From: Winfield Hill
Subject: Re: Transistors circuits by the thousands
Date: 1 Jan 2003 18:26:43 -0800
Organization: Rowland Institute
X-Newsreader: Direct Read News 4.20
> Fred wrote:
>> I'm looking at the datasheets for a 2n3904. It specifies an hfe
>> of at least 20 at ic=0.1mA. The "typical" performance curve shows
>> a gain of 70. I need the hfe to be at least 50.
>> Does anyone have experience with this? Do you always use worst
>> case values for high volume designs?
> Typically, yes. ;-)
> Discrete BJTs are often binned for beta - for example, one might use
> (for example) a slightly cheaper and more common C1815Y if one can,
> a GR if one can't, and probably avoiding the O and BL extremes.
> You just design for and order the standard beta range you need, no
> big deal, really, and it's spelled out on the data sheets.
Right, don't a complete production design without selecting transistors
that have a sufficient specified minimum gain. True, the Philips 2PC1815
gives you graded gain ranges, 200 minimum for the GR version. But the
Toshiba part only allows for a minimum gain of 70 at 2mA. So be careful.
BTW, Toshiba's data sheet provides more detail than the Philips version.
Nonetheless, most manufacturers do offer graded-gain parts. For example
Zetex makes the Super E-Line series, with some parts having high gains
even at high currents. Like the ZTX-1051A, with a minimum beta of 290
at 10mA ($1.14 from DigiKey, http://www.zetex.com/3.0/pdf/ZTX1051A.pdf )
and 45 min at 10A! Note, this beast has a rather high 27pF capacitance.