My another proposal, more classic - with FB, THD=0.002879%
NOTE: R2 allow for gain adjust. If supplied from single 20V just put 1:1 divider on input, e.g two 200k resistors + 0.1u input cap.
NOTE: R2 allow for gain adjust. If supplied from single 20V just put 1:1 divider on input, e.g two 200k resistors + 0.1u input cap.
@Zbig , thank you very much for your great input. I will check your circuits within the next days. I have a lot of new ideas and will now simulate and test several approaches and try to understand what's best for my requirements. Currently I'm thinking about the dual mu-follower idea, because than the second stage provides inverting as well.
Much higher gain, indeed. So I added second pnp in differential configuration. It took a while to adjust RIAA network, but within 1dB difference it is done ((out, green line).
At the end I've added aux (red line) output separated from feedback. Out has +1dB at 20kHz while Aux has -1dB at 20kHz
Preamp has k=12 at 1kHz.
Last edited:
Is this only simulation, or you also have measurements ?
And you have a reliable source for 2N3819 ?
Thanks,
Patrick
It was mainly simulations, except npn amplifier with grounded emitter, which I've built and mesaured.
JFET (2N3819) was randomly selected from LTSpice library.
JFET (2N3819) was randomly selected from LTSpice library.
For my own education, I simulated the circuit in post #100.
I added an output decoupling cap of 10uF, a realistic load of 10k, use BC327-40 as PNP, and apply an input of 5.6Vpp (or 2Vrms).
Results are as follows :
Gain = -0.996
H2 = 0.0489%
H3 = 0.0210%
THD = 0.0558%
Patrick
I added an output decoupling cap of 10uF, a realistic load of 10k, use BC327-40 as PNP, and apply an input of 5.6Vpp (or 2Vrms).
Results are as follows :
Gain = -0.996
H2 = 0.0489%
H3 = 0.0210%
THD = 0.0558%
Patrick
Last edited:
Seems that you've reached maximum input voltage. J1 drain current goes close to 0mA, and J1 drain-source voltage goes below saturation voltage.
It is too much for this amplifier, compensation wont't work in such conditions
Last edited:
Nope. Highest value (+4dB) is only 3.472V peak-to peak or 1.228 Vrms, See wiki
Change R2 into 12k and it will accept input 4Vpp with THD=0.005%
Change R2 into 12k and it will accept input 4Vpp with THD=0.005%
Last edited:
0.005% is only when the load is >>10k.
For 10k load, it is more like 0.024%.
In any case, results in spice is normally more optimistic.
And 2N3819 is obsolete, at least the Onsemi / Fairchild variant.
Cheers,
Patrick
For 10k load, it is more like 0.024%.
In any case, results in spice is normally more optimistic.
And 2N3819 is obsolete, at least the Onsemi / Fairchild variant.
Cheers,
Patrick
2N3819 was picked randomly. If you have model for other more common JFET send me PM.
Compensation requires high impedance to keep balance at full swing, when drain current becomes very low.
It will work perfectly with JFET/MOSFET next stage.
Compensation requires high impedance to keep balance at full swing, when drain current becomes very low.
It will work perfectly with JFET/MOSFET next stage.
While the "aux" output provides the inversion the OP wanted, the DC bias for the aux OP will vary a lot with FET characteristics. Getting the right AC and DC on two outputs usually requires two feedback loops. If P-channel FETs and/or bjts are allowed, you can make a decent discrete op-amp. Otherwise, I think a Rs=Rd phase splitter with a FET follower is the way to go.