Question about using equalizing resistors to protect rectifier diodes in power supply

Thread Starter

harryec

Joined May 5, 2018
6
I am reading the excellent Radio Theory Handbook Beginner to Advanced by Ron Bertrand and I've come up on something that I just can not understand. I was hoping that someone might be able to provide some guidance. He talks about how in high voltage power supply circuits that rectifier diodes are often connected in series to increase their overall PIV. The suggestion is that equalizing resisters are needed to force even voltage distribution across the diodes. These resisters are parallel to the diodes. My question is... doesn't that defeat the purpose of the diodes? I see diodes like one way valves and these resisters seem to be bypassing the valves allowing current to flow the wrong way. What am I missing??
 

ebeowulf17

Joined Aug 12, 2014
3,274
I am reading the excellent Radio Theory Handbook Beginner to Advanced by Ron Bertrand and I've come up on something that I just can not understand. I was hoping that someone might be able to provide some guidance. He talks about how in high voltage power supply circuits that rectifier diodes are often connected in series to increase their overall PIV. The suggestion is that equalizing resisters are needed to force even voltage distribution across the diodes. These resisters are parallel to the diodes. My question is... doesn't that defeat the purpose of the diodes? I see diodes like one way valves and these resisters seem to be bypassing the valves allowing current to flow the wrong way. What am I missing??
Interesting, I would've thought that made sense if you reversed series and parallel, maybe?

***EDIT: No, I was thinking of power sharing, not reverse voltage. Ignore my first comment.

Can you share a schematic? It's so much easier to understand circuits we can see, not just verbal descriptions.
 

crutschow

Joined Mar 14, 2008
25,388
The resistors are to equalize the reverse voltage across them, otherwise the one with the lowest reverse leakage would try to block all the voltage, which could cause it to go into reverse breakdown.

The resistors do conduct in the reverse direction, of course, but their value can be made high enough (likely a megohm or more) that the reverse current is generally not that significant in the application.
The resistor just has to be sized to carry the diode maximum reverse leakage current at its rated voltage at its highest operating junction temperature.
 

Thread Starter

harryec

Joined May 5, 2018
6
The resistors are to equalize the reverse voltage across them, otherwise the one with the lowest reverse leakage would try to block all the voltage, which could cause it to go into reverse breakdown.

The resistors do conduct in the reverse direction, of course, but their value can be made high enough (likely a megohm or more) that the reverse current is generally not that significant in the application.
The resistor just has to be sized to carry the diode maximum reverse leakage current at its rated voltage at its highest operating junction temperature.
Right.... I hadn't really calculated how much current would be leaking back... not much. Now it makes perfect sense... thank you!!
 

Thread Starter

harryec

Joined May 5, 2018
6
Interesting, I would've thought that made sense if you reversed series and parallel, maybe?

***EDIT: No, I was thinking of power sharing, not reverse voltage. Ignore my first comment.

Can you share a schematic? It's so much easier to understand circuits we can see, not just verbal descriptions.

I probably shouldn't do that since its in this guys book and I don't have permission to copy. I think I got my answer. The resisters are like 470K each so total reverse resistance is fairly large. I would leak current back... but not much due to high resistance. I was just hitting me wrong... but when you think about it ... its just a trade off... and it protects the diodes.

Thanks for your kind response to my post!
 

Bordodynov

Joined May 20, 2015
2,658
And I have my own, alternative opinion on this matter. I think that when using modern silicon diodes there is no need to use parallel resistors. Take for example high-voltage rectifier diodes, which contain dozens of diodes in their design. I strongly doubt that these diodes are shunted by resistors. In any case, their spice models do not contain resistors. High-voltage diodes have a not very sharp breakdown and when approaching the breakdown mode the current gradually increases. This effect redistributes the voltage. Moreover, the currents of all diodes are equal.
See
Draft19.png
 

Attachments

Bordodynov

Joined May 20, 2015
2,658
Such an effect (degradation) will be observed in the case of an uneven distribution of the electric field. I observed this effect in the emitter junction (voltage is 7 volts). The emitter had a striped (rectangular) structure. I observed a glow at one of the corners of a rectangular emitter in the breakdown. Relatively high-voltage diodes can have in their structure special security rings, which equalize the electric field. I think that the habit of putting parallel resistors appeared at the dawn of microelectronics and for germanium diodes. Then it could be actual.Such an effect (degradation) will be observed in the case of an uneven distribution of the electric field. I observed this effect in the emitter junction (voltage is 7 volts). The emitter had a striped (rectangular) structure. I observed a glow at one of the corners of a rectangular emitter in the breakdown. I observed the localization of the breakdown in a microscope (plasma glow).
Relatively high-voltage diodes can have in their structure special security rings, which equalize the electric field. I think that the habit of putting parallel resistors appeared at the dawn of microelectronics and for germanium diodes. Then it could be actual. By the way, I analyzed the selenium rectifiers (diode bridges).They used many elementary diodes connected in series with each other.There were no resistors.Each diode was a thin square plate.
 
Last edited:

Bordodynov

Joined May 20, 2015
2,658
http://enciklopediya-tehniki.ru/promyshlennost-na-v/vypryamitelnyy-stolb.html

The most common are silicon rectifier poles. They do not need to shunt individual diodes, since the current-voltage characteristic of a silicon rectifier diode does not have a falling pre-breakdown section and a uniform distribution of the reverse voltage is achieved automatically. The merits of silicon rectifiers include the fact that silicon rectifier diodes allow for higher reverse voltages than germanium ones.
It was a quote.It mentions that the germanium diode has a falling section of the breakdown (the Losev site).Therefore shunt resistors are necessary for germanium diodes.I came to the conclusion - For some high-voltage diode, diodes with a working voltage of 400V are used. The necessary voltage is determined by the number of such diodes connected in series.
 
Last edited:

Bordodynov

Joined May 20, 2015
2,658
But continuous reverse breakdown of the diode junction could degrade it's reliability.
Leveling resistors do not increase reliability! The weakest diode will be the first to break through. The resistor does not limit the diode current There is another example - the rectifier bridge. There are two pairs of diodes included in the serial circuit. Why, then, in this case, do not put aligning chains. If the breakdown is reversible, and for the silicon the diode is so until it overheats. For capacitors, this is justified because the breakdown in them is irreversible Example of a high-voltage diode BY724. Has a leakage of 3uA at 24000 volts and at a temperature of 120 degrees Celsius. RLeak = 24000V / 3uA = 8Gig_Ohm!? What leveling resistors can we talk about! A similar question was discussed at one Russian forum. I certainly put in my five cents. To which he received the answer: We put aligning resistors in the equipment parallel to the diodes in the 70's - 80's of the last century, and we will continue to install these resistors. I believe that this is conservatism. And people, when they once read a textbook in their youth, and it stuck in their heads forever. The most interesting thing is that on the Internet formulas for the calculation of these equalizing resistors are often encountered.
 

BR-549

Joined Sep 22, 2013
4,938
If we could clone PN junctions.......we wouldn't need such things. The closest we get now....is to use a common substrate. A little closer to it.

I here there are serious attempts to manufacture a constant/standard junction.
 

Bordodynov

Joined May 20, 2015
2,658
I'm 99.9% sure that high-voltage diodes have separate substrates for each elementary diode. Although all of them (substrates) can be on a common, insulating plate.
 
Last edited:

crutschow

Joined Mar 14, 2008
25,388
The resistor does not limit the diode current
Never said it did.
It's there to equalize the reverse voltage across the diodes.
There is another example - the rectifier bridge. There are two pairs of diodes included in the serial circuit. Why, then, in this case, do not put aligning chains.
You don't need them in a bridge because normally the reverse rating of each of the diodes is selected to be greater than the peak reverse voltage applied to the bridge, thus making it impossible for the peak reverse voltage rating of the diode to be exceeded.
 

Bordodynov

Joined May 20, 2015
2,658
Never said it did.
It's there to equalize the reverse voltage across the diodes.
You don't need them in a bridge because normally the reverse rating of each of the diodes is selected to be greater than the peak reverse voltage applied to the bridge, thus making it impossible for the peak reverse voltage rating of the diode to be exceeded.
When I think about the reliability of the device, then I mean reliability in emergency (emergency cases). After all, in a household outlet because of a thunderstorm there can be overvoltages of several kilovolts. If there are no additional protective elements, then the overvoltage will be on the rectifier bridge diodes. I swami agree that the diodes are selected with a margin. But can not this be done in a consistent chain of diodes. In the example (link), the author used 35 diodes with a working voltage of 1000 volts, although he needed "only" 30,000 volts.
 

crutschow

Joined Mar 14, 2008
25,388
When I think about the reliability of the device, then I mean reliability in emergency
Well that's quite different from the usual meaning of reliability in an electronic system.
That usually refers to the mean-time-between-failure (MTBF) when the system is being used within its rated limits.
 

ebp

Joined Feb 8, 2018
2,332
My impression is that the use of resistors to shunt and equalize leakage current has been maintained as a "good thing" almost exclusively in the amateur radio domain, and it dates from the days when semiconductor diodes were inferior to modern diodes and choices for off-the-shelf high(ish) voltage parts were limited. I have certainly seen nothing to suggest any such thing is done internally in true high-voltage rectifiers that are "stacks" of diodes (of course the manufacturers would be able to assure the individual diodes came from the same batch, which may be advantageous)

look for this topic at
http://www.voltagemultipliers.com/html/FAQs/FAQ_diodes.html#Parallel Resistors and Capacitors

One issue with using equalizing resistors is that a resistor suitable to put across a modern diode rated at up to a few amps is likely to cost more than the diode. Most ordinary low-power leaded resistors aren't rated for more than a few hundred volts. One of the numerous Vishay divisions will happily sell you high voltage resistors, but you may be left sad at the price.
 

ronsoy2

Joined Sep 25, 2013
71
The highest voltage "single junction" silicon diodes are in the 1200 to 1500 volt range. This is because of problems in the diode with arc over. All diodes (rectifiers) of higher voltage use more than one junction stacked in series. These individual junctions are cut from the same wafer for matching.
Next, for high voltage applications where multiple rectifiers will be put in series, rectifiers called "avalanche specified" are used. These diodes are rated for repetitive avalanche in the reverse direction of a certain amount of current. This current is vastly higher than the leakage current of the diodes so it instantly equalizes the voltages across the string.
Note that all rectifiers are not specified for repetitive avalanche, and if not could fail if subject to reverse avalanche. Thus, with avalanche diodes no equalizing resistors are necessary.
Another consideration is frequency. At low frequencies (<20khz) resistors will more or less be satisfactory for equalizing non-avalanche diodes (1N4007,etc) But at higher frequencies the capacitance and reverse recovery currents of the diodes will overwhelm any reasonable attempt at resistor equalizing.
Equalizing resistors are not satisfactory at all for square wave voltages applied to the diodes. Reverse recovery will determine the matching of the rectifiers in that condition.
 
Top