BC337 hfe distribution

Just measured hfe from some BC337-40 at about 5 mA collector current. Unfortunately nothing very interesting there

Results: Sample size: 960 pcs minimum hfe: 388 maximum hfe 513 (datasheet limits are 250 and 630) avarage: 452 avarage deviation: 27 median: 463 mode: 477 Distribution is a bit strange looking with two spikes Histogram here:

formatting link

-ek

Reply to
E
Loading thread data ...

Maybe two different batches of parts?

Reply to
Joel Koltner

"Joel Koltner" kirjoitti viestissä:fmLvn.199280$ snipped-for-privacy@en-nntp-02.dc.easynews.com...

Possibly, alltought they were all from same bag. Or maybe just from different wafers? How many BC337 fit one wafer? Probably quite many...

-ek

Reply to
E

Perhaps they take all the middle ones for a tighter-spec version?

--

John Devereux
Reply to
John Devereux

Roughly 80K _untested_ die (if wafer is 8"). The wafer is probably auto-probed and binned into various device numbers.

Discrete device production is so automated... it's fascinating to watch the automated handling equipment :-) ...Jim Thompson

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

"Jim Thompson" kirjoitti viestissä: snipped-for-privacy@4ax.com...

They have still not managed to get their processes so much under control that they can make what is needed instead of what happens to come out? BC337, BC547,

2N2222 etc. have been around for quite some time... (Google didn't know how long)

-ek

Reply to
E

That looks awfully like a mix of two production runs......

Reply to
Sjouke Burry

I think it's more a case of, "for the vast majority of applications, what comes out is already good enough." Indeed, it makes no sense to try to tighten up the specs of a 2N2222 since, by the very nature of selecting it for a design, the designer is saying that tolerances can be quite wide and his design is still going to work anyway.

And keep in mind that BJTs have a lot less spread in their device parameters than FETs anyway. In the simple hybrid-pi small-signal model of a BJT, no device parameters whatsoever come into play! :-) (...which is not true for FETs...)

At the end of the day I imagine it comes down to cost, though -- if it cost next to nothing to trim hFE of a 2N2222 to 1%, it would be happening. (Note that trimming a particular parameter like hFE to 1% *might* imply that you have to trim various parameters in your process far more tightly -- this is what can get expensive...)

Reply to
Joel Koltner

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

Very true, my brother builds machines that makes the wafers....

Reply to
TTman

First hit for "2N2222 history", a few pages in:

formatting link

Introduced by Motorola in 1962.

Anyone know how much it retailed for that year?

With the loose specs on a 2N2222, you'd think a fairly loose run could be binned into all sorts of parts. Maybe they did, once upon a time. 'Course, the 2N2222 spec (and the other three in the series) is so loose as to avoid that, at least initially.

I've heard at least one suggestion that modern 2N2222s should be avoided because, due to the lax specifications, modern processes could be dangerously faster than the original, leading to parasitic oscillation where you aren't expecting it.

Personally, I use 2N3904 and 2N4401, which have proper spec sheets.

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms
Reply to
Tim Williams

I always was amused by the machines that handle the little "D-shaped" plastic transistors... a shaking machine "tub" causes them to flip over on their heads and "walk" up a ramp, flat side against the wall of the ramp. Then another handlers grabs 'em by the leads and tests and bins them. ...Jim Thompson

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

You mean none that you recognize. The transistor hybrid-pi model came about before computers and process extraction, thus is strictly an electrical performance model. But I know a guy (Jim Dunkley) who can extract BJT parameters from diffusion profiles.

...Jim Thompson

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

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

Right -- essentially the first BJT model you encounter in school.

I realize that while this is a good enough model to design, I dunno, audio amplifiers or oscillators and maybe get a start on a 741, anyone doing BJT-based design today (at least for money :-) ) is using far more sophisticated models that do rely heavily on the actual device parameters.

Reply to
Joel Koltner

BJT parameters (for Gummel-Poon) are all based on _electrically_ measurable performances.

Unlike BJT, CMOS is up into the 60-count of model types developed since the original LEVEL=1 :-)

PSpice supports LEVEL=1, 2, 3, 4, 5 (EKVv2.6), 6 (BSIM3v2.0), and 7 (BSIM3v3.2)

Other simulators use different numbering conventions... HSpice lists them up to (IIRC) ~ LEVEL=58, never dropping any model numbers that were flops :-) ...Jim Thompson

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

Hi Jim,

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

Has that ever been a problem for you, yet -- that PSpice hasn't added any of the newer models? Or do all the foundaries still provide older models even once the new ones are available?

It looks good to pointy-haired managers when they can say, "We have 58 transistor models, our competitors only have (the) 8 (that are actually used anymore)!"

---Joel

Reply to
Joel Koltner

I wouldnt waste your time. Transistors will be obsolete according to HP. ;-)

I never heard of this (memristors)?

"Hewlett-Packard scientists on Thursday are to report advances in the design of a new class of diminutive switches capable of replacing transistors as computer chips shrink closer to the atomic scale. "

formatting link

"The most advanced transistor technology today is based on minimum feature sizes of 30 to 40 nanometers ? by contrast a biological virus is typically about 100 nanometers ? and Dr. Williams said that H.P. now has working 3-nanometer memristors that can switch on and off in about a nanosecond, or a billionth of a second. "

Reply to
Hammy

Wow, that's slow. I have 2N3904s that avalanche faster than that!

'Course, they're referring to the memristance effect, which is fundamentally different. When it comes to moving ions in a crystal, that's not too bad.

Heh...speaking of moving ions... wouldn't the gate threshold voltage hysteresis of sodium impurities have a similar effect? But since it's on a gate, it would be more of a transmemristance effect? :-) I wonder if a couple sodium ions buried into a regular DRAM cell could have any useful effect. (Note it's probably a good idea to deposit sodium only in the active cells, and not the word line switches...)

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms
Reply to
Tim Williams

In Europe, traditionally a letter is appended at the end of the type designator to signify the current gain group. If you specify just "BC337", it is quite possible that transistors belonging to a specific group may have been taken out.

In a similar way, if you order 5 or 10 % resistors/capacitors, there might be deep dips in the distribution close to the nominal E12/E24 values, as these are selected out and sold as 1 % components.

Reply to
Paul Keinanen

"E" schrieb im Newsbeitrag news:hpnv2c$5it$ snipped-for-privacy@news.eternal-september.org...

Hello ek,

Have you measured half of the transistors on the other day when the room temperature was a few degrees different? Current gain of transistors heavily depends on temperature.

Best regards, Helmut

Reply to
Helmut Sennewald

"Helmut Sennewald" kirjoitti viestissä:hppqkr$hm6$03$ snipped-for-privacy@news.t-online.com...

No, they are all measured within three hours. I just looked at the data file in 250 pcs slices and distribution looks about same in each slice.

-ek

Reply to
E

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.