With matched JFETs, 2 transistors and two resistors would do. (And the noise is lower, too).
So which low noise PJFET do you intend to use ?
Available in SOT23 ?
And the Gnd wire does not need connection, or you need 3 wires ?
Cheers,
Patrick
https://www.trendsetter.com/v/vspfiles/assets/datasheets/Linear%20Integrated%20Systems/LSK489%20Application%20Note.pdf
Memory Distortion Philosophies - Part 4 : Circuits
bjt jfet CFP diff input stage
Bob Cordell, Lavardin, just old forgotten names...
Last edited:
I'm not claiming anything, it is claimed by itself and LtSpice, and of course a fet trz behaves as a resistor!...It was made with this purpose.The op-amp and a nested feedback just take care of this transistor's nonlinear behavior as it isn't such a perfect resistor.
As you could easily see, i used those fets with lower transconductance used in cassette players as switching parts in the signal path right after the tapehead as they have lower noise at lower currents, but higher RDSon which is much easier to be controlled by feedback.
Actually the simulations are better with lower transconductance transistors than like with bf862 and by the way...there's no noise spec of bf862 in the audio range...The only good thing about bf862 is that in a batch of 20trz, they all have the same specs(i measured a full lot of 25 bf862 and with identical source resistor they sink identical currents) which i believe can be interpreted as it might provide a more credible simulation model.
I have 4 different models for bf862 and i tried them all in my circuits, they show almost the same results as the op-amp is the one to take care of their nonlinearities anyway.
Simple truth is that i stayed over night until 6am listening to my player and it is a different player now, but the lower noise and perceived increased dynamics might be due to some other mods too... and this topic is not about player mods.I just wanted to show that the circuit works.
I have no idea if in a mc cart amp circuit bf862 would be better than a lower transconductance as 2sk117 as the latter has also lower noise at lower currents...
Last edited:
The original aim of my circuit was to put two op-amp sections in paralel as transconductance stages after a DAC so that i would have higher output drive and lower voltage noise.The offset was different in any two section of an op-amp so i needed a buffer to separate their inputs.
Simple truth is that i stayed over night until 6am listening to my player and it is a different player now, but the lower noise and perceived increased dynamics might be due to some other mods too... and this topic is not about player mods.I just wanted to show that the circuit works.
I have no idea if in a mc cart amp circuit bf862 would be better than a lower transconductance as 2sk117 as the latter has also lower noise at lower currents...
One thing i can tell you about one fet model: bss 135 (through hole version) must be very close to reality.I used this model for a preamp where i actually used bsp135 (smd version) and the simulations showed me a very narrow area where all the circuits parameters collapsed but they collapsed in a very predictable pattern within that region.When i tried it in the real world, that region had the same behavior just that the results in real life were a bit milder than the sim model which was a positive thing.I think that some of the models are actually very accurate and they are purposely modified for the worse so that the circuit won't fail miserably.Unfortunately i wasn't able to sim with the original bsp smd transistor model.My computer got nuts when tried it, probably because of the included thermal model not present in the through hole part model.
You have the circuit here:
1V2 phono riaa preamplifier
These was the first circuit where i used just the transistor transconductance to make the whole amplification , there's no resistor setting it , just the idle current is the one setting the amplification and LtSpice didn't lie.Well...it was a very complex interaction that asked for hundreds of simulations for it too...but in the end i was able to narrow everything to only to 2 working pathways that were very credible.
Simple truth is that i stayed over night until 6am listening to my player and it is a different player now, but the lower noise and perceived increased dynamics might be due to some other mods too... and this topic is not about player mods.I just wanted to show that the circuit works.
I have no idea if in a mc cart amp circuit bf862 would be better than a lower transconductance as 2sk117 as the latter has also lower noise at lower currents...
One thing i can tell you about one fet model: bss 135 (through hole version) must be very close to reality.I used this model for a preamp where i actually used bsp135 (smd version) and the simulations showed me a very narrow area where all the circuits parameters collapsed but they collapsed in a very predictable pattern within that region.When i tried it in the real world, that region had the same behavior just that the results in real life were a bit milder than the sim model which was a positive thing.I think that some of the models are actually very accurate and they are purposely modified for the worse so that the circuit won't fail miserably.Unfortunately i wasn't able to sim with the original bsp smd transistor model.My computer got nuts when tried it, probably because of the included thermal model not present in the through hole part model.
You have the circuit here:
1V2 phono riaa preamplifier
These was the first circuit where i used just the transistor transconductance to make the whole amplification , there's no resistor setting it , just the idle current is the one setting the amplification and LtSpice didn't lie.Well...it was a very complex interaction that asked for hundreds of simulations for it too...but in the end i was able to narrow everything to only to 2 working pathways that were very credible.
Last edited:
somebody said the same thing recently , but because of noise problems you need a certain minimum gain like 5x...10x.At that level the distortions are too low to be a problem for such gains.
I made a simple suggestion. You only need to add a .op command to your AC analysis, then View: Spice Error Log. You see no voltages and no Id, the FET is biased at the origin and is a resistor equal to 1/go. The model is correct 1/gds = 22 Ohms or so.
Name: j1
Model: bf862
Id: -8.97e-23
Vgs: 2.22e-18
Vds: -1.96e-21
Gm: -1.37e-22
Gds: 4.59e-02
Cgs: 8.20e-12
Cgd: 8.20e-12
Last edited:
If you can prove that all your money worries are over for the rest of your life. Nobel will beckon.
Jan
You're simply refusing to believe after i have uploaded you a movie on Youtube in real time.
I do understand that you are very high caliber design engineers, but i'm not baffled by the fact that design engineers are too lazy to solder a fet transistor, an op-amp and some resistors and capacitors because I AM A TEST TECHNICIAN!
I actually hold a TEST ENGINEER badge, but i got it in the UK where you don't need a university diploma to be considered an engineer as in the rest of the EU.
So just because others called me a test engineer for years , doesn't mean i have the legal right to consider myself one outside the UK.
I respect the fact that you are true valuable engineers, i will try to learn more about simulations if needed,but
Please respect the fact that I AM NOT a LIAR!
As I said, you are at loss for the moment.You could build and measure this simple circuit with your fancy analyzers that you have and i don't and draw a very simple conclusion .
I am still listening to my i/v converter which is fine without even a bias divider and that is because VIRTUAL ground is not REAL ground.
At the same time you should know that dac output is not a real zero ohm output either.So maybe i just got under the skin of an imperfect theoretical model.
This is my real job that i do for almost two decades, guessing how to fix electronic equipment without schematics and "solid knowledge".
I do understand that you are very high caliber design engineers, but i'm not baffled by the fact that design engineers are too lazy to solder a fet transistor, an op-amp and some resistors and capacitors because I AM A TEST TECHNICIAN!
I actually hold a TEST ENGINEER badge, but i got it in the UK where you don't need a university diploma to be considered an engineer as in the rest of the EU.
So just because others called me a test engineer for years , doesn't mean i have the legal right to consider myself one outside the UK.
I respect the fact that you are true valuable engineers, i will try to learn more about simulations if needed,but
Please respect the fact that I AM NOT a LIAR!
As I said, you are at loss for the moment.You could build and measure this simple circuit with your fancy analyzers that you have and i don't and draw a very simple conclusion .
I am still listening to my i/v converter which is fine without even a bias divider and that is because VIRTUAL ground is not REAL ground.
At the same time you should know that dac output is not a real zero ohm output either.So maybe i just got under the skin of an imperfect theoretical model.
This is my real job that i do for almost two decades, guessing how to fix electronic equipment without schematics and "solid knowledge".
Last edited:
That's a good trick, I asked you simply to remove the FET and put in a 22 Ohm resistor and observe the difference. I have no intention of building an I to V converter of any kind, I have no use for it. If you want help making a composite FET/op-amp I to V accept the help, everyone here is generous with it. Insisting that a FET connected as a resistor in series with an op-amp circuit has some special effect that we "have to listen to" to judge is not going to go too far.
BTW it is not productive to bring your anger and frustration here this is a place for DIY audio.
Well that's all a lot of prose, with no value for the discussion at hand. I asked a simple tech question and you spend 100-s of words talking around it.
Lets forget it, OK? You enjoy the sound of your thing, that's great!
Jan
I simply don't want to do it.
I just told that i have done this thing that could be applied with some reserves in an mc circuit and you made me a liar both here and in my topic telling that YOU KNOW that my circuit doesn't work.
Be free to believe what you want!
I used 4 models for bf862 available here and in other places.I also used the LT Spice included models for the op-amps(Lt1115 and AD8034), Then i built it and verified, now enjoying it.
Ask LtSpice engineers to do whatever you want to do.I don't get payed enough to argue with you.
Last edited:
It actually does in my dac circuit: it acts as a buffer to the capacitor placed at the dac output where i can even use a more complicated filter if i can deal with the phase response(not yet unfortunately), but that is just another discussion and i'm sorry that i've polluted this subject about the mc cart.
The real benefit of this circuit is in paralleling more input transistors and op amps for less input noise , higher total buffered capacitance for filtering dac output transients and higher output drive as i needed to get around the offset problem.
I was a bit tired yesterday and now i got your point.
Sorry!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.