Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Steve Conner, Wed Aug 20 2014, 08:47AM

Sounds like a job for silicon carbide Schottkies.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Dragon64, Wed Aug 20 2014, 05:28PM


You'll probably need to hook them in series as most schottky are 1200V and lower.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Ash Small, Wed Aug 20 2014, 06:44PM

They have some 20kV, 25mA, 100nS reverse recovery ones here. How fast do you need?



They have faster, lower voltage diodes too, ie: 15kV, 100mA, 30nS reverse recovery.

Don't expect E-bay prices, though

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
teravolt, Wed Aug 20 2014, 10:16PM

what type of high voltage are you rectifying flyback, tube output, etc. maybe they don't have to be so fast.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
tarakan2, Thu Aug 21 2014, 05:57AM

I am experimenting with an Avramenko Fork.


Except I am using HV power supplies. This is a real, working circuit and it is interesting.


Sorry, I will make an English translation of this video soon.
I am a little busy at work.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Steve Conner, Thu Aug 21 2014, 06:58AM

This is just the same sort of "one wire" power transmission that Tesla experimented with. The circuit is completed by stray capacitance between the load and ground. (Draw a capacitor from each end of the load to ground, draw in the ground connection of the signal generator, and you have a full-wave doubler.)

If you want to build it cheaply, the EHT rectifier tubes used in old tube TV sets (with batteries for the heaters) would be ideal. Vacuum tubes have no reverse recovery charge, they will rectify almost as fast as you like. Maybe someone else can suggest some part numbers. Watch out for X-rays.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Ash Small, Thu Aug 21 2014, 10:19AM



Here's one. Can't seem to find a datasheet for it, though.

Here's the datasheet for a common one, the KY50 (U25):
]u25.pdf[/file]

Another is the KY80 (U26):

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Patrick, Thu Aug 21 2014, 09:39PM

we always see the 30nS and 100nS time intervals quoted in datasheets and we always see 60Hz and 120Hz numbers for common(standard) rectifiers for line use. But maybe we here at the 4HV should figure out how rate of rise correlates to a diode spec requirement.

HV waves tend to not reverse instantly, so what nano-second revesrse recovery is needed for heat and not popping with smoke? This way we dont have to buy excessivly spec'ed and cost componets.

is there a way to experiment or calculate this? obviously, when seriesed, it gets more complicated.


This was the purpose for my HV O-scope Probes and the years of work they took to get right.

I was/am most interested first, in square wave HV from SMPS (high power ferrite) transformers being rectified. And second, feeding CW type multipliers. so diodes become important.

(i do like the HER108, but its 1kv only.)

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Ash Small, Fri Aug 22 2014, 12:03AM



I concurr. If you have inductance, you have time before current reverses.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Patrick, Fri Aug 22 2014, 12:12AM

so Ash, if we have multiple series'ed diodes to get upto our HV, then we should have some nH in stray incductance hanging around. if we lay up a simple RLCD || TVS then we would have scalabe module that could be tested...



you could put these modules together, pretty long. there would be 3 discrete componets, the resistor, diode and TVS, the induction and capacitence would be stray, from the PCB.

edit: i meant to put the resistor in parallel, as a volteg divider with 500k or so... ugh!

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Steve Conner, Fri Aug 22 2014, 06:55AM

Diode reverse recovery is quite simple. When the diode has been conducting, the junction is left full of electrons and holes. When the voltage across it reverses, it keeps on conducting in the "wrong" direction until these charge carriers are cleared out.

That is why reverse recovery behaviour is quoted as a charge, because you have to get a fixed amount of charge cleared out to turn the diode off. You can do it as a huge spike of current, or a smaller current for a longer time.

The amount of charge is not fixed though, it depends on how much forward current the diode was conducting in the few microseconds before turn-off.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Ash Small, Fri Aug 22 2014, 10:42AM

Thanks for reminding us Steve.

I've just been reading up on Schottkys again, and apparently 1700V silicon carbide Schottkys have been available for a few years now.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Proud Mary, Fri Aug 22 2014, 12:48PM

EY51 is a super little directly-heated EHT diode rectifier, with ratings up to 500 kHz. You can find them easily enough on ebay for £5.

X-rays will start to get through the glass envelope > 14 kV (my experimental finding)

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Patrick, Fri Aug 22 2014, 10:56PM

so the reverse recovery can be treated as resistive heat loss after the blocking moment is complete?

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Ash Small, Sat Aug 23 2014, 02:30PM


I worked this all out in a thread about a year ago, at least as far as how NPN transistors work, and a diode is a transistor with the base and emitter connected.

You have to upset the equilibrium that exists after the junction is formed by removing a few valence band electrons from the collector side of the junction before electrons will start to cross the junction, and until those valence band electrons are replaced, current can flow.

Of course, in reality, they are constantly being replaced and others are being removed, but this replacement process takes time, due to the random nature of electron movement/energy, etc.

Most electrons remain in the conduction layer, rather than dropping into the valence layer, due to energy levels, etc.

So called 'hole theory' is the mathematical model that best describes it.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Steve Conner, Sat Aug 23 2014, 02:44PM

Yes and no. For the first part of the recovery, the diode looks like a dead short. If the diode is supposed to be handing over conduction to a transistor, this can cause high current and resistive losses in the transistor. This is a common problem in boost converters and bridges driving loads with a leading power factor.

What happens next depends on the diode. Modern "soft recovery" ones have a smoothly damped turn-off that causes resistive losses in the diode itself. However some older fast recovery diodes would snap off pretty much instantaneously. Gold-doped signal diodes like the 1N4148 still do this, the turn-off is so fast that they can generate energy up to hundreds of MHz. This behaviour is deliberately enhanced in the "step recovery diode" which is used as a pulse generator.

Schottky diodes have no reverse recovery charge at all. They do have junction capacitance though, which can cause similar problems.

Tube rectifiers essentially behave the same as Schottkies.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Patrick, Sat Aug 23 2014, 03:46PM

heres what i really meant before my brain stalled above...



R= 10M ohm, for voltage balancing with its seriesed peers.
C= stray, keep small.
L= stray, keep small.
TVS= breakover at 800v? (i should have drawn this as unidirectional right?, since theres another diode. )
D= HER108 (1kV, 1 amp) or spec'ed better as needed.

EDIT: i think they raised the price on the HER108 since i bought my 750 pieces. there now 14.9 cents.

Then build these out in iterations, possible in oil or epoxy, i think 600 V per module would be reliable, and resistant to blow out.

My interest in all this is for high effciency EHV switch mode power supplies using ferrite transformers at 100-300kHz. But they would require full wave rectification.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Ash Small, Sat Aug 23 2014, 05:52PM


I imagine the reason why Schottky diodes, which have junctions formed from a metal and a semi-conductor, and gold doped diodes have such short reverse recovery times is due to the metals in the junction, or more specifically, the differences in the valence band electrons associated with the metal atoms.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Proud Mary, Tue Aug 26 2014, 08:21AM

I've got loads of 1970s TV EHT diodes which I'll give away for the price of a donation to 4HV.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Steve Conner, Tue Aug 26 2014, 10:01AM

That's very kind of you Proud Mary.

Ash, Schottky diodes have no reverse recovery because they have no minority carriers. Conduction is by electrons, the same as a tube rectifier, with no holes blundering around to slow things down.

Gold-doped diodes have fast recovery because the gold atoms act as recombination centres to help the minority carriers dissipate quickly.

In modern fast-recovery diodes the same effect is achieved by deliberately disrupting the crystal lattice a little with a high-energy electron beam.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Ash Small, Wed Aug 27 2014, 02:22PM

Wikipedia says: "The majority carriers are quickly injected into the conduction band of the metal contact on the other side of the diode to become free moving electrons. Therefore no slow, random recombination of n- and p- type carriers is involved, so that this diode can cease conduction faster than an ordinary p–n rectifier diode."

The electrons associated with the metal are certainly more mobile than the valence band electrons in the semi-conductor.

Wikipedia also says:"Through various techniques (e.g., doping or gating), the semiconductor can be modified to have an excess of electrons (becoming an n-type semiconductor) or a deficiency of electrons (becoming a p-type semiconductor). In both cases, the semiconductor becomes much more conductive (the conductivity can be increased by one million fold or more). Semiconductor devices exploit this effect to shape electrical current." on the Semiconductor page.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Steve Conner, Wed Aug 27 2014, 02:32PM

That's what I said, in fact I checked the same Wikipedia article while writing my reply.

Re: Do you know any good HV diodes that perfrom well at upper KHz - lower MHz rf frequencies?
Ash Small, Wed Aug 27 2014, 03:46PM


I'm not disputing what you said Steve, I'm just trying to understand the full mechanism of what's actually happening.

It's only the majority carriers that are actually mobile, and an area that is depleted of electrons has an overal positive charge, and vice versa. I now understand the 'voltage drop' bit, which is dependant on the potential gradient across the junction. I'm also beginning to get a better picture of the covalent bonds in semi-conductors, and why they take time to occur.