1N4007(1N400x series) is diffused junction normal recovery diode, not fast.
At 1MHz have 25pF capacitance, and it is ridiculous to measure such diode at 4MHz.
Carrier in difused region have, in simply words, some localised zones. At fast changing electic field, there is a lot of nonpredictable situations.
At 1MHz have 25pF capacitance, and it is ridiculous to measure such diode at 4MHz.
Carrier in difused region have, in simply words, some localised zones. At fast changing electic field, there is a lot of nonpredictable situations.
As it turns out, whenever you are moving electrons, it is not the electrons moving, it is you..as in, there is no spoon. Nucleo-centric world view...worry bout dem quarks, gluons, plasmas and such...
Turn on noise? What is that?
Schottky's have no turn off noise? What was it I was measuring back in the day?? The bog standard TRR using controlled inductance current slew will show parasitics just as well as the constant current switching TRR test.
Capacitance is a heavy function of voltage. In forward conduction, it's Huuuge.
I'm not sure what you mean here. Are you saying lateral lack of predictability, vertical, or are you speaking about an entire product line?
John
Here are the raw data files. One of the graphs is labeled wrong due to what I think is a software bug.
Air source pump. There's a big white steel box outside and a little room with lots of dials and pipes down stairs. The hell I know what it all does but the house is toastie in winter - when the kids visit they complain that its too warm.
Ah yes the fens. Nice and quiet now . . . just how I like it! No more biz meetings, customer visits, budgets, gap closing, strategy sessions, **** licking or green tea . . .
The plot shown is 500 uS per division. The 4MHZ is a readout of the noise. It is not a local RF source rather generated in the test circuit.
As it turns out, whenever you are moving electrons, it is not the electrons moving, it is you..as in, there is no spoon. Nucleo-centric world view...worry bout dem quarks, gluons, plasmas and such...
Turn on noise? What is that?
Schottky's have no turn off noise? What was it I was measuring back in the day?? The bog standard TRR using controlled inductance current slew will show parasitics just as well as the constant current switching TRR test.
Capacitance is a heavy function of voltage. In forward conduction, it's Huuuge.
I'm not sure what you mean here. Are you saying lateral lack of predictability, vertical, or are you speaking about an entire product line?
John
You got it right finally. I do not climb stairs, rather I step on the first one and push the Earth down to my level and then repeat the process with each subsequent one until the destination arrives at me!
OK turn on surge, not noise. Ever do the transformer ringing setup? Drive it with a square wave through a small capacitor. Rings on rise and fall.
from Wiki "The most important difference between the p-n diode and the Schottky diode is the reverse recovery time (trr), when the diode switches from the conducting to the non-conducting state. In a p–n diode, the reverse recovery time can be in the order of hundreds of nanoseconds to less than 100 ns for fast diodes. Schottky diodes do not have a recovery time, as there is nothing to recover from (i.e., there is no charge carrier depletion region at the junction). The switching time is ~100 ps for the small-signal diodes, and up to tens of nanoseconds for special high-capacity power diodes. With p–n-junction switching, there is also a reverse recovery current, which in high-power semiconductors brings increased EMI noise. With Schottky diodes, switching is essentially "instantaneous" with only a slight capacitive loading, which is much less of a concern."
Yes switching off a current does produce noise. Schotkys have no surge current due to recovery time.
Ah yes the fens. Nice and quiet now . . . just how I like it! No more biz meetings, customer visits, budgets, gap closing, strategy sessions, **** licking or green tea . . .
And first under water! I do like it tho. Some of my family have moved to near hunstanton and I do like visiting.
Better.
Sigh...wiki....
PN's(diffusion junctions) are minority carrier, schottky's are majority carrier.
PN's recover by diffusion and recombination. Schottky's sweep the charge out.
PN's down to below 100 nSec??? I was testing 15 nSEc power product in an actual product line back in 1981, with testing support for 5 nSec power product the diffusion guys were developing. 1N914's were half nanosecond devices back in '81 as well, my test rig was quarter nano capable, with bandwidth to about 10 gig.
The wiki site needs updating, both in numbers and actual tech...it's not even up to 1981 understandings.
sigh...
John
JN,
You seem to be missing that no new information exists on anything that happened after one graduates from (or leaves) college. All new knowledge stops. (Been burned by that mistake more than once!)
ES
I've no idea what you meant by that last post, nor contextually how it has anything to do with an erroneous wiki page.
John
Interesting, thanks.
I note the 10mA, -10mA, -.1mA values used, typically it was .5/-1/-.25 .
btw, TRR is relatively consistent over 5 orders of magnitude in current density. A power device measured at half amp forward, one amp reverse, quarter amp threshold will measure the same at 1/1000th the currents.
To me, the most important number in that spec is the 100 ohm value they mention. That is the compliance resistance being used to force the constant current in the reverse direction. A lower resistance causes the recovery tail to be longer, so the .1mA number would extend out in time. This due to the drive current falling off of constant current mode as the diode voltage increases. Running a higher series resistance holds the reverse current longer.
John
I do recall something, but twas a while ago...
I seem to recall something about diffusing a well around the periphery to control reverse leakage and breakdown. Power shottky's back in the day were rather high leakage devices, so much so that reverse leakage dissipation would sometimes tilt the decision over to PN junctions even though the shottky's had lower vf's.
Perhaps that was what Pease was talking about.
Or, I wonder if it was an avalanche control junction. With a geometry like the buried junction zeners, where they keep the edge of the tub from avalance for noise control and/or robustness. The 67mm device I'm looking at bevels the edge to control Bvr..
Hmmm.
John
Last edited:
It may have been in one of his bully pulpit columns, or perhaps somewhere in Troubleshooting Analog Circuits.I do recall something, but twas a while ago...
I seem to recall something about diffusing a well around the periphery to control reverse leakage and breakdown. Power shottky's back in the day were rather high leakage devices, so much so that reverse leakage dissipation would sometimes tilt the decision over to PN junctions even though the shottky's had lower vf's.
Perhaps that was what Pease was talking about.
Or, I wonder if it was an avalanche control junction. With a geometry like the buried junction zeners, where they keep the edge of the tub from avalance for noise control and/or robustness. The 67mm device I'm looking at bevels the edge to control Bvr..
Hmmm.
John
Just use a standard bridge and make sure you snubber it.
In an earlier life designing instrumentation products, a very experienced and I'd say somewhat gifted SMPSU designer demonstrated the RFI generated by 50Hz rectifiers. Snubbing them completely solved the problem.
Oh, I almost forgot, that was back in 1987.
In an earlier life designing instrumentation products, a very experienced and I'd say somewhat gifted SMPSU designer demonstrated the RFI generated by 50Hz rectifiers. Snubbing them completely solved the problem.
Oh, I almost forgot, that was back in 1987.
For regular pn junction diodes, the low trr comes to the hefty price of forward drop (that is, increased power dissipation). Gold doping impact is huge, look at this diode, 1.25V forward drop @50mA is the price for a ton of recombination centres to quickly eliminate the minority carriers from the depletion region, at turn off.
And watch out for reverse leakage. The venerable 1N4148 and its ilk have a bunch and are thereby rendered unsuitable for some applications.
Connect enough of them in parallel to get the forward current you seek. Each one recovers in 0.9 nsec -- so the parallel connection recovers in 0.9 nsec.
- Status
- Not open for further replies.