2N2369 reverse Vbe avalanching, consequences?

Folks,

I'd like to use a MMBT2369 in a design where there are no protection diodes. Mainly because the not so well performing FET that was in there had one integrated. So, in order to get very low storage time I'd like to use the gold-doped 2369 but of course there can be the occasional ESD event. There are 1k to 10k resistors up front, limiting the current. How low can the hfe go upon repeated zenering on the B-E junction?

I could live with a hfe as low as 5. But not if it eventually drops to 1 or zero.

It's a glorified level translator and I'd like to push transitions to less than 100nsec.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg
Loading thread data ...

I don't think anyone has published the rate at which hFE decays, with respect to current magnitude and rate (is it a linear charge dose, square law, or..?), and how that translates to normal parameters.

If you'd like to take the effort to do it, I'm sure a few people would be appreciative, but I'm guessing you don't have that kind of time in the project...

Otherwise, might you be able to find a "self biased" transistor, that's fast enough, and which for some bizarre reason, just happens to have a reverse diode integrated already? (Worthy of note: mere 2N3904s will get transitions under 30ns, but you might not be able to afford the quiescent current to get there.)

You could also hack a SOD-323 (maybe) or MLF or 1206 diode in, as well. Crappy, but can probably still be managed via pick-n-place.

Tim

Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website:

formatting link

Folks,

I'd like to use a MMBT2369 in a design where there are no protection diodes. Mainly because the not so well performing FET that was in there had one integrated. So, in order to get very low storage time I'd like to use the gold-doped 2369 but of course there can be the occasional ESD event. There are 1k to 10k resistors up front, limiting the current. How low can the hfe go upon repeated zenering on the B-E junction?

I could live with a hfe as low as 5. But not if it eventually drops to 1 or zero.

It's a glorified level translator and I'd like to push transitions to less than 100nsec.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Tim Williams

On 12/08/2015 01:35, Joerg wrote:> Folks, > > I'd like to use a MMBT2369 in a design where there are no protection > diodes. Mainly because the not so well performing FET that was in there > had one integrated. So, in order to get very low storage time I'd like > to use the gold-doped 2369 but of course there can be the occasional ESD > event. There are 1k to 10k resistors up front, limiting the current. How > low can the hfe go upon repeated zenering on the B-E junction? > > I could live with a hfe as low as 5. But not if it eventually drops to 1 > or zero. > > It's a glorified level translator and I'd like to push transitions to > less than 100nsec. > Hi Joerg,

Last year I did some evaluation on a circuit where a jelly bean npn bjt (I used mostly 2N3904 and BC848) experienced occasional Veb reverse "zaps". The zap was a 330nF capacitor charged to 30V discharged into the b-e junction via series 470 ohm. In actual use the zap could occur several times per hour and I wanted to see what effect this could have over the expected product lifetime of 10-20 years, i.e. about 1-2 mega zaps.

I rigged a test applying those zaps at 100Hz rate and sampled hfe at intervals. There was a rapid initial drop in hfe of 5-10% which gradually levelled off with increasing zap number and seemed to taper off at hfe about 20% lower than initial. So hfe 400 would after "years" of simulated use reduce to 320. I even tried an ancient 2N2924 from the junk box and saw hfe fall from 89 to 67.

The big problem is that I am quoting from memory as due to my sloppiness I cannot now find my notes from this and cannot post the result tables. If I find them soon I will post or re-do the experiments.

All this was strictly room temp so I cannot say what extremes of temperature would do. Also I did not try a 2N2369 / 2369A. I read somewhere that the 2N3904 is also gold-doped to speed it up even if not as extreme as the 2369?

What I do remember very clearly is that the biggest hfe fall was seen after the first zap.

My zaps are quite low-energy and I am willing to believe that prolonged or stronger zenering the b-e jct will more profoundly degrade hfe. I have heard audio types say that zenering makes the device noisier as an amplifier and can wreck mic front-ends. My app was not noise sensitive, almost logic level.

Since in the end-application any hfe over 20 would have been enough I was happy to leave the design as it was.

Does that help you?

beste

piglet

Reply to
piglet

Found these snippets in Pease's Troubleshooting Analog Circuits:

p66 - where I saw Au doped 2N3904 "The popular 2N930 and 2N3707 have low leakage, typically. Some 2N3904s do, too, but some of these are gold-doped and are leakier."

p78 - about Veb zenering and ESD, mentions 2N2369 "... but in some cases ESD damage can still happen. Delicate RF transistors such as 2N918s, 2N4275s, and 2N2369s sometimes blow up "when you just look at 'em" because their junctions are so small."

I think it is a very good thing you have several k-ohm between the b-e jct and the outside world. In my limited ESD experience that is often enough for HBM provided the resistor bodies are long enough that the ESD does not arc across them, 1206 has worked for me.

piglet

Reply to
piglet

Nah, don't want to turn this into a science project :-)

I have a 3904 in there right now and it's really slow. Turn-off takes well over 200nsec to happen.

Turn-key assemblers generally won't agree yo do that.

[...]
--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

Now that's a word. Good news. In our case it's systems that do not have to last forever so we should be ok then. Only thing is, the 2N3904 is so sluggish in turn-off that it's borderline. But if the digital guy is happy with the speed I'll just leave it at that for now.

At least the 3904's I have sure don't appear gold-doped.

I have 10k in 0805 up front so ESD zaps will be muffled. It's only ESD I am concerned about and there won't be regular zapping.

Yes, thanks, it sure does.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

The resistor in series probably will save the day. However, in an automotive environment... aka HOT, under hood, Arizona... dead in a few weeks (and only 1mA reverse current).

Note: Plastic package made me suspect a chemical degradation. Can you still get 2N2369's in a can? ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

I've wondered about this for some time. Great info, Piglet, and thank you.

Reply to
John S

Sure:

formatting link

They also have the high rel version from Micro-Semi in stock but that's over $7 in singles. Higher bling factor due to its golden tab.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

What's the expected range of input voltage/current and load current, or a snippet of the schematic? That should only be so slow if you don't have a B-E resistor.

I'd be worried that a 2N2369 wouldn't be much faster, and that you need an altogether smaller (lower Ccb and Cbe) part like MMBTH10 or beyond. Which being into the RF range, will be even more ESD sensitive..

Holy crap, it just occurs to me (looking at the datasheet) the 2N2369 claims its storage time at Ib(on) = Ib(off) = Ic = 10mA -- an hFE of 1!! That's harder turn-off than you can even get from a cascode! It's not obvious how much that will change under ordinary circumstances, since it is an unusual transistor, after all.

Don't suppose it would be yet possible to shoe-horn in a more sensible, say, logic inverter?

formatting link
Or the original FET direction, just find a substitute.

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
 Click to see the full signature
Reply to
Tim Williams

Zenering the B-E junction increases noise as well but in my experience everything plateaued fairly early on. Forward biasing the junction to the extent that significant heating occurs will anneal "damage" and lower noise. I don't recall what the effect on hfe was.

The 2N2369 is popular in avalanche mode pulse generators (~90V vce) among hobbyists.

--
Grizzly H.
Reply to
mixed nuts

The 2N2369 is an old-fashioned device... gold-doped... which kills storage time. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

Well,, if zenering them degrades the beta below what is tolerable,just anneal out the damage (i think 185-250F for a few hours)

Reply to
Robert Baer

  • Acts exactly like as if radiation damage happened. Key words "dose rate", "total dose". Can be annealed out (for the most part, especially for less than moderate total dose).
Reply to
Robert Baer

...and already 'aged' 25 years, on the shelf.

RL

Reply to
legg

We have 10k in series with the base and 10k in the emitter. The values have almost no impact on turn-off time.

Kinda too late at this point :-)

It works ok now with the 3904, not pretty but good enough for prototyping.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

Wasn't that the old trick on certain HP equipment where reverse Vbe was exceeded and then the tech extinguished a cigarette on the metal can?

Of course now that smoking is no longer PC that won't work. But it wouldn't work on platic anyhow.

[...]
--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

The damage is done due to dopant _migration_... no amount of "annealing" is going to help... might even make it worse. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

If it's an electromigration issue, the reversal of bias followed by heating can undo the damage. That's mainly a surface-contamination effect, not the bulk dopants. Radiation damage is different, there's either alphas (helium nuclei) or dislocations inside the bulk material that cause strains/leakage; annealing can remove the dislocations (or allow the helium to diffuse out) without moving any heavy dopant atoms.

Reply to
whit3rd

We're talking _gold_, once it moves you're dead. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

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.