|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
216.0.238.6
In Reply to: RE: Eli & Max... posted by Triode_Kingdom on July 23, 2014 at 21:16:36
Valve Wizard Merlin Blencowe seems to think caps can do the job...
Edits: 07/24/14Follow Ups:
up to 1040V using a high voltage probe. Results are shown in the attached graph. at 1040V, the difference between the highest and lowest diodes with caps across them was a MAX of 1.6V.
Pretty darn close!
I don't believe you can accurately measure the PRV seen by individual diodes using that method. As for Merlin Blencowe, he's simply incorrect on this point. Don't get me wrong; if you're happy with it and don't want to make changes, that's fine with me. Nevertheless, the need to use resistors to balance voltages across SS diodes is founded in well-established engineering going back more than four decades. I would personally never build a series string without them.
--------------------------
Buy Chinese. Bury freedom.
The newer ARRL handbook recommends against using voltage balancing resistors as they seem to cause more failures than they prevent.
According to the 2005 ARRL Handbook (page 17.9) "shunting resistors are generally not needed across diodes in series rectifier strings. In fact, shunt resistors can actually create problems because they can produce a low-impedance source of damaging current to any diode that may have reached avalanche potential."
I know how this might sound, but I no longer consider the ARRL an authoritative source on electrical theory. Walter Maxwell years ago had significant differences of opinion with ARRL staff and was practically banned from publishing under their logo. Walter was a highly regarded satellite/microwave scientist best known in the amateur ranks for his work and publications on choke baluns. I bought his last hardback book on the subject back in the mid-'90s, sold at a discount after ARRL banished him from the ranks.As someone who works frequently with RF magnetics in my day job, I can also point out several areas of ARRL-approved construction having to do with baluns and core materials that violate basic magnetic theory, and over which I have also voiced concerns. The errors have existed for more than 30 years, and like Maxwell, I had no luck ever convincing ARRL staff that any of those designs - still being published AFAIK - should be revisited.
With this as (partial) background, I hope you'll understand why I say it will take more than a quote from the ARRL handbook to convince me. I continue to use the Handbook for easy reference to formulas and the like, but I consider all else to be questionable at best. If there is another source for this new concept, I'd be interested to know more. Otherwise, well, you get the point... :)
Jack
--------------------------
Buy Chinese. Bury freedom.
Edits: 07/24/14
I don't want to let go of this without at least a cursory examination of the quotation you posted:
"shunt resistors can actually create problems because they can produce a low-impedance source of damaging current to any diode that may have reached avalanche potential."
This claim raises at least one troubling question. If a diode in the string has reached avalanche potential, in other words, its PRV has been exceeded, are the resistors really the issue of concern? Isn't this a theoretical problem that doesn't exist in properly designed circuitry? If a diode avalanches, and the current path through the resistor causes junction failure, it only means the designer failed to install a sufficient number of diodes relative to the total PRV of the string.
I am also skeptical of the author's claim that the resistors provide a low-impedance source for current flow. Values for these resistors are typically in the vicinity of 470k. This seems almost insignificant as a parallel value when considering the impedance of a diode that has avalanched.
--------------------------
Buy Chinese. Bury freedom.
Here's a discussion on the topic among some hams, with a couple references cited as well. Apparently among the ham community balancing resistors went out with drum brakes.All I can say for sure is that in 7 months time I haven't had any issues using 10nF caps for the balancing duties, in a 2500VAC bridge.
Edits: 07/24/14
I don't think that discussion contradicts what I've been saying on this point. According to one of the contributors, "if you have the equipment, check a few of your diodes and see if they will support say 10 microamps reverse ok. If not, then they should have equalizing resistors." The same person states earlier that "the diodes made today mostly don't require them. Back resistances and switching times are pretty well controlled." He goes on to say that rather than using resistors, he would add diodes to the string.This agrees completely with my previous post, where I said "The fact that you haven't experienced a failure might simply be a testament to the consistency of the devices, plus the fact each diode in the string is overrated."
The bottom line here - in my opinion - is that the balancing resistors provide a safety mechanism that isn't present otherwise. If the particular diodes are well-matched (and impervious to small reverse currents) such that the resistors aren't required, so be it. At that point, it simply becomes a matter of design philosophy. One has to balance A) the possibility of over-voltage or line spikes that might cause the PRV to be exceeded due to whatever mismatching remains, and B) a generalized concept of adding some number of additional diodes for protection. I see nothing to indicate that equalizing resistors cause any harm, so I use the number of diodes indicated by the expected PRV, regardless of the diodes in use, and I add the resistors to be sure they have the best chance for a long and useful life. :)
Jack
--------------------------
Buy Chinese. Bury freedom.
Edits: 07/24/14
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: