Yes, I'm looking for a NPN type that has:
1. as high hfe as possible
2. high transconductance at around 100mA
3. capable of more that 1W
4. Will be run cascoded - no particular voltage requirements.
5. Reasonably easy to get.
Any tips?
Thank's
1. as high hfe as possible
2. high transconductance at around 100mA
3. capable of more that 1W
4. Will be run cascoded - no particular voltage requirements.
5. Reasonably easy to get.
Any tips?
Thank's
All bipolar transistors have the same transconductance "gm". It equals Ic/Vt where Vt is the "thermal voltage" (kT/q), which is 25.8 millivolts at room temperature. So at 100mA and room temperature, all bipolar transistors have gm = 3.88 siemens.
Mouser.com sells the KSD1691G (link) for $0.53. Its datasheet guaranteed beta is 200-400 ; I measured a few of them at 8mA and found Beta=320. Max permitted current is 5 amps and it's rated to dissipate 1.3 watts without a heatsink, 20 watts with a heatsink.
Mouser.com sells the KSD1691G (link) for $0.53. Its datasheet guaranteed beta is 200-400 ; I measured a few of them at 8mA and found Beta=320. Max permitted current is 5 amps and it's rated to dissipate 1.3 watts without a heatsink, 20 watts with a heatsink.
Thank's, I will check these up.
Yes, I know the old formula appr. 40 * Ic from my school book. But I didn't know how universal it is. Does it apply for power BJT's as well? For all currents? Power devices seem to have a more resistive Vbe/Ic curve at higher currents.
Yes, I know the old formula appr. 40 * Ic from my school book. But I didn't know how universal it is. Does it apply for power BJT's as well? For all currents? Power devices seem to have a more resistive Vbe/Ic curve at higher currents.
The formula is very general, and stems from first principles, which is why it is exponential in nature and applies to almost anything: Si transistors, of course, Ge transistors too, even vacuum or electrolytic devices.
In practice though, there are corrections due non-idealities like ohmic contacts, imperfect semiconductors, etc.
Basically, you need to compute the "ideal" 1/gm resistance, and add the parasitic series ohmic resistances.
In practice, given the non-idealities of real junctions for Si, a better substitute for the basic relation Re=26/Ie (at room temperature, Ie in mA) is 26*Ie-0.8 (according to practicians like H. Schreiber), if you need a good accuracy on a wide range.
The exponent can vary from -1 to -?, depending on the perfection of the device.
For a LM394, it is probably close to -1, and for some generic low power HV transistors, it could be -0.7.
There are now lots of superbeta transistors around: the ZTX series is not the only one anymore: there are BSP in european nomenclature, and 2SC, 2SD in asiatic series: transistors for power conversion for flash in disposable cameras are typical examples
Have not searched, have not simulated, have not attempted measured-vs-simulated comparison. Sorry.
The 2SC6099 I have been using are ranged 300-600, 0.8 up to 15 watts with thermal attention. In use the ones I have are 500+.
If you really want to push the hFE envelope you might look at 'muting' transistors like the 2SD2153 and 2SD2537 - can be had at 820-1800 range. 0.5 to 2 watts with thermal attention. I don't think you'll be able to find the 1200-2700 grade ones easily.
If you really want to push the hFE envelope you might look at 'muting' transistors like the 2SD2153 and 2SD2537 - can be had at 820-1800 range. 0.5 to 2 watts with thermal attention. I don't think you'll be able to find the 1200-2700 grade ones easily.
Yes, 2SD2153 was a real hfe-cruncher! But it's unavailable at Farnell.
But the ZTX960 is absolutely perfect for me. And it's included in my simulator as well.
Thank's
But the ZTX960 is absolutely perfect for me. And it's included in my simulator as well.
Thank's
Perhaps anyone knows some N-channel mosfet with high transconductance at 100mA? At least higher than the usual IRF610.
Voltage may be low. Current < 1A.
Voltage may be low. Current < 1A.
Toshiba SSM3K123TU - gm about 2S @ 100mA. Incidentally FZT688B has typical hfe 1k @ 100mA.
Last edited:
Apologies for reviving such and old topic, but I used the search function first and this topic checked all my requirements, but for one. I’m looking for a high hfe transistor able to handle at least 400V for a high voltage power supply, any suggestions!
I measured Beta=220 on a handful of ZTX558 (max Vce = 400V). If that's insufficient you might consider connecting two of them in a darlington pair.
High gain and high voltage tend to be mutually exclusive, as high gain requires a narrow base region and high voltage requires a wide base region (for the CB depletion zone to grow into without punch-through). Cascoding it the solution usually.
Last edited:
One of many teachings of Andy Grove as well as Mark here is that maximum hFE is inverse square root of breakdown voltage.
SuperBeta transistors 2V max. 40V parts maybe hFE=1,200. Other-Mark's 220 @ 400V is very good.
Yeah, Darlingtons (maybe with separate collectors), cascodes, MOSFETs....
Thanks gents, I'll have a look at those ZTX parts, or try a darlington, these are meant for gyrator duty for a high voltage supply, so that could certainly work.