Hi,
I can follow those rules/ideas that Early creates distortion and that some topologies can eliminate some of that distortion.
Why the distinction between Low Early and High Early voltages for different devices?
I seen it stated a few times that a better device is available and someone reminds us that it has Low Early voltage as if that is something that over rides all other benefits.
I can follow those rules/ideas that Early creates distortion and that some topologies can eliminate some of that distortion.
Why the distinction between Low Early and High Early voltages for different devices?
I seen it stated a few times that a better device is available and someone reminds us that it has Low Early voltage as if that is something that over rides all other benefits.
AndrewT said:
Of course, lower Early voltage meas more nolinearities, so more distortions. Even a cascode can do only so much to linearize a gain stage.
But beyond distortions, the Early voltage has also a "linear" (that is, "small signal") negative impact. It defines the device output impedance, r0= Va/Ic. This is limiting the simple common emitter gain to gm*r0, ot Va/(kT/q), that is, if the there's no external load, then the maximum gain is limited internally. Of course we want this internal gain to be as large as possible. r0 is in parallel with the external load and for low Va devices, to keep the gain constant, you need to increase the external load, which is not always convenient or even possible.
It has been pointed out that the real figure of merit for a transistor in regard to Early effect is the product of beta and Early voltage.
For example, a general purpose transistor with a beta of 100 and an Early voltage of 100 V will have a product of 10,000.
The reason why this figure of merit can be seen to be important is that the ideal output impedance of a cascode transistor is:
Rout = Beta * VA/Ic
So if you know the figure of merit and you know the operating current, you know the output resistance.
For a device with a FOM of 10,000 and an operating current of 1 mA, this turns out to be 10M ohms.
Cheers,
Bob
For example, a general purpose transistor with a beta of 100 and an Early voltage of 100 V will have a product of 10,000.
The reason why this figure of merit can be seen to be important is that the ideal output impedance of a cascode transistor is:
Rout = Beta * VA/Ic
So if you know the figure of merit and you know the operating current, you know the output resistance.
For a device with a FOM of 10,000 and an operating current of 1 mA, this turns out to be 10M ohms.
Cheers,
Bob
Njw0281/ 0302
Talking about high-quality OPS BJT devices, any idea how the NJW0281/NJW0302 devices are? They claim 10% gain matching between PNP and NPN complementary pairs, and have very high gain (hfe 75-150) compared to other 150W OPS devices. This should translate to lower VAS loading, I guess?
Mouser quotes $1.40 for an NJW0281or 0302. Quite affordable. At these prices, there may be no need to ever use the 2SA1943/5200 pairs or any of the usual alternatives.
Talking about high-quality OPS BJT devices, any idea how the NJW0281/NJW0302 devices are? They claim 10% gain matching between PNP and NPN complementary pairs, and have very high gain (hfe 75-150) compared to other 150W OPS devices. This should translate to lower VAS loading, I guess?
Mouser quotes $1.40 for an NJW0281or 0302. Quite affordable. At these prices, there may be no need to ever use the 2SA1943/5200 pairs or any of the usual alternatives.
You can order these from Mouser: NJW0302 and NJW0281. No idea about European resellers, but then most of the big chaps like Mouser, Digikey, etc, have European subsidiaries, don't they? Have you checked findchips.com to see other resellers? Mouser doesn't have them in stock. Lead time is 7 weeks estimated.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.