anode voltage, mostly due to space charge effects. The datasheet you linked only gives emission curves at 65 kV, here's a datasheet for a similar Toshiba tube with more comprehensive information: https://etd.canon/eng/product/read_binary.php?cid=100200100100&pid=20140730_1058&fn=PE-D-045_2017-06.pdf
I don't have the equipment to measure such a high voltage, however I have an oscilloscope and I measured the voltage at primary, so I know the voltage at secondary because I made the transformer and I know the turns ratio, here the voltage is around 5.5kV, then it goes to a 12-stage voltage multiplier that should give up to 66kV.
-
The anode voltage will sag heavily with load, even an old style iron transformer has a substantial impedance due to the very long thin windings of the secondary. If you're using a voltage multiplier with a high frequency drive then you'll get even more sag. You need to be able to measure the anode voltage and adjust the drive according to the filament voltage. Also you need to be careful not to energize the anode without powering the filament or the voltage across the tube can rise to dangerous levels.
Here's some info on a Kodak inverter type dental xray head I reverse engineered several years ago:
https://www.repairfaq.org/sam/xraysys.htm#xraytro (https://www.repairfaq.org/sam/xraysys.htm#xraytro)
-
The anode voltage will sag heavily with load, even an old style iron transformer has a substantial impedance due to the very long thin windings of the secondary. If you're using a voltage multiplier with a high frequency drive then you'll get even more sag. You need to be able to measure the anode voltage and adjust the drive according to the filament voltage. Also you need to be careful not to energize the anode without powering the filament or the voltage across the tube can rise to dangerous levels.
Here's some info on a Kodak inverter type dental xray head I reverse engineered several years ago:
https://www.repairfaq.org/sam/xraysys.htm#xraytro (https://www.repairfaq.org/sam/xraysys.htm#xraytro)
Can you please explain how voltage decreases more with a high freq transformer? I thought that a high freq driver for the tube would prevent the voltage from decreasing, mainly because as the tube discharges the multiplier's capacitors, you are recharging them fast. My transformer is a big ETD59 made of ferrite, and frequency is 110kHz (high in my opinion)
I have contacted the manufacturer and I'm waiting for the response, I will post here any useful information they give me.
-
Voltage multipliers store energy, so their final output voltage varies depending on load and the input frequency. One way to increase their output voltage is to increase the frequency of the input.
Can you please explain how voltage decreases more with a high freq transformer?
I believe what is meant is that at higher frequencies, more power is lost in the magnetic loop of the transformer. You are correct that a higher frequency reduces the sag in the voltage multiplier, but the transformer has inductance. It seems like you can't win!
-
The tube manufacture is probably not going to be able to help you, it's obvious that your anode voltage is sagging, fixing that is your responsibility. Your frequency is plenty high but the supply is still going to present a substantial source impedance. You *must* measure the anode voltage, there is no way around this. If you are going to be playing with an xray tube you need to know the anode voltage, not only to get the tube operating correctly but for safety reasons. As the voltage increases, the penetration and scatter of the radiation increases. Did you look at the Kodak schematic I linked to? It's not difficult to build a voltage multiplier, you can use a load of 1M resistors in series.
-
I believe what is meant is that at higher frequencies, more power is lost in the magnetic loop of the transformer. You are correct that a higher frequency reduces the sag in the voltage multiplier, but the transformer has inductance. It seems like you can't win!
What I meant is that using a voltage multiplier (which is typically only used with a high frequency transformer) is going to result in greater sag. Voltage multipliers are very saggy, that's just the way they are, at least when you have multiple stages with reasonably sized capacitors.
-
Your SMPS might have poor regulation as the tube anode current comes up and then filament voltage sags. Sub in a known resistor (2V/1.75A graph) say 1-2 ohms and see if you have enough power from the PSU. Add a 1/2 turn to the secondary. Poor (filament) connections can also do it, they can heat up and a thermal IR camera pic can show these.
This is one ETD59 with two secondary (HV, filament) windings? If so, 110kHz is high so your measurements are beyond multimeters and hard to be accurate at HF.
The x-ray tube is bit strange, it's has a "grid" that looks like an acceleration (screen) grid.
The grid resistor graph shows lines at 35k and 40k with "Rg-3k" and "Rg+3k" labels added.
So 47k ohms seems a bit high. It affects the focus and the conductance of the tube.
But the low filament voltage I think is the problem.
-
This is a test I have done today, you all know what it is.
Power consumption: 33W
Camera: Nikon D300
ISO: auto
Exposure: 0.5 seconds.
The thing is that, despite the "low" power (because I was expecting more), it's working very good.
-
The image looks pretty muddy but that might be due to the camera photographing the dim phosphor screen. It's also pretty clear that your anode voltage is low given the low penetration, you can't even see through the IC package. Get the power up to the proper level and you should get a much better image.
-
The image looks pretty muddy
I had to resize to upload it here, the original seems better.
It's also pretty clear that your anode voltage is low given the low penetration, you can't even see through the IC package.
Well keep in mind that it's just 0.5s of exposure.
Get the power up to the proper level.
That's why I'm here lol.
-
What's the capacitor value that you are using in your multiplier? Can you confirm that your voltage multiplier is 12 stages (i.e. 12 stages of two capacitors and two diodes each)? A twelve stage multiplier will be very saggy under load, and I can guarantee that the output voltage is much lower than the theoretical. The transformer output voltage can also not be calculated simply by the ratio, due to leakage inductance and parasitic secondary capacitance. Assumptions are as good as the understanding of the system, so if you want to estimate the voltage, you have to account for your drive circuit behavior, transformer parasitics, voltage multiplier sag and the operating characteristics of your x-ray tube. Just measuring the voltage is usually significantly easier.
Results look good so far.
-
Interested to see pics of the whole setup. TIA.
-
All the capacitors in the multiplier are 10nF. I have simulated the whole machine in Spice (xray tube as biggg resistor) and everything is fine & the tube works at 120W.
This setup is very temporary as I'm planning to mount everything in a 3d printed housing.
Edit:
Can you confirm that your voltage multiplier is 12 stages (i.e. 12 stages of two capacitors and two diodes each)? A twelve stage multiplier will be very saggy under load, and I can guarantee that the output voltage is much lower than the theoretical.
The multiplier has 12 capacitors and 12 diodes, that's a 12 stage multiplier as far as I understand (am I wrong? :-X)
-
I don't see much lead shielding around that box....
-
May we assume that you operate the apparatus from a remote room and you've got a geiger counter to make sure your own exposure is minimized? And also that noone else is in the device's vicinity when you operate it (any pets?). Really looks like shielding is NIL.
-
May we assume that you operate the apparatus from a remote room and you've got a geiger counter to make sure your own exposure is minimized? And also that noone else is in the device's vicinity when you operate it (any pets?). Really looks like shielding is NIL.
I understand the concerns. There are no neighbors around here, I have a Geiger counter and I'm operating the machine with a relay and a long cable, far from it, in another room.
-
May we assume that you operate the apparatus from a remote room and you've got a geiger counter to make sure your own exposure is minimized? And also that noone else is in the device's vicinity when you operate it (any pets?). Really looks like shielding is NIL.
Lets hope so.
His set-up probably isn't producing very energetic x-rays but, even so, I would not want to hang around it without shielding and with no dosimetry.
I'm almost as amazed that he has a CW multiplier with a nominal 65kV output built on perfboard.
-
The less energetic (or "soft") X-rays are even more dangerous since they are less penetrating and are absorbed by the soft tissue of the body more or less completely. Hence they do more damage to the body than "hard" or higher energy X-rays. That's also the reason that most of the X-ray heads for general medical use include a filter of 0.5...2mm of aluminium to block the low-energy portion of the X-ray spectrum. Things are different in a mammography machine and that's also the reason that these kind of examinations are considered more risky.
When expeimenting with X-rays, you've simply got to know what you're doing. But it seems that @Neukyhm has taken appropriate precautions.
-
Neat "setup" plan to make one as well,i`m lucky as i have alot of lead plates lying about.
-
Hi everyone, I have no reply from the manufacturer (cries), but I have tried to increase the filament current by just 0.20V more, so now it's 2.7V (it was 2.5V before). Also I reduced the ISO to 200 and 1 second exposure and now the pictures are much better. I also noticed that the tube draws more current from my power supply.
-
I encourage you to measure the HV, as I've build multipliers with those same caps/diodes and found they don't perform well.
The capacitors are out-of-the-box -15% at around 8.5nF and even less with DC bias on them. The diodes I have not tested but they are designed to be encapsulated (8, 10mm vs 15, 20mm) and I had them arcing across the body when I pushed it. I measured their forward voltage as much less than the datasheets, so the stack seems small.
My point is, I would make measurements to confirm the HV stays up under load, instead of upping the filament voltage to get good anode current.
Less X-rays generated if the HV sags.
-
I encourage you to measure the HV, as I've build multipliers with those same caps/diodes and found they don't perform well.
The capacitors are out-of-the-box -15% at around 8.5nF and even less with DC bias on them. The diodes I have not tested but they are designed to be encapsulated (8, 10mm vs 15, 20mm) and I had them arcing across the body when I pushed it. I measured their forward voltage as much less than the datasheets, so the stack seems small.
My point is, I would make measurements to confirm the HV stays up under load, instead of upping the filament voltage to get good anode current.
Less X-rays generated if the HV sags.
If I'm lucky I will be able to measure the spectrum of the x-rays at my college :P
That way I can know the voltage.
-
Why don't you just measure the anode voltage? You don't need anything particularly exotic to do so.