Author Topic: Linear solid state amplifier BJT?  (Read 1234 times)

0 Members and 1 Guest are viewing this topic.

Offline ELS122Topic starter

  • Frequent Contributor
  • **
  • Posts: 914
  • Country: 00
Linear solid state amplifier BJT?
« on: June 30, 2022, 09:58:14 pm »
a dumb question but, is there any silicon device that acts like a triode tube? really linear?
for some reason I was under the impression that JFET's were linear, but they only are when the voltage across them is really low.
and I don't want to have like 500mV headroom before saturation and cutoff.

so is there some BJT that is linear throughout it's operating voltages?

and yes I know that I can add DC feedback from the drain to gate, but where's the fun in that  :D
« Last Edit: June 30, 2022, 10:00:32 pm by ELS122 »
 

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 7934
  • Country: us
  • Retired, now restoring antique test equipment
Re: Linear solid state amplifier BJT?
« Reply #1 on: June 30, 2022, 10:12:01 pm »
The vacuum triode is the only common device that, under certain reasonable conditions, is theoretically linear.
To achieve a linear voltage gain, an ideal triode needs to be operated into a high load resistance, ideally a constant-current load.
The vacuum pentode into a "normal" load resistance is similar to a triode into a low load resistance, very roughly 3/2-power, due to the constant voltage on the screen grid.
A BJT as a voltage amplifier, with a "normal" load resistance, is inherently exponential.
A FET (J or MOS) as a voltage amplifier, with a "normal" load resistance, is approximately quadratic.
Degeneration (cathode, emitter, or source resistor, respectively) will help linearize a single device by providing local feedback.
A BJT may have a reasonably constant current gain over a wide range of collector current:  to make a voltage gain, you can add a resistor in series with the base and a load resistor in the collector.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21606
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Linear solid state amplifier BJT?
« Reply #2 on: June 30, 2022, 10:14:28 pm »
Yes, but not very common, mostly a relic.  The Static Induction Transistor is the nearest analog to a triode.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: ELS122

Offline Benta

  • Super Contributor
  • ***
  • Posts: 5839
  • Country: de
Re: Linear solid state amplifier BJT?
« Reply #3 on: June 30, 2022, 10:22:07 pm »
Yes, but not very common, mostly a relic.  The Static Induction Transistor is the nearest analog to a triode.

Tim
Wow! Never heard of that one before. Thanks!
 

Offline mag_therm

  • Frequent Contributor
  • **
  • Posts: 633
  • Country: us
Re: Linear solid state amplifier BJT?
« Reply #4 on: June 30, 2022, 11:05:13 pm »
Is a Triode  a standard for linearity?

For triode common cathode with no feedback:
Ia=A+BVg+cVg^2. To get ABC these days, one way is with curve fitting along load line drawn on published Ia vs Vg curves
Duncan Amps has a large data base .
These parameters can be used for both single tone linearity Vg =ASin(W1*t) and two tone intermodulation, Vg =ASin(W1*t)+BSin(W2*t)

A cathode follower has excellent linearity but no gain.

A grounded grid is a high frequency connection, but can have improved linearity too:
iRL/e_in = (1+u)RL/(Rl+ra+(1+u)Rin)   note the u on numerator and denominator.

I converted an old mixer with ECC83  to grounded grid to match  4000 ~12000 Ohm  from acoustic, bass and Stratocasters
Sound quality is second to LOUD and  your ears will be ringing when the teens are in the basement here.

Of course, all the above can apply to BJT and FET.
 

Offline ELS122Topic starter

  • Frequent Contributor
  • **
  • Posts: 914
  • Country: 00
Re: Linear solid state amplifier BJT?
« Reply #5 on: June 30, 2022, 11:35:58 pm »
Yes, but not very common, mostly a relic.  The Static Induction Transistor is the nearest analog to a triode.

Tim

why aren't they produced more? is a much more linear BJT for high voltages and absurdly high frequencies not useful enough to prompt production for them?
or are they too OP, not wanting hobbyists to be able to mess with mil-spec products  ;D
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21606
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Linear solid state amplifier BJT?
« Reply #6 on: June 30, 2022, 11:54:30 pm »
Because they suck? Triodes suck too... why don't they use 'em anymore, right? ;D

I don't know any SIT characteristics offhand, but evidently they simply aren't competitive to MOSFETs and BJTs.  So, yeh.

Your concern is misplaced -- linearity is a non sequitur for amplifying devices.

As long as the operating range is continuous and one-to-one, it doesn't really matter.    What does matter is how much gain is available in general, what saturation voltage is, and device capacitance or inductance.  We can fix everything else by simply using a couple more devices and applying negative feedback to clean it up, and maybe a few more to re-distort it back to whatever other characteristic we want.

Whereas no amount of feedback will gain you back the pitiful efficiency that triodes have, or the high impedance and relatively low bandwidth of vacuum tubes in general.  (You can use a triode in positive-grid operation to reclaim some efficiency -- which costs some distortion due to grid rectification loading the driver -- but this takes more effort, and again probably needs feedback to manage the distortion.)

RF amps are the most stringent, as there's no time (that is to say, phase margin) for anything more than local feedback -- usually shunt feedback as a output-to-input resistor in a common-source configuration (often called "neutralization", but it's not just compensating for reactance when it's used in this way).  In this case, it does help that conventional devices are much better behaved than they possibly could be -- an exponential transfer function is a far sight better than some lumpy sort of thing you might imagine.  So we get the basic calculus advantage of, small signals span less curvature so have less distortion; and a few tricks like push-pull or other phase shift hybrids (for narrowband RF mostly) to cancel out some harmonics/IMD.  And if you don't mind spending a lot of time figuring things out, you can predistort the signal to perfectly cancel it out with the final's distortion, which is how particularly high linearity power amps are made, for cell towers etc.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline ELS122Topic starter

  • Frequent Contributor
  • **
  • Posts: 914
  • Country: 00
Re: Linear solid state amplifier BJT?
« Reply #7 on: June 30, 2022, 11:58:06 pm »
Because they suck? Triodes suck too... why don't they use 'em anymore, right? ;D

I don't know any SIT characteristics offhand, but evidently they simply aren't competitive to MOSFETs and BJTs.  So, yeh.

Linearity is a non sequitur for amplifying devices.  As long as the operating range is continuous and one-to-one, it doesn't really matter.    What does matter is how much gain is available in general, what saturation voltage is, and device capacitance or inductance.  We can fix everything else by simply using a couple more devices and applying negative feedback to clean it up, and maybe a few more to re-distort it back to whatever other characteristic we want.

Whereas no amount of feedback will gain you back the pitiful efficiency that triodes have, or the high impedance and relatively low bandwidth of vacuum tubes in general.  (You can use a triode in positive-grid operation to reclaim some efficiency -- which costs some distortion due to grid rectification loading the driver -- but this takes more effort, and again probably needs feedback to manage the distortion.)

RF amps are the most stringent, as there's no time (that is to say, phase margin) for anything more than local feedback -- usually shunt feedback as a output-to-input resistor in a common-source configuration (often called "neutralization", but it's not just compensating for reactance when it's used in this way).  In this case, it does help that conventional devices are much better behaved than they possibly could be -- an exponential transfer function is a far sight better than some lumpy sort of thing you might imagine.  So we get the basic calculus advantage of, small signals span less curvature so have less distortion; and a few tricks like push-pull or other phase shift hybrids (for narrowband RF mostly) to cancel out some harmonics/IMD.  And if you don't mind spending a lot of time figuring things out, you can predistort the signal to perfectly cancel it out with the final's distortion, which is how particularly high linearity power amps are made, for cell towers etc.

Tim

but audiophools would love them, same as vacuum tubes, yet vacuum tubes are still produced for the consumer market while SIT's aren't.
maybe the idea of sand being inherently much worse and incomparable to tubes is what keeps SIT's from being produced still.

I guess for now I'm stuck with using tubes, or I could learn how to make a proper BJT amplifiers  :D
« Last Edit: July 01, 2022, 12:00:02 am by ELS122 »
 

Offline mclute0

  • Regular Contributor
  • *
  • Posts: 68
  • Country: us
  • Somnium aut moriar!
Re: Linear solid state amplifier BJT?
« Reply #8 on: July 01, 2022, 12:16:08 am »
like SIT_2SK77B

 

Offline ELS122Topic starter

  • Frequent Contributor
  • **
  • Posts: 914
  • Country: 00
Re: Linear solid state amplifier BJT?
« Reply #9 on: July 01, 2022, 12:18:23 am »
found some 2SK180 output SIT trannies for from 300 to a grand  ::)
 

Offline mag_therm

  • Frequent Contributor
  • **
  • Posts: 633
  • Country: us
Re: Linear solid state amplifier BJT?
« Reply #10 on: July 01, 2022, 12:22:21 am »
As I recall, there were presentations about SIT in 1984 at the world power electronics conference in France.
It sounded too good to be true,  I think  they were used in prototypes ~ 2000~ 3000kW locomotive drives in Japan in the low kHz.

But nothing eventuated as far as devices on the market.
Then ...later....IGBT came along.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21606
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Linear solid state amplifier BJT?
« Reply #11 on: July 01, 2022, 12:22:41 am »
Meh -- at least tubes look nice.  If you put a WE300B in a TO-247 package, no one's going to buy it.  Even if the electrical characteristics were perfectly identical.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: ELS122

Offline WatchfulEye

  • Regular Contributor
  • *
  • Posts: 109
  • Country: gb
Re: Linear solid state amplifier BJT?
« Reply #12 on: July 01, 2022, 12:27:13 am »
Lateral FETs might be worth a mention. Not as linear as SITs, but not quite as esoteric and expensive.

Still actively produced and in stock in a convenient TO247 package for a fiver.
 
The following users thanked this post: ELS122

Offline ELS122Topic starter

  • Frequent Contributor
  • **
  • Posts: 914
  • Country: 00
Re: Linear solid state amplifier BJT?
« Reply #13 on: July 01, 2022, 12:31:24 am »
Meh -- at least tubes look nice.  If you put a WE300B in a TO-247 package, no one's going to buy it.  Even if the electrical characteristics were perfectly identical.

Tim

yeah, in style, tubes can't be beat.

do you perhaps know any terrible BJT's, that have a really big linear region relatively?
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21606
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Linear solid state amplifier BJT?
« Reply #14 on: July 01, 2022, 06:21:17 am »
BJTs are almost all ideal; hFE can be "linear" in the sense that it remains high even at low Ic, but that really doesn't matter much.  Ic is still exp(Vbe), give or take R_E and R_B.

You can always make a "worse" BJT by degenerating the emitter (adding R_E), or other things.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 6733
  • Country: pl
Re: Linear solid state amplifier BJT?
« Reply #15 on: July 01, 2022, 09:38:00 am »
Ic(Vbe) characteristics of power BJTs look quite linear above a few hundred mA due to internal emitter resistance, whether it's parasitic or introduced deliberately to improve current sharing across the die.

Of course an external degeneration resistor achieves the same, and possibly with better thermal stability of gain.
 
The following users thanked this post: ELS122


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf