[SOLVED] How can I simulate transistor's avalanche in a pulse generator circuit

Hi all, I'm now make a board to produce a sub-nanosecond pulse using avalanche transistor.
And I made it successfully.The second Fig is the waveform I tested using a 4GHZ bandwidth oscilloscope.

But when I want to produce a more narrow pulse, I don't know which parameter should be adjusted.
I had simulated this circuit with ADS, and it works well. But I doubt it's result because what I used transistor module in this circuit doesn't has parameters about avalanche.
I think it's only works as a simple switch, not contains avalanche breakdown.
Simulate model is download in https://www.infineon.com/cms/en/pro...nel=ff80808112ab681d0112ab6b2b24074e#ispnTab5

This show's that simulation model download from infineon is not contain parameter about avalanche.

In avalanche transistor Wikipedia page, says follows:
NXP Mextram home page A very rich repository of documents about the Mextram bipolar junction transistor SPICE model, capable of avalanche breakdown behavior simulation.
But I can't find specific transistor model in NXP Mextram page.

so,how can I find a transistor simulation model that contain avalanche parameter and the success simulate avalanche breakdown to produce a pulse?
Thanks a lot!

"capable of" is different than "properly implemented"
with foundry-flow-realistic simulated response. This
latter requires that somebody at the foundry (or the
aftermarket) step up and do work fitting params to
data.

Your specific transistor model is thus likely to be
found elsewhere than the compact-model web page.
If you're lucky you might find some remarks about
the early or more enthusiastic adopters of the model,
and chase through -their- web-accessible material.

However, SPICE models are part of the "secret sauce"
according to most (if not all) foundries and seldom
found public, without you engaging with the foundry
on their legal terms.

I'm now make a board to produce a sub-nanosecond pulse using avalanche transistor. And I made it successfully.

Click to expand...

I'm quite sure that the waveform doesn't show avalanche breakdown. just linear amplifier operation.

If you put it to the test and find the model you have
fails to demonstrate the avalanche {whatever} effect
you're expecting (or can demonstrate on a physical
sample) then you would want to look at macromodel
approaches, like add a shunt zener C-B with its BV
and RBV fitted realistically, a high Vf and as little Cj
and leakage as possible. Many times I've found this
necessary, like when trying to model ESD GGNMOS
clamps.

However I suspect that what you are after with an
"avalanche transistor" is probably not the breakdown,
but some greater pulsed current gain - maybe like an
avalanche photodiode but base modulated instead?
This you might have to address by something like a
cccs for excess avalanche gain, "sniffing" the zener
breakdown current. In any case you may need to do
more work than you'd like, to get the behavior you're
interested in to show up.

Unless Googling "avalanche transistor SPICE model"
turns you up an Easter egg. Which it does, plenty.

**broken link removed**

for example.

Now why an avalanche transistor is any better for a
5V, 1nS pulse than (say) a GaAs or GaN HEMT I have
zero idea. Certainly there are many solutions for a
simple impulse amplifier in this voltage range.

Hi FvM,
So the transistor doesn't work in avalanche region?
But I read some articles say that when supply voltage is bigger than BVceo and less than BVcbo, it works in avalanche region.

[Laser pulser for a time-of-flight laser radar]

[A. Ameri, G. Kompa, and A. Bangert, ''Balanced Pulse Generator for UWB Radar Application'', 8th European Radar Conference (EuRAD), pp. 198-201]
For BFP420, BVceo = 4.5V and BVcbo = 15V , so when supply voltage is 8V , it should be worked in avalanche region.
Is anywhere wrong here? I'm really confused about it.
Thanks a lot for reply!

- - - Updated - - -

Hi dick_freebird,
Thanks a lot for transistor simulation module.
I had found the spice model of FMMT413 on diodes.com, but when I use ADS to
analysis it's I-V curve, it seems not occur breakdown either.And also in LTSPICE and multisim.


What I want to is using avalanche transistor to produce a pulse with 1ns width and then using
SRD to shape it to about 200ps, finally have a pulse with 200ps width and about 5V amplitude
in an 50 Ohm load. It was used to drive a laser diode.

The reason why I use avalanche transistor not GaAs or GaN HEMT is that I find more article
about pulse generator using avalanche transistor, SRD, tune diode, and NLTL.:smile:
So is there any discrete HEMT component I can use to make a pulse generator? I would like
to have a try.
Thanks a lot for above detail reply~

Pulse generator design with avalanche transistor and SRD is 70th or 80th technology, used at times when no really fast switching transistors had been available. You see it e.g. in the pulse generation of Tektronix sampling oscilloscopes. But avalanche transistors can only switch on fastly, switch-off is much slower.

I read some articles say that when supply voltage is bigger than BVceo and less than BVcbo, it works in avalanche region.
...
For BFP420, BVceo = 4.5V and BVcbo = 15V , so when supply voltage is 8V , it should be worked in avalanche region.

Click to expand...

1. Basically correct, but the referred voltages are actual Vce0 and Vbe0 of a transistor, not the maximum ratings in datasheet, which is the guaranteed (minimal) value. Actual breakdown voltages can be up to double the maximum ratings. It's reasonable to measure Vcb0 to get an idea of useable avalanche voltage range.

2. It's not guaranteed that a particular transistor type shows avalanche breakdown at all.

Pulse generator design with avalanche transistor and SRD is 70th or 80th technology, used at times when no really fast switching transistors had been available.

Click to expand...

I have found HEMT transistor QPD1009(discrete GaN on SiC. HEMT which operates from DC to 4 GHz), is it can be used as a simple switch as avalanche transistor do?
If it used the circuit I mentioned above, switch-off time is also not very short because it's manly determined by R and C values in the charge loop.
I don't find many useful information about use HEMT to produce a pulse, sorry to bother you but if you could offer more information about new technology about pulse generator,
I would be very grateful!

Besides when to say avalanche transistor,

2. It's not guaranteed that a particular transistor type shows avalanche breakdown at all.

Click to expand...

Does this mean that I must use Semiconductor curve tracer to make sure that the transistor I use shows avalanche breakdown?