Diodes Help

[daniel_pflug]

Hey, I recently acquired a 70kV dental xrt and was going to craft a homemade supply from it. I've the necessary ballast resistors and voltage/current resistors set and a massive 20A variac ready to go, but am no good with the rectification process. I've seen many successful fusors with strings of microwave oven diodes, but didn't know if these would suffice at these voltages. I am aware that the full potential can only be reached using a pulse method, but I figure at 75% or less on the variac and lots of mineral oil I should be able to supply continuous current.

Basically what I'm asking is for a basic schematic for diode placement for the rectification as well as how to implement it into the HV line. Also, can full wave rectification be achieved using both secondary windings of this center-tapped transformer? Thanks for the help!

**I'll show what I've already started drawing up when I get to my computer at home, but it's not pretty haha

[Jeroen Vriesman]

Putting diodes in series is not always a good idea.

For example, some people use a string of cheap 1N4007 (1kV/1A diode), thinking they can use it for N x 1kV. That is not true, the reverse voltage will not be evenly distributed over the diodes, so the diode with the lowest reverse leakage current (very variable and temperature dependent) will be destroyed, it will probably zener for a moment and the next diode will also die.

The way to fix that is use special diodes which are made to put in series, these are diodes which will zener somewhere above their rated voltage. for example the DSA2-16A, the rated reverse voltage is 1600V, but it will start to zener around 1750V (2mA max). This will limit the maximum voltage over one diode in a chain.

Usualy "high voltage diodes" (> 2kV) are made of diodes in series which will zener like the DSA2-16A to limit the reverse voltage over one diode, the number of diodes can be estimated by 2 * (forward voltage at low current)

So you can put high voltage diodes in series if they are a chain of diodes themselves and are exactly the same type.

[daniel_pflug]

So theoretically could I rectify each of the secondaries through a 50kv 1A bridge setup then combine the dc lines out?

[Jeroen Vriesman]

It is center-tapped, so the center goes to ground, you don't need a bridge for full-wave rectification with a center tapped transformer, just two diodes which can handle the voltage, see FAQ.

Or here: http://www.eng.uwi.tt/depts/elec/staff/ ... fwrc1.html (reverse diodes for negative supply, and make sure the center tap is connected to ground).

[charlie_mccartney]

Go to a used appliance store and pickup two broken microwaves and use the diodes from there and capacitors if you choose.

[Rich Feldman]

Not so fast, Charlie.
He's asking about an x-ray transformer, not a NST. Higher voltage by a factor of at least four.
And as I explained here, viewtopic.php?f=6&t=8720#p59997
even the canonical NST fusor circuit exceeds the reverse voltage rating for typical MWO diodes
(when there's no plasma loading it down). Many demo fusioneers can attest to successful operation anyway.

[Richard Hull]

Series diodes work great on 60hz xfrmrs. Two or three microwave diodes in series will handle most apps. For high frequency apps, you will need faster diodes in most cases and these are rather special and costly. Semtec is one of the better high voltage diodes out there and I use 3 of their 20kv 500ma silly-con diodes in series on each leg of my 60 kv ct, 60hz, x-ray transfomrer, but 5 microwave diodes would have worked just fine in each leg. No blowouts under oil since 2004 on fusor IV.

Richard Hull

[charlie_mccartney]

Sorry there Rich my ADD got the best of me. Just glad he didn't take my advice.

[Rich Feldman]

No problem Charlie, we all get that way from time to time.

... He's asking about an x-ray transformer, not a NST. Higher voltage by a factor of at least four.

Oops, should have said "higher voltage by a factor of at least three".
That's because the nominal kV value of an XRT is the secondary PEAK voltage.
The 70 kVp unit in OP would be called 50 kV (RMS) if specified for non-x-ray use.

Daniel:
Your XRT probably has two 25 kV (RMS) windings.
Normally the "start" end of each winding is tied to GND, perhaps with a mA meter in series on one side.
There are, of course, other useful ways to connect them.
But don't assume that there's 35 kVp worth of insulation between that end and the core.

[Johannes westman]

Im currently bidding for a 50,000V 8mA High Voltage X-ray Tube Power Supply - would HVM12 diodes in series work for that? I'm specifically thinking about what Jeroen Vriesman mentioned about reverse voltage.

[Chris Bradley]

Just buy a few 40 or 20kV diodes from the HV store guy on ebay. Easy to get this stuff now. No need for long strings of diodes.

In the days when we struggled to get this stuff, ganging up strings of diodes was just a consequence of a limited supply versus an experimental ambition! These days, your every desire is catered for!!

[Johannes westman]

would love a link to his store. I've tried searching, but I cant find it, and I don't know how to search for stores specifically.

[Chris Bradley]

http://www.ebay.co.uk/usr/high-voltage-hv

[Jeroen Vriesman]

If you look at the datasheet of the HVM12, you will see a forward voltage of 14V at 350mA.

That means that it is 28 diodes in series, as explaned before, these are zener diodes.

Take a look at figure 3 in the datasheet "Typical reverse characteristic", you can see that the zener voltage is just above the rated voltage (is has to be, the 28 diodes will differ a bit), you can see that the zener voltage is about 110% of the rated voltage.

So the HVM12 is just 28x 470V (12E3 * 1.1 / 28) zener diode.

To answer your question: yes you can put Nx HVM12 in series and this will give you a maximum reverse voltage of Nx12kV, becasue the HVM12 is a 13.2kV zener diode, it would not work with normal diodes

It would be a good idea to use HVM12's of the same manufacturer and age.

One extra note: is you rectify an AC voltage of X volts rms with one diode string, you might think that the max reverse voltage is sqrt(2) * X, this is true, but if you use a capacitor (even a cable is a small capacitor) the reverse voltage becomes 2 * sqrt(2) * X volts during the phase when the diode is blocking the current (voltage between peak AC and charged capacitor).

When the capacitance is small, the zener current will prevent this, but the datasheet does not give any info about the maximum zener current (it only states the maximum current at the rated reverse voltage: 5 uA, but that is not the maximum allowed reverse current).
The maximum zener current is limited by the amount of power the diode can dissipate, 350mA @14V = 5 W, so for example, if your forward current is 35mA avg (70mA half the time). there is 4.5W left to dissipate. At 13.2 kV that is 0.34 mA avg. ( 0.7 mA for one half of the AC period).