The failure mode when a non-avalanche diode goes into avalance breakdown is that the avalanche path - which starts off very narrow - doesn't get any b roader, so it gets hot fast, and the conduction moves from avalanche mode t o pure conduction through molten silicon.
That kind of thermal runaway is remarkably quick, because you don't have to heat up very much silicon at all.
Not that I've heard of, yet.
-- Bill Sloman, Sydney
have
sfor series connection? Think so.
nd
gs
e highest voltage gets the most self-heating, which - in the first instance - increases its leakage current, and consequently increases the voltage ov er all the other diodes, lowering the leakage current to something only sli ghtly higher than the value it started off at, but that's going to mean mor e self-heating in all the other diodes, which could eventually run away.
describe as stabilizing... the voltages will tend to balance.
string overheating.
.Only when he is wrong, which happens rather too often. This is a typical ex ample.
-- Bill Sloman, Sydney
Low Vf schottkys run away at rated reverse voltage. Try it. Maybe not starting from room temperature, but who are you to say what's "room temperature" inside, say, an unvented box in the hot Arizona sun?
Tim
-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Design Website: https://www.seventransistorlabs.com/
Yes. And he's willing, even eager, to be irrational. His emotions overpower his ability to think logically. Most people are like that.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
We use some small RF-type schottkies at two or almost 3x their data sheet reverse voltage abs max. No problems.
Have you ever seen a schottky fail because it was operated within data sheet limits? I have killed some with big forward current surges, charging capacitors.
If I were to design a box to run outside on a telephone pole in a hot climate (which I have actually done a few times) I would bother to estimate the interior temperature. The big hazard isn't over-temperature, it's the locals using the boxes for target practice. After that, lightning.
NASA claimed that our outdoor acoustic monitors (around the Mississippi test facility) were intermittently oscillating at subsonic frequencies in warm weather. It turned out the be the mating calls of bull alligators.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
It's amazing that we can have a debate over something this simple.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
John Larkin wrote
Well now, I was about to put 2 schottkies is series, or brige configuration actually, to see what happens.
-140 V | ------------- | | a a
k k |---V meter---| a a
k k | | /// ///
No 140 V supply at hand now, maybe later, anyone?
(takes out the popcorn)
69883925, (is that your borg ID number?)
As luck would have it I have an hour or so to kill. (Our good voltmeters are on the way to a workshop/ conference.. so ~only good to a few mV)
V tot. V diff (V) (mV)
120 16 mV 130 27 140 47 150 75 160 123 165*** ~700 mV (fluc) hit 100uA current limit in supplyI'm not sure what the numbers mean... care to elucidate?
George H.
My irony meter just pegged.
"Sloman's "reasoning" suggests that a single diode can't be used at its rated reverse voltage." isn't a proposition for which John Larkin could produce any rational support, and he wants to complain that I can't think logically.
He does get very peevish when he doesn't get the flattery to which he thinks he is entitled. He's not all that great at thinking logically even when he isn't feeling peevish.
If he could manage joined up logic he might be less of a sucker for denialist web-sites.
-- Bill Sloman, Sydney
So far. As long as the manufacturer keeps on supplying that part number wit h parts that didn't quite make the next voltage rating up the range you wil l be fine, but if the demand gets high enough that they actually have to ma ke parts to that specification life might get more difficult.
Good. It's always wise to keep ambient temperatures in mind, and the extra heat that you get from direct sunlight.
rSo far the NRA hasn't gone in for the right to carry a concealed flame thro wer, but I wouldn't put it past them - they do have to keep on acttracting the attention of their lunatic fringe
Did you have to find a shade of paint that was less attractive to bull alli gators?
-- Bill Sloman, Sydney
e:
eand/or
'em
ponents,
fI suspect that they mean that your diodes all came from the same batch.
At Cambridge Instruments I got involved with a fold-back current limited su pply built around an LM305, and had to put in a select on test resistor to get get the curren limit to kick in at the desired current.
Final test kept on asking me to drop the select on test because we bought o ur LM305's in six month batches, and for six months at a time they'd fit th e same resistor on every machine. They'd go quiet for amonth or so when the y went over to the new batch, but it took a couple of years before the mess age sank in.
-- Bill Sloman, Sydney
ve
r series connection? Think so.
ighest voltage gets the most self-heating, which - in the first instance - increases its leakage current, and consequently increases the voltage over all the other diodes, lowering the leakage current to something only slight ly higher than the value it started off at, but that's going to mean more s elf-heating in all the other diodes, which could eventually run away.
The voltages don't balance - they just don't movever far away from an perfe ctly uniform drop over each diode. It's the reverse voltage to reverse curr ent relationship that does the almost balancing.
The self-heating is a small positive feedback (if the reverse currents rema in small) and that impairs the voltage balancing rather than improving it, but not by much (as long as the reverse currents and the self-heating are s mall).
ing
That would take a very high reverse current. With the 1N5711 the difference between the typical and worst case reverse current seems to shrink dramati cally as you get close to 150C so you would really have to abuse them to ge t into trouble.
-- Bill Sloman, Sydney
You wanted a float?????
Good work man!
Neither do I, both links could be at ? almost zero or even supply, but trust you also diddit with a resistor divider.... It should convince the unbelievers.
have
for series connection? Think so.
d shighest voltage gets the most self-heating, which - in the first instance
- increases its leakage current, and consequently increases the voltage ove r all the other diodes, lowering the leakage current to something only slig htly higher than the value it started off at, but that's going to mean more self-heating in all the other diodes, which could eventually run away.
fectly uniform drop over each diode. It's the reverse voltage to reverse cu rrent relationship that does the almost balancing.
main small) and that impairs the voltage balancing rather than improving it , but not by much (as long as the reverse currents and the self-heating are small).
tring
ce between the typical and worst case reverse current seems to shrink drama tically as you get close to 150C so you would really have to abuse them to get into trouble.
Thanks Bill, you know I use to not mind interacting with you about technical things. It's not fun anymore. What you say is all correct. But it's almost all stuff I said/learned/ or corrected myself about upstream.
I'm starting with three in series, and 1 meg to ground.
George H.
A name would be easier to remember and respond to.
Nope. I guess I just want to use the diode string for a minimum capacitance switch... I'm trying to switch the the high side of a ~3pF photo diode. It may turn out to be a bad idea. (I was thinking the other night that if I reset the spad before it's 'totally' self quenched, it may just turn back on.)
Let 'em do their own measurements. :^) But thanks... I now feel free to lower the resistance on the ' bottom of the string.. it might help drain the charge out of the switch (pnp) faster... TBH I'm not sure that's important or not.
George H.
e
series connection? Think so.
acteristic.
pendent on current in reverse region, not vice versa.
That is exactly the point. He is talking about boosting the operating volt age of the diode string. The exponential behavior of the reverse character istic means a small difference between diodes will result in a significant difference in voltage at the same current. With the voltage divided betwee n the diodes unevenly increasing the applied voltage will result in one rea ching its breakdown well before the others.
You keep talking about stability and this is not about stability. This is about having an even voltage distribution.
Rick C.
Power schottkies certainly can suffer thermal runaway. I had a design a few years back with multiple boost converters feeding a common 24V rail, with a load share controller wrapped around the whole thing to roughly balance the output currents. The output diode I used was, IIRC, a MMBT2535, because we had lots of them in stock.
One of the boards I had on test fried itself. It didn't take long to find the failure mechanism, the output diode wasn't very well heatsunk and ran fairly hot, maybe 80C or so. The reverse leakage added to the dissipation of the switching FET, which added more heat to the diode. Eventually the diode failed short-circuit, followed by the FET.
I tried a quick test with a MMBT2535 reverse-biased with 24V from a bench supply and warmed it gently with a hot-air gun. It didn't take much, maybe 40-50C before it went into runaway and unsoldered itself from my test leads.
I replaced all the output diodes on my board with MMBT2545, nearly identical part but reverse leakage at rated voltage is about an order of magnitude lower. No problems so far.
:
ence between the typical and worst case reverse current seems to shrink dra matically as you get close to 150C so you would really have to abuse them t o get into trouble.
It's not supposed to be about fun, but rather about clarifying what's going on.
That's bound to happen. I was bit pleased with myself when I worked out tha t the three to one difference in leakage current at 50V could be translated into 6V less reverse voltage drop across a worst case leakage current dido e in a stack with typical leakage curren diodes. If you assume that the "be st" leakage current diode has nine times less leakage current than the wors t case, that's an upper limit of 12V between the reverse voltages across an y diode in a stack.
It's not all that well founded, but it's a whole lot better than nothing, w hich is where we started out.
Real strings, built with diodes from the same batch, are going to be much b etter - as you have demonstrated by experiment, which is another good thing .
I have been posting more comments than I should have done - I'm stuck in a gite in Pommard, Burgundy, and a bit bored. When we've come to Burgundy bef ore - and we've being doing it for some thirty years now, off and on - we'v e been more enthusiastic tourists on earlier visits, but we've got older an d less energetic.
-- Bill Sloman, Sydney
Do you have any links to the part numbers above? I can't find any of either you mention with a search. Thanks.
Maybe something like this,
(I searched for schottky diode 2545) GH
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.