I have a class-A small signal TIA challenge to be solved with BJTs.
From your experience, which transistor models are similar or better to the BC550 in noise figure, but much higher voltage VCE0?
Maximum input signal, for example, is 5mVp@1KHz delivered by a pickup coil. If I design for 80dB amplification, the output signal is already 50Vp which in class-A it requires VCE0>100V.
This is not a power amplifier, it is a preamp.
Which direction&part do you advise and recommend?
From your experience, which transistor models are similar or better to the BC550 in noise figure, but much higher voltage VCE0?
Maximum input signal, for example, is 5mVp@1KHz delivered by a pickup coil. If I design for 80dB amplification, the output signal is already 50Vp which in class-A it requires VCE0>100V.
This is not a power amplifier, it is a preamp.
Which direction&part do you advise and recommend?
Cascode the low noise, low voltage transistors. Arrange your component value choices & bias currents, so the first stage's gm and RL combine to give a first stage gain around 16 dB. Then simply use high voltage and normal noise transistors in all other stages. This idea is 50 years old and its purest example is the LM308 opamp from National Semiconductor.
Or look up high power (high voltage) solid state power amps with MPS6571 (N.Pass) or MPSA18 (J. Bongiorno) input transistor pairs. 15V-20V devices. Notice the cascoding.
Or look up high power (high voltage) solid state power amps with MPS6571 (N.Pass) or MPSA18 (J. Bongiorno) input transistor pairs. 15V-20V devices. Notice the cascoding.
Of course I can use LM308 (like any GP op-amp) for a TIA. But will be the noise performance equivalent, better or worse to a TIA made with a BC550-based Frontend preamp which works just fine until 60dB as BucksBunny mentioned?
Or, I misunderstood you? You mean that I should clone the LM308 internal details in the discrete world?
Or, I misunderstood you? You mean that I should clone the LM308 internal details in the discrete world?
BTW, this is my first TIA attempt for low input Z, like coils+audio. No experience in this corner.
Previously I played a bit with TIA for high input Z, like charge sensitive preamps until 100MHz.
This weekend I connected the charge preamplifier(s) to the pick-up coil: and yes, the noise result was very disappointing.
So, I need a design for low Z but no experience which parts exactly or topology to use. The starting point for me is the Revox A77/B77 frontend as inspiration.
Previously I played a bit with TIA for high input Z, like charge sensitive preamps until 100MHz.
This weekend I connected the charge preamplifier(s) to the pick-up coil: and yes, the noise result was very disappointing.
So, I need a design for low Z but no experience which parts exactly or topology to use. The starting point for me is the Revox A77/B77 frontend as inspiration.
Clear questions:
In the higher voltage direction: would be right to replace the 25-50V parts BC109/BC549/BC550, with the 120V part KSA992-D, or the 160-180V parts 2N5550-2551 or 2N5401, or with the 300V part MPSA92-D? The noise figure jump from 1.4dB to 8dB almost negates the reasoning to take the voltage jump for higher TIA gain.
Or, maybe I should better go the ZTX route, say in the same topology mentioned ?
Or, the transformers IN and OUT combined with ZTX parts?
In the higher voltage direction: would be right to replace the 25-50V parts BC109/BC549/BC550, with the 120V part KSA992-D, or the 160-180V parts 2N5550-2551 or 2N5401, or with the 300V part MPSA92-D? The noise figure jump from 1.4dB to 8dB almost negates the reasoning to take the voltage jump for higher TIA gain.
Or, maybe I should better go the ZTX route, say in the same topology mentioned ?
Or, the transformers IN and OUT combined with ZTX parts?
as said, I start with a card from Revox A77. Their FrontEnd/TIA values for Repro cards are already optimized for a pick up coil, I think.
So far, this beats already pants down my ORTEC high-Z TIA preamps: of course because of input impedance mismatch.
To improve over the Revox FrontEnd/TIA performance on Repro cards, maybe I should use different starting point, like Studer and Ampex Frontends? Such card(s) would be more rare and expensive, and I wonder myself if this effort will pay off.
2SA872A/2SC1775A from Hitachi if you can find them, these are only 0.3W though, or eventually the 2SA992/2SC1845
as second choice wich have the advantage of a higher 0.5W TDP although i personaly prefer the former, there s also the 2SA970/2SC2240 wich are also 0.3W, all these device have a 120V VCE.
as second choice wich have the advantage of a higher 0.5W TDP although i personaly prefer the former, there s also the 2SA970/2SC2240 wich are also 0.3W, all these device have a 120V VCE.
Aren't you mixing up input impedance and source impedance? A TIA is either a minor brain infarction or a transimpedance amplifier, in the latter case it by definition has a low input impedance.
What kind of coil do you mean? What's its impedance over the frequency range of interest? Why do you care about 100 MHz for audio?
If you can still get them somewhere, the 2SC2547 and 2SA1085 are pretty good low-noise high-voltage transistors. Unfortunately they have been out of production for quite a while.
Thanks to all for your great inputs! I'll study and order the new parts, and also look around in drawers: I remember seeing some old Hitachi ones with these numbers a few years ago. I wonder if they are still there and if the numbers match.
I'm not an EE expert. The bit of EE-training I have is either via hobby or tangential to the main job/training (physicist).
So, my hobby challenge for the Easter holiday: build myself a flux measurement device (e.g. pickup coil) for audio frequencies.
For the source input I have made a pickup coil with 150uH/10Ohm, fT(LR)=10KHz.
Previous personal experience, related to this topic:
That's my bit of previous TIA experience.
Surely I am wrong, otherwise I would not write this here now:
This weekend I recognized&confirmed in measurements the TIA's behaviour on the A77's frontend for heads (their Repro board is a cheap one to buy as a start). As said, this board is planned for being the FE for my pickup coil.
Measurement results:
But, I do not have a proper analysis for this topology how it does so for a low-Z source.
My confusion goes also at the fundamenal level: as it would contradict the TIA definition, according to my understanding. On the other side, it would fit the TIA definition according to Marcel
So far, so good. The planned approach is to push the R_f beyond 60dB, but I need to change the PS and the BC550 transistors. I do have good regulated dual PS for 120V. Should be enough for R_f=100dB and I hope to see the noise floor of the pick-up coil.
Nevertheless, extra help for a proper analysis of A77's repro-FE topology (only the first stage) will be very welcomed here by me, and I guess by many forum friends as well. Thanks again everyone for your time and kind help!
I'm not an EE expert. The bit of EE-training I have is either via hobby or tangential to the main job/training (physicist).
So, my hobby challenge for the Easter holiday: build myself a flux measurement device (e.g. pickup coil) for audio frequencies.
For the source input I have made a pickup coil with 150uH/10Ohm, fT(LR)=10KHz.
Previous personal experience, related to this topic:
- use Charge sensitive preamplifiers for gamma proportional detectors (hence was valid the BW 100MHz which I mention, but I don't care about this BW in my present self-challenge). These products are made by ORTEC, Canberra, FAST and Nano. Good and very good results, depending on name.
- made Opamp-based TIA's as frontends for photodiodes as proportional gamma detectors. Pretty much textbook. Happy with very good results until 160dB(BW=200KHz).
That's my bit of previous TIA experience.
If I understand right, the 'standard' TIA has high Z_in and low Z_out? Maybe I am wrong.
Surely I am wrong, otherwise I would not write this here now:
This weekend I recognized&confirmed in measurements the TIA's behaviour on the A77's frontend for heads (their Repro board is a cheap one to buy as a start). As said, this board is planned for being the FE for my pickup coil.
Measurement results:
- SNR is proportional to R_f (the feedback R)
- was checked by removing all the NAB EQ feedback parts from this board and replace them by a R feedback.
But, I do not have a proper analysis for this topology how it does so for a low-Z source.
My confusion goes also at the fundamenal level: as it would contradict the TIA definition, according to my understanding. On the other side, it would fit the TIA definition according to Marcel
So far, so good. The planned approach is to push the R_f beyond 60dB, but I need to change the PS and the BC550 transistors. I do have good regulated dual PS for 120V. Should be enough for R_f=100dB and I hope to see the noise floor of the pick-up coil.
Nevertheless, extra help for a proper analysis of A77's repro-FE topology (only the first stage) will be very welcomed here by me, and I guess by many forum friends as well. Thanks again everyone for your time and kind help!
Why I want to build myself a good FE for small pick-up coils:
- I intend to use a 5um Tungsten wire, placing it in front of the REC and Play heads gap, one for each head, align them properly in microscope.
- so, for each head I will have only 1/2 turn (straight wire). The inductance of simple wire is expected to be in the 5nH range (ultimately, no coil to link more field and get more signal).
- I can model&simulate precisely this physical situation
- hopefully I will obtain a flux-meter calibration setup, equally precise, as I am tired to be a 'consumer' of MRL calibration tapes.
- this is also a nice hobby project for Easter holidays
Generally a voltage input is high-Z (though not required), a current input is low-Z (though not required), a voltage output is by definition low-Z and a current output is by definition high-Z
Therefore:
Voltage amps are high Z in, low Z out,
Transimpedance are low Z in, low Z out,
Transconductance are high Z in, high Z out.
Current amps are low Z in, high Z out...
There are of course exceptions. Low and high are relative to the source or load... Not all outputs are like voltage sources or current sources, for instance with RF we match impedance of output to next input and often use 50 ohms everywhere, so low and high all mean 50 ohms!
Therefore:
Voltage amps are high Z in, low Z out,
Transimpedance are low Z in, low Z out,
Transconductance are high Z in, high Z out.
Current amps are low Z in, high Z out...
There are of course exceptions. Low and high are relative to the source or load... Not all outputs are like voltage sources or current sources, for instance with RF we match impedance of output to next input and often use 50 ohms everywhere, so low and high all mean 50 ohms!
Yeah, but I know also a bit about the 50 ohm world
So, for proportional detectors the source was pretty high (does 1pF qualify as high?) and the output of TIA was pretty low (if 50Ohms qualify as low).
For photodiodes about the same thing. Both have TIAs, at least if my textbooks are right.
As for Play-head's (coil) FE, the source impedance is anything between 300Ohm and 15kOhm. The output is also in a range.
That's why I prefer to identify amplifiers by their best fit to unique equations and properties
In my eyes (not 100% factual or true) the FE in A77 repro board is very close to a TIA
So, for proportional detectors the source was pretty high (does 1pF qualify as high?) and the output of TIA was pretty low (if 50Ohms qualify as low).
For photodiodes about the same thing. Both have TIAs, at least if my textbooks are right.
As for Play-head's (coil) FE, the source impedance is anything between 300Ohm and 15kOhm. The output is also in a range.
That's why I prefer to identify amplifiers by their best fit to unique equations and properties
In my eyes (not 100% factual or true) the FE in A77 repro board is very close to a TIA
neh. I disagree.