Component variation.

"Boki" a écrit dans le message de news: snipped-for-privacy@i40g2000cwc.googlegroups.com...

Yes. Welcome to the electronics design _real_ world.

So, boki, accept the fact that any given 2n7000 you purchase may be anywhere within that min-max range, or if it's beyond the range, which it could be, then and only then may you return it for a refund. This reality means that we come up with designs that can live with the full range of variation, in order to insure that all manufactured production units will operate the first time and every time, without any further attention or tweaking.

That said, most parts do not show such a wide range of operation, and sometimes critical designs do rely on that fact, employing pre-production sorting, etc., to work with a tighter tolerance. In AoE figure 3.14 we show the full measured variation of Vgs vs Id over a 9-decade drain-current range, for 20 MOSFETs from four different production runs spread over two years. Boki, you should study this graph. The VN01 part we used in the graph is very similar to the 2n7000. One thing, if you were to evaluate 2nd-source parts from different manufacturers, you'd probably see a *much* wider tolerance range than we observed. Now and then a manufacturer changes his fab facility, etc., and we see a step shift in his tolerance-distribution mean.

Fred Bartoli =E5=AF=AB=E9=81=93=EF=BC=9A

It seems that customer asks the 10mA for LED current is reasonable. He care about the beta variation on transistor.

Boki.

He shouldn't, if you design a decent circuit with its performance that doesn't depend on a transistor beta, except that it be higher than say 50.

I got it, thank you so much for your advice.

Best regards, Boki.