The review of the "DIY" Pass ACA Class A Power Amplifier reminded me that the design of my DIY preamp had been influenced by Nelson Pass, in 1997. I had to rely on subjective listening due to missing measuring equipment, and the first version of the line stage with just an opamp sounded "sterile".
While searching the internet I came upon NPs DIY pages and got the idea to use a MOSFET as a simple class A voltage follower behind the opamp, but within the feedback loop. This allowed me to set the bias current (R43) so high that the output could directly feed my Sennheiser HD580 (300 Ohm). I liked the sound and stayed with it.
Here is the schematic:
Notes:
Since the THD of the ACA Class A power amp was so incredibly high I just had to check the behaviour of this line stage. I used the RME ADI-2 PRO fs to feed the input and catch the output. I really wanted to know how much negative influence the MOSFET would have.
Let's start with THD at 2Vrms in and volume set for 2Vrms out:
Oops, not too bad! THD is quite good, seems the MOSFET is not a problem. Noise is dominated by PS hum (despite the overkill PS ) - may have to check the power supply caps or the cables (there are just too many ...). Note that the preamp was connected to my hifi system - I was too lazy to take it out.
Same measurement but volume set for output voltage of +20 dBu = 7.75 Vrms, just before clipping:
THD is some 10 dB worse, could be caused by common mode distortion. Noise floor looks cleaner now.
Same voltages for in/out, this time measuring IMD CCIF (19+20 kHz):
IMD itsself is quite OK at -110 dB, but there is a lot of stuff which shouldn't be there (PS noise).
Next is multitone:
Again we see the PS noise and distortion (or noise?) at higher frequencies.
Frequency response:
Frequency response is 20 Hz to 20 kHz +0/-0.08 dB. -3dB limit is at 100 kHz but may be limited by the RME although I used 786 kHz samplerate. Anyhow more than good enough for my aging ears.
Finally a 1 kHz square signal, generated by the RME at 768 kHz samplerate and captured by REWs scope:
And here zoomed in at the transition from low to high:
The slight overshoot is not problematic, it's not critically damped but not really ringing either. The cable was very short though.
Conclusion: a very simple line stage which can feed a high impedance head phone (300 Ohm) with around 200 mW power (not measured).
While searching the internet I came upon NPs DIY pages and got the idea to use a MOSFET as a simple class A voltage follower behind the opamp, but within the feedback loop. This allowed me to set the bias current (R43) so high that the output could directly feed my Sennheiser HD580 (300 Ohm). I liked the sound and stayed with it.
Here is the schematic:
Notes:
- chances are high that I replaced the OPA2134 by a LM49710 when I modified the phono stage a few years ago - can't remember, getting old
- the input is fed via a 50 kOhm Alps pot (not shown)
- the power supply (not shown) uses double linear voltage regulators (totally overkill if you ask me today, but in 1997 I was influenced by the big Naim power supplies for their highend preamps ):
- 2 * 24V toroid transformers in a separate housing
- rectified and regulated to +/- 24V with 7824/7924 (also in the separate housing)
- regulated to +/- 15V with LM317/337, per channel (on the PCB)
Since the THD of the ACA Class A power amp was so incredibly high I just had to check the behaviour of this line stage. I used the RME ADI-2 PRO fs to feed the input and catch the output. I really wanted to know how much negative influence the MOSFET would have.
Let's start with THD at 2Vrms in and volume set for 2Vrms out:
Oops, not too bad! THD is quite good, seems the MOSFET is not a problem. Noise is dominated by PS hum (despite the overkill PS ) - may have to check the power supply caps or the cables (there are just too many ...). Note that the preamp was connected to my hifi system - I was too lazy to take it out.
Same measurement but volume set for output voltage of +20 dBu = 7.75 Vrms, just before clipping:
THD is some 10 dB worse, could be caused by common mode distortion. Noise floor looks cleaner now.
Same voltages for in/out, this time measuring IMD CCIF (19+20 kHz):
IMD itsself is quite OK at -110 dB, but there is a lot of stuff which shouldn't be there (PS noise).
Next is multitone:
Again we see the PS noise and distortion (or noise?) at higher frequencies.
Frequency response:
Frequency response is 20 Hz to 20 kHz +0/-0.08 dB. -3dB limit is at 100 kHz but may be limited by the RME although I used 786 kHz samplerate. Anyhow more than good enough for my aging ears.
Finally a 1 kHz square signal, generated by the RME at 768 kHz samplerate and captured by REWs scope:
And here zoomed in at the transition from low to high:
The slight overshoot is not problematic, it's not critically damped but not really ringing either. The cable was very short though.
Conclusion: a very simple line stage which can feed a high impedance head phone (300 Ohm) with around 200 mW power (not measured).
Last edited: