FAQ - HV diodes selection and testing

If you have a question about this topic, the answer is probably in here!
Post Reply
User avatar
Richard Hull
Site Admin
Posts: 10873
Joined: Fri Jun 15, 2001 1:44 pm
Real name: Richard Hull

FAQ - HV diodes selection and testing

Post by Richard Hull » Sat Jul 05, 2014 4:39 pm

If you roll your own supplies, you will need suitable HV diodes. All such diodes must be of the silicon type. Much older surplus selenium diode stacks are no good, in general. Avoid them like the plague. They can be identified due to their being most often found in a glass or phenolic tube about 3 to 6 inches long. The tube is very slinder. I have included an update post below in this thread with images of both selenium and slicon HV diodes.

HV silicon diodes are almost always black in color and in a molded plastic or epoxy cylindrical or square body with a cathode band clearly placed at one end encircling the entire body of the diode. They are almost always over 1 inch in length and can be up to 1 foot long at extreme voltages. In general, the longer the body, the higher the voltage.

Modern silicon diodes that are suitable come in two varieties.....slow and fast recovery. 99% of all HV silicons you will encounter are slow recovery designed for frequencies of the power line, 60 hz. They can be used up to about 5khz, however, with relative ease. Such diodes are relatively plentiful, easily obtained and inexpensive until you hit 20kv PIV ratings that can handle high currents above 100ma.

The typical recovery time for such diodes are in the 1us to 500ns recovery time range

The amateur fusioneer should demand diodes of at least 100ma current capability. More would be better. Among the cheapest and finest of the slow recovery diodes are the microwave oven diodes. These are typically in the 10-12kv range and can handle 250 ma or more. I have bought them at hamfests for $3.00 each, but the list price at an appliance store can be $20.00 each. The more normal street price for new diodes of this type are $6.00 or a little more. (e-bay) These diodes can be series connected for higher voltages when used at line frequencies, provided you don't go over three in series where special circuitry might be needed. I would limit simple series connection to two diodes.

The best HV diodes for slow recovery are made by Semitech and cost a small fortune new. Thus far 100% of my Semitech finds have been surplus at hamfests. They show up on E-bay on occasion.

HF Diodes

High frequency or fast recovery diodes, at high voltages, are hyper expensive and are rarely found surplus. A UF4007 is cheap but will only handle 1kv at 1 amp. Once the piv voltage on a fast diode exceeds 5kv you are on a special order at list price regime. Surplus finds are extremely rare.

Typical recovery times are all well below 100ns. If you don't have a spec sheet for a specific number diode in your hand, assume it is slow recovery.

These diodes are what you will need if you are designing or building a modern switcher type supply that runs at 20khz or more. Such HV supplies are rare in the fusor voltage and current regime and even rarer in a negative hot supply, (grounded positive), demanded by the fusor. Such complete supplies of this rare type have been found on E-bay, but if complete and working would cost many hundreds, surplus and many thousands of dollars brand new.

In general, all modern HV supplies are manufactured as switching supplies. They are lighter, cheaper, easier to control and construct at the OEM level.

High frequeny diodes of PIV ratings over 20kv can cost $50.00 each or more. Thus, even the manufacturers tend to achieve voltages over 30kv with multiplier strings of lower voltage, fast recovery diodes. These diodes work very well in multipliers and far lower capacitor values can be used for higher current ratings.

Modern microwave ovens all use switcher supplies and have fast recovery HV diodes in them and if you can rob them from newer defective microwaves you are in good shape if a swicther supply is your game. In general, these are of a much lower PIV rating (5-7.5kv) and cost a good deal more if purchased new.

Testing

All hv diodes will NOT respond to the diode test function on a modern digital multimeter, unless the diode is dead shorted.

The following test is for true HV diodes over 5kv rating

For true testing, you will need a fully variable 0-30 volt DC supply, a 1 watt 10 kohm resistor and your multimeter. Hook your test diode in series with the 10k ohm resistor feeding the DC from your supply into each end of this series connection.

Connect your DC meter across the diode. Turn on the supply and slowly advance the voltage to your highest voltage, noting the voltage on your multimeter once full voltage is reached and write it down. Now, lower the voltage to zero and reverse connect only the diode in your circuit. Reconnect the meter across the diode. Take the voltage all the way up and record this new voltage value. Your testing is done!

1. If you read a very low voltage or virtually zero volts across the diode in both diode connect directions, the diode is shorted. Throw it in the trash now!
2. If you read between 4-7 volts in one direction and nearly full supply voltage in the other then the diode is probably good.

Once you have what you think is a good diode, try and look its number up on-line or at the manufacturer's website for more info. and, hopefully, a spec or data sheet on it.

Testing for unmarked HV didoes PIV and current capability is usually destructive in nature. Unless you have 10 or 20 of the same item in hand you can't be sure of what you have even if the above testing shows they are all in good shape. If you are willing to sacrifice one or two of them to destruction then you can, with good and complete gear and skills, discover what they are rated for, but it can be dangerous and expensive to run such tests for the uninformed about electronics and the ill-equipped. As such, we will not explain further.

Now you are ready to hunt for good didoes based on your planned construction of your HV supply.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
Retired now...Doing only what I want and not what I should...every day is a saturday.

User avatar
Bob Reite
Posts: 277
Joined: Mon Aug 26, 2013 1:03 am
Real name: Bob Reite
Location: Wilkes Barre/Scranton area

Re: FAQ - HV diodes selection and testing

Post by Bob Reite » Sun Jul 06, 2014 4:30 am

A 30 volt supply will be good for testing HV diodes up to about 15 KV peak reverse voltage rating. Diodes with a higher peak voltage rating have a larger forward voltage drop and may still test as 'open' with only 30 volts. I use a 48 volt dc supply, mainly because that's what I have around. If you own an old style "capacitor checker" such as the Heatkit IT-11 or the Sprague Tel-Ohmike, you can use its leakage function to supply a higher reverse voltage. The capacitor checkers are nice because the leakage testing supply is current limited.

http://hvstuff.com has high voltage 100 nanosecond diodes for sale. A 100 mA 20KV diode goes for around $4.00 from there. However it has a rather small body, if the diode is going to be operated above 4KV it will need to be immersed in transformer oil. A rule of thumb for high voltage parts operated in air is 10,000 volts per inch between terminals of the device. So for 20KV the diode body needs to be at least two inches long for operation in air. Note that I have not bought or tested the diodes from hvstuff.com, hopefully they do live up to their specs.
The more reactive the materials, the more spectacular the failures.
The testing isn't over until the prototype is destroyed.

User avatar
Richard Hull
Site Admin
Posts: 10873
Joined: Fri Jun 15, 2001 1:44 pm
Real name: Richard Hull

Re: FAQ - HV diodes selection and testing

Post by Richard Hull » Sun Jul 06, 2014 11:22 pm

Some images of the types of diodes First the seleniums type you do not want to buy.
Then several images of various silly-con type that are OK.

Go to a lot of hamfests and cool finds like these are around.

For a higher voltage DC test supply, a full wave bridge attched to a variac would yield a 0-140 volt diode test supply but I would go to a 50k ohm series resistor and be real careful not to kill myself while using the higher voltage.

Richard Hull
Attachments
sel.JPG
selenium type HV diodes - Not good for fusor use. Too low a current
D1.JPG
Top - Semitech super 10KV 1 amp diode door knob stackable. Silec molded bar 8kv @1amp, Microwave diode 12 kv @ 250 ma. Unitrode superb small doorknob types 7.5kv @ 500ma. Also stackable
D2.JPG
Giant BBC door knobs stackable string Each 10kv @ 750ma. RCA 10kv bridge 100 ma. 20kv molded IR bar type @ 250ma
D3.JPG
one of two Semitech boards I bought that make up a 25kv, 250ma bridge. I paid $10.00 for both boards at a hamfest.
D4.JPG
Close up of Semitch diode on the boards above.
D5.JPG
Giant 14 inch long pair of IR diodes to make a 30KV 250MA bridge. Paid $25.00 for the pair. (hamfest)
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
Retired now...Doing only what I want and not what I should...every day is a saturday.

Post Reply