Forums

diode recovered charge

Started by John Larkin September 6, 2014
I got curious about the amount of reverse-recovery charge in PN
diodes, as a function of forward current (and time of fwd bias) and
diode size/type.

Question is, are the LT Spice diode models realistic? We'll have to
test some diodes to see. Since diode recovery for a given part number
depends a lot on the manufacturer, we should stick to sole-source
parts and tweak the Spice model to align with reality.

The ES1D below, straight from the LT Spice diode list, seems to have a
definite step-recovery behavior, which probably isn't realistic. It
stores 115 nC when biased to 1 amp forward. The recovered charge is
not a function of ON time, also unrealistic.


Version 4
SHEET 1 1104 680
WIRE 128 80 16 80
WIRE 336 80 208 80
WIRE 336 112 336 80
WIRE 16 128 16 80
WIRE 336 224 336 176
WIRE 416 224 336 224
WIRE 640 224 496 224
WIRE 784 224 640 224
WIRE 944 224 784 224
WIRE 1008 224 944 224
WIRE 1056 224 1008 224
WIRE 16 256 16 208
WIRE 336 256 336 224
WIRE 496 256 496 224
WIRE 416 272 416 224
WIRE 448 272 416 272
WIRE 640 272 640 224
WIRE 784 272 784 224
WIRE 944 288 944 224
WIRE 448 320 416 320
WIRE 336 368 336 336
WIRE 416 368 416 320
WIRE 416 368 336 368
WIRE 336 400 336 368
WIRE 496 400 496 336
WIRE 640 400 640 336
WIRE 784 400 784 352
WIRE 944 400 944 352
FLAG 16 256 0
FLAG 336 400 0
FLAG 496 400 0
FLAG 640 400 0
FLAG 784 400 0
FLAG 944 400 0
FLAG 1008 224 Qd_nC
SYMBOL diode 352 176 R180
WINDOW 0 -69 44 Left 2
WINDOW 3 -84 14 Left 2
SYMATTR InstName D1
SYMATTR Value ES1D
SYMBOL voltage 16 112 R0
WINDOW 0 51 57 Left 2
WINDOW 3 -16 -79 Left 2
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V1
SYMATTR Value PULSE(-10 100 10u 5n 5n 20u)
SYMBOL res 112 96 R270
WINDOW 0 -45 58 VTop 2
WINDOW 3 -53 57 VBottom 2
SYMATTR InstName R1
SYMATTR Value 1
SYMBOL res 320 240 R0
WINDOW 0 -76 43 Left 2
WINDOW 3 -79 78 Left 2
SYMATTR InstName R2
SYMATTR Value 1m
SYMBOL g 496 240 R0
WINDOW 0 46 61 Left 2
WINDOW 3 23 107 Left 2
SYMATTR InstName G1
SYMATTR Value 1000
SYMBOL cap 624 272 R0
WINDOW 0 57 22 Left 2
WINDOW 3 62 54 Left 2
SYMATTR InstName C1
SYMATTR Value 1n
SYMBOL res 768 256 R0
WINDOW 0 56 40 Left 2
WINDOW 3 60 70 Left 2
SYMATTR InstName R3
SYMATTR Value 1g
SYMBOL diode 960 352 R180
WINDOW 0 -53 48 Left 2
WINDOW 3 -55 17 Left 2
SYMATTR InstName D2
SYMATTR Value Dx
TEXT 664 112 Left 2 !.tran 25u
TEXT 600 64 Left 2 !.model DX D(Vfwd=0)
TEXT 552 -32 Left 2 ;Diode Reverse Charge Tester
TEXT 592 16 Left 2 ;J Larkin    Sep 6, 2014
TEXT 936 176 Left 2 ;1 volt per nC

-- 

John Larkin         Highland Technology, Inc

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

On Sat, 06 Sep 2014 12:31:46 -0700, John Larkin
<jlarkin@highlandtechnology.com> wrote:

> >I got curious about the amount of reverse-recovery charge in PN >diodes, as a function of forward current (and time of fwd bias) and >diode size/type. > >Question is, are the LT Spice diode models realistic? We'll have to >test some diodes to see. Since diode recovery for a given part number >depends a lot on the manufacturer, we should stick to sole-source >parts and tweak the Spice model to align with reality. > >The ES1D below, straight from the LT Spice diode list, seems to have a >definite step-recovery behavior, which probably isn't realistic. It >stores 115 nC when biased to 1 amp forward. The recovered charge is >not a function of ON time, also unrealistic. > >
Adding UIC to the transient analysis does make Qrr depend on Ton, but not much. -- John Larkin Highland Technology, Inc jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
SPICE is notoriously bad at reverse recovery.  I've heard of some models 
which claim to produce realistic results, but not seen.

Probably would help with transistors too.  I typically see circuits going 
about 1/2 to 1/3 as fast as SPICE would claim.

Tim

-- 
Seven Transistor Labs
Electrical Engineering Consultation
Website: http://seventransistorlabs.com

"John Larkin" <jlarkin@highlandtechnology.com> wrote in message 
news:nfpm0ah35hr2e8s968lrrd5ti67n52m3ej@4ax.com...
> On Sat, 06 Sep 2014 12:31:46 -0700, John Larkin > <jlarkin@highlandtechnology.com> wrote: > >> >>I got curious about the amount of reverse-recovery charge in PN >>diodes, as a function of forward current (and time of fwd bias) and >>diode size/type. >> >>Question is, are the LT Spice diode models realistic? We'll have to >>test some diodes to see. Since diode recovery for a given part number >>depends a lot on the manufacturer, we should stick to sole-source >>parts and tweak the Spice model to align with reality. >> >>The ES1D below, straight from the LT Spice diode list, seems to have a >>definite step-recovery behavior, which probably isn't realistic. It >>stores 115 nC when biased to 1 amp forward. The recovered charge is >>not a function of ON time, also unrealistic. >> >> > > Adding UIC to the transient analysis does make Qrr depend on Ton, but > not much. > > > -- > > John Larkin Highland Technology, Inc > > jlarkin att highlandtechnology dott com > http://www.highlandtechnology.com >
On Sunday, 7 September 2014 10:12:00 UTC+10, Tim Williams  wrote:
> SPICE is notoriously bad at reverse recovery. I've heard of some models =
which claim to produce realistic results, but not seen.
> =20 > Probably would help with transistors too. I typically see circuits going=
about 1/2 to 1/3 as fast as SPICE would claim. The transistor models that you get with LTSpice and from manufacturers are = Gummel-Poon models. What the manufacturers use in-house, are VBIC models, which LTSpice can run= if you can find the parameters. The manufacturers won't give them to you -= they are commercial-in-confidence. You could - in theory - work them out f= or yourself, but I've never seen any actual VBIC model parameters for a rea= l transistor. http://www.designers-guide.org/VBIC/ I raised this point here, a few years ago, when I was trying to get an LTSp= ice model of a Baxandall class-D oscillator built with bipolar transistors = to "squeg" with a relatively high feed inductance. The simulations I ran made it fairly clear that it was something about the = operation of a bipolar transistor in inverted mode (using the emitter as if= it were a collector and the collector as if it were an emitter) that produ= ced "squegging" but equally that the Gummell-Poon model didn't capture that= particular behavior, because my LTSpice circuits always settled down to un= iform oscillation where the real circuits had never settled down to a const= ant amplitude oscillation. --=20 Bill Sloman, Sydney
On Sat, 6 Sep 2014 19:12:00 -0500, "Tim Williams"
<tiwill@seventransistorlabs.com> wrote:

>SPICE is notoriously bad at reverse recovery. I've heard of some models >which claim to produce realistic results, but not seen.
Looks that way. I may have an application where an accurate power diode reverse recovery model really matters. We may wind up characterizing candidate diodes and then making some weird nonlinear capacitor behavioral model, nothing like a normal diode model. -- John Larkin Highland Technology, Inc jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
On Sat, 06 Sep 2014 20:25:21 -0700, John Larkin
<jlarkin@highlandtechnology.com> wrote:

>On Sat, 6 Sep 2014 19:12:00 -0500, "Tim Williams" ><tiwill@seventransistorlabs.com> wrote: > >>SPICE is notoriously bad at reverse recovery. I've heard of some models >>which claim to produce realistic results, but not seen. > >Looks that way. I may have an application where an accurate power >diode reverse recovery model really matters. We may wind up >characterizing candidate diodes and then making some weird nonlinear >capacitor behavioral model, nothing like a normal diode model.
PIN diode Attenuator? Cheers
John Larkin wrote:

> > I got curious about the amount of reverse-recovery charge in PN > diodes, as a function of forward current (and time of fwd bias) and > diode size/type. > > Question is, are the LT Spice diode models realistic? We'll have to > test some diodes to see. Since diode recovery for a given part number > depends a lot on the manufacturer, we should stick to sole-source > parts and tweak the Spice model to align with reality.
Come on, you heard it from Thompson. Spice has no charge issues. ;-)
On Sat, 06 Sep 2014 21:46:57 -0700, miso <miso@sushi.com> wrote:

>John Larkin wrote: > >> >> I got curious about the amount of reverse-recovery charge in PN >> diodes, as a function of forward current (and time of fwd bias) and >> diode size/type. >> >> Question is, are the LT Spice diode models realistic? We'll have to >> test some diodes to see. Since diode recovery for a given part number >> depends a lot on the manufacturer, we should stick to sole-source >> parts and tweak the Spice model to align with reality. > >Come on, you heard it from Thompson. Spice has no charge issues. ;-)
I wish the Trr simulation were a bit better (I use LTspice) but I just exagerate the TT parameter and it seems to work "OK". I just thought I wasn't doing things right, but maybe it just doesn't work as good as it should ? This is a very imporant issue for correct simulation IMO. Seems most of my design problems are designing around all these defective parts ! :) boB
On Sat, 06 Sep 2014 23:36:27 -0400, Martin Riddle
<martin_rid@verizon.net> wrote:

>On Sat, 06 Sep 2014 20:25:21 -0700, John Larkin ><jlarkin@highlandtechnology.com> wrote: > >>On Sat, 6 Sep 2014 19:12:00 -0500, "Tim Williams" >><tiwill@seventransistorlabs.com> wrote: >> >>>SPICE is notoriously bad at reverse recovery. I've heard of some models >>>which claim to produce realistic results, but not seen. >> >>Looks that way. I may have an application where an accurate power >>diode reverse recovery model really matters. We may wind up >>characterizing candidate diodes and then making some weird nonlinear >>capacitor behavioral model, nothing like a normal diode model. > >PIN diode Attenuator? > >Cheers
No, a sort of nonlinear transmission line terminator, low impedance for smallish signals and high impedance for big ones. Kilovolts. We could do this with a series string of PN rectifier diodes, high-voltage PIN diodes, or maybe even Z5U nonlinear capacitors. But we'd need accurate models of whatever we use, to let us simulate the whole complex mess. Playing with the standard 1N914 in LT Spice, it shows no forward recovery (ie, no turn-on delay). It does store reverse charge, but it snaps off instantly, probably just junction capacitance limited, when the charge is exhausted. Power diodes do that in Spice, too. This is interesting: https://dl.dropboxusercontent.com/u/53724080/Diode_TurnOn/1N914_Spice.jpg https://dl.dropboxusercontent.com/u/53724080/Diode_TurnOn/1N914_d.JPG Spice has no turn-on delay and gets the reverse charge wrong by maybe 3:1 or so. Could be worse, I guess. That particular 1N914 makes a decent SRD, not as fast as the Spice model but still interesting. Enough fun for now. Gotta go to Safeway. -- John Larkin Highland Technology, Inc jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
On 06/09/2014 20:31, John Larkin wrote:
> > I got curious about the amount of reverse-recovery charge in PN > diodes, as a function of forward current (and time of fwd bias) and > diode size/type. > > Question is, are the LT Spice diode models realistic? We'll have to > test some diodes to see. Since diode recovery for a given part number > depends a lot on the manufacturer, we should stick to sole-source > parts and tweak the Spice model to align with reality. >
You might like to look at how Simetrix deal with this (they offer a node limited free version of their simulator).