We're paying astronomical prices for 2N2432 and 2N2945's. Does anyone happen to know of any alternatives worth trying?
Graham H
We're paying astronomical prices for 2N2432 and 2N2945's. Does anyone happen to know of any alternatives worth trying?
Graham H
You mean, short of redesign? There are probably tons of ways to do it better nowadays... mosfets, jfets, CMOS analog switches, IC chopamps.
But Zetex is making some super-low Vce_sat transistors that might work as choppers.
John
Thanks John No option to redesign, customer is happy with everything but the price, design is 35 years old. We have just to get the cost of the parts down. We've got some polycarbonate capacitors costing 160UKP (100+). It goes on.
Graham H
I've measured Vsat of ordinary 2n3904s, inverted: around 2-3mV. (Caution: large variation between manufacturers)
Is that good enough?
James Arthur
I sure hope "P" is pence, not pounds.
John
If you diddle Ib, a 3904 can get literally zero saturation in inverted mode. But I think these other critters have millivolt Vce_sat in common-emitter operation.
John
happen
Cool. Vce_sat=0 at what collector currents? I diddled i(b), & think my i(c) was around 1-to-10uA ... not sure.
The application wasn't clear to me--I thought he meant low-level analog signal chopper.
I _love_ Zetex's lo-sats. Inverted mode gain is huge too.
James
happen
I once built a rack full of 16-bit dacs, made from discretes and hand-selected wirewound-resistor R-2R ladder networks. Each bit was...
+Vref = +10 | | | c drive--------------+----------b NPN | e | | | | | +-------Rladder | | | | | e +--------- b PNP c | | | -Vref = -10The transistors are essentially over-driven emitter followers. I tweaked the drive currents on the MS four bits or so to get exactly zero saturation. Overdrive a bit and you get "oversaturation", which would have the NPN emitter going a few mV above +10. Transistors with decent inverted beta are good at this.
The dacs were used to drive a huge wall-sized rear-projection display in some situation room in the Pentagon or something.
John
Perhaps a little talk about reality to the customer is in order. Do you want a lower price?...the we can provide a replacement unit that is better (give list) at a significantly lower price (and give quote).
So...if one is picky about a transistor being upside down, then have the assembly done in Australia...
John
POUNDS!
This analogue computer has UKP3500's worth of them.
Graham H
Robert
This is an ongoing project and time is against us. A new design would cost big numbers and involve flight qualification. At this time there's not much money about for new designs.
Graham H
No. Thanks for checking.
Need
happen
Gotcha--coolness! I hadn't thought of that, I was thinking strictly:
| [RL] | | . | === GND
James Arthur
At what kinds of collector currents?
--James Arthur
happen
I think the term "inverted saturation" has been used to express the above circuit, with the emitter and collector swapped, and my circuit, the over-driven emitter follower.
I think your sketch above is "ordinary" saturation.
Some old transistors were truly symmetric, agnostic as to emitter and collector. Ditto jfets.
There are some enhancement-mode PHEMTS that might make decent substitutes for low-sat NPNs.
John
happen
Don't the electrons spin backwards down there?
John
happen
Yes yes, I screwed up my drawing. Please swap ce.
That explains their high emitter-base breakdowns compared to today's. Sometimes those would be handy.
I too think of your DAC as using over-driven followers, which I hadn't considered as a way of achieving a precision zero-offset. Slick.
As far as inverted mode, here are few samples from my junk box, measured 8/5/2004:
================================= Inverted Mode Saturation Voltages of selected common transistors =================================
Fig. 1 Low Currents, i(c)=26uA ====== V+ device Vsat -+- ------------------- ---- | 2n3904, ITT 2 mV () | 26uA " , Motorola 2 mV () V " , TI, 1960's 65 mV | BC548 1 mV 140uA +-----> Vsat --> |>' .--()()--| D.U.T. | |\\ === | GND === GND
Fig. 2 Higher Currents (~500uA) ======
+5v -+- | [10k] | +-------> Vsat |>' V1 >----[1.5k]---| Device Under Test |\\ | === GNDdevice V1 Vsat
-------------- ---- ---- Philips BC548 2.9V 3.3mV 3.0 3.2 3.6 3.1 5.0 3.1
National Semi PN2222a 5.0V 2.2 mV
Motorola MPSA18 5.0V 2.6 mV Motorola BC547B 5.0V 2.8 mV ITT 2n3904 5.0V 15 mV ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
I'd expected the high-gain MPSA18 to stand out but it surprised me and didn't.
I still imagine the Zetex lo-sat parts would shine, but haven't tested.
Cheers, James Arthur
Yes, and they all have ADD. Have you ever tried to medicate ranting electrons? :(
-- http://improve-usenet.org/index.html Goggle Groups, and Web TV users must request to be white listed, or I
I think you may be paying for the package, more than for the wafer process. Unfortunately there weren't any JEDEC 'choppers' in plastic. (There were quite a few EIAJ plastic parts made for this function.)
As you'd be faced with re-qualification for part package changes like this, I think you're stuck. All hermetic subs will have the same cost penalty.
RL
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.