Theorectically the transconductance is gm = qIC/(kT) and re = kT/(qIE), where kT/q is Boltzmann's constant times the absolute temperature divided by the elementary charge, which is somewhere between 25 mV and 26 mV at room temperature. As the collector DC current IC is quite close to the emitter current IE, the transconductance and re are practically each other's reciprocal.
Your approximation takes 25 mV for kT/q.
Your approximation takes 25 mV for kT/q.
Dang sir, now that is an explanation I can understand, my figures come from AoE 2nd/3rd edition as they are the only decent references I have here in hard copy. Something about a book that's easier in some regards.
I've got to figure out this base resistance next. All of these little hidden "resistors" can really make a huge difference in precision design.
I've got to figure out this base resistance next. All of these little hidden "resistors" can really make a huge difference in precision design.
Although I saw responses to the other questions, the answer to "Bob who?" is Cordell.
Also, elsewhere around here is some fantastic discussion with another group of sharp folks looking at creating 990 versions with JFET inputs.
You might not be looking for JFET inputs for your particular application, but if you're really in learning mode, you're doing yourself a disservice by not looking though those discussions.
Keywords to search for to help find it are: SWOPA, FET990.
Also, elsewhere around here is some fantastic discussion with another group of sharp folks looking at creating 990 versions with JFET inputs.
You might not be looking for JFET inputs for your particular application, but if you're really in learning mode, you're doing yourself a disservice by not looking though those discussions.
Keywords to search for to help find it are: SWOPA, FET990.
I've shelved it for short while so I can understand some of the more complex interactions. I spent weeks trying to understand 990 and then set up a full prototype with sufficiently functionality to test the application in an afternoon with op amps. So my next job is (after I've done some more reading and some bench testing) is to make up a PCB with everything pre-wired but ready to drop the components into place.
Thanks @PB2 that will be amazingly helpful! I'm not 100% on SPICE, it's a useful tool for sure but even with decent models, there's nothing like the real thing as SPICE can't allow for bad board layouts, for example... or ground loops. I leaned a bit too hard once and got bitten. Badly. That was a hard and expensive, but valuable lesson.
The other end of this job is "powered" by an LSK170 (or LSK389) with a bootstrapped source follower driven by an AB "power" amp. I can only speak from experience but the 170 has performed excellently so far. I can use a 389 in parallel to reduce the noise but it's so low with a typical condenser that it hardly seems worth the extra. It's my "design" but basically pinched from a design by Rode, I just selected the component value's for sanity and FET current.
Thanks @mlloyd1 I'll have a gander at that that though, but the AoE is pretty clear that for my specific application (with very low input impedance) that I need a bipolar (or several bipolar transistors). I'll admit it's very tempting though, large die JFETs are very quiet indeed. Esp. when compared to single-rail, "budget" op amps... I used some LM324s for testing purposes - and the amount of (I assume primarily Johnson noise) coming out of that thing with just 40dB of gain is horrendous! This thing sounds like an untuned FM receiver with the mute function disabled.
It does what I intended it to do (proof of concept) but I think we'll have a battle to get low-noise quads from my usual place - not a biggie really, but it's depressing how quiet even a TDA2822 is and that's donkey's years old; not to mention surprisingly powerful for such an elderly design intended for portable transistor radios.
I got a note to say some of this discussion has been removed so I can only apologise if I've missed anything.
Thanks @PB2 that will be amazingly helpful! I'm not 100% on SPICE, it's a useful tool for sure but even with decent models, there's nothing like the real thing as SPICE can't allow for bad board layouts, for example... or ground loops. I leaned a bit too hard once and got bitten. Badly. That was a hard and expensive, but valuable lesson.
The other end of this job is "powered" by an LSK170 (or LSK389) with a bootstrapped source follower driven by an AB "power" amp. I can only speak from experience but the 170 has performed excellently so far. I can use a 389 in parallel to reduce the noise but it's so low with a typical condenser that it hardly seems worth the extra. It's my "design" but basically pinched from a design by Rode, I just selected the component value's for sanity and FET current.
Thanks @mlloyd1 I'll have a gander at that that though, but the AoE is pretty clear that for my specific application (with very low input impedance) that I need a bipolar (or several bipolar transistors). I'll admit it's very tempting though, large die JFETs are very quiet indeed. Esp. when compared to single-rail, "budget" op amps... I used some LM324s for testing purposes - and the amount of (I assume primarily Johnson noise) coming out of that thing with just 40dB of gain is horrendous! This thing sounds like an untuned FM receiver with the mute function disabled.
It does what I intended it to do (proof of concept) but I think we'll have a battle to get low-noise quads from my usual place - not a biggie really, but it's depressing how quiet even a TDA2822 is and that's donkey's years old; not to mention surprisingly powerful for such an elderly design intended for portable transistor radios.
I got a note to say some of this discussion has been removed so I can only apologise if I've missed anything.