BC337 hfe distribution

BC337-40 is a binned-by-hfe part number, so you can't expext a 'natural' hfe distribution curve.

For such a natural distribution, you'd need a part number not subject to selectuive processing, 'probably' made by a firm that does not offer that service for any of it's produce.

RL

RL

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

Yes, it is hfe binned part. Unfortunately I don't have many non-hfe binned transistors, but found some MPSA42 high voltage bipolars. hfe distribution is actually pretty similar in those than in bc337. Min-max:118-142, relative avarage deviation: 2.6%

Maybe I need to buy some more of them to get some idea about batch-to-batch variation

I can only speak for NXP, but it's more like 200k on a 6" wafer (I can look up the actual number).

Not quite. They auto-bin whole wafers into different selection groups of the same type. Meaning that entire wafers end up in one group.

Absolutely. More than 10pcs/sec eutectic die bonding and wire bonding, and a whole assembly line still many hours for a single BAV99 wafer.

robert

Yes they have.

They do. Maybe it hasn't been always that way, but nowadays when a batch starts into the line as BC547 it'll come out as 547.

robert

For applications requiring gain spreads that are narrower than those offered by the vendor, you already seem well situated to address the issue, by direct measurement.

Before accepting hfe as a critical design parameter, you'd best examine effects on measurement method-induced and general environmental effects on this feature, reversible or otherwise.

This information may be more economically obtained from a good reference library, or even a simple google search......

RL

He who designs by hfe, shall die by hfe :-)

Designing by symmetry, I don't ever fret over hfe until I get down near the unity range ;-) ...Jim Thompson

Also don't forget that certain parameters may shift over the lifetime of the part while staying in spec. An extremely unwanted property of electronic devices but sometimes hard to get rid of.

robert

If it matters, hfe rises with temperature AND with aging; Vce falls with age. It's not certain that 'staying in spec' is gonna happen, you know, that's what the 'lifetime of the part' phrase is intended to warn of.

Excellently careful wording, Bob; you could teach to marketers. :-)

Hfe with aging? What's the mechanism?

Same Q...

Google didn't help...

Lots of it is down to raw materials technology.

Over the years the purity of the silicon surely has improved. The fewer the impurities, the more predictable the end product.

Originally transistor specs were mostly quite loose, and tight-specced devices were selected. Nowadaytds even the devices with a wide spec actually have quite a narrow spead of actual performance.