current amplifying phono stage

[BrianL]

Scott,

Yes a current-feedback amp would be interesting. I was thinking
of that on the way to work this AM. The only problem I know
is that the brand I tend to have access to has pretty high
1/f noise corner which may not be the best for this application.
After all the parts are intended for high-speed applications
Maybe a MA-centric semiconductor company would have
something more suitable.

Now the only problem with trying this is I don't have a MC
cartridge. In fact, no working turntable at the moment.

Thinking about this some, I believe that J Peter Moncrief of
International Audio Review advocated a transimpedance
stage for MC preamps years ago. And didn't Richard Marsh
publish a design in TAA? So this is not an unheard-of concept.
(both were single-ended, not "balanced")

[john curl]

Briant, you are correct. This is an early idea, that we ultimately 'scrapped' over the years. I contacted EMT and Ortofon about optimized cartridge loading about 30 years ago. Their responses were vague. Also, AJ Vandenhul recommended low Z loading on EMT cartridges that he retipped.
As the years went by, optimum loading has tended to go to higher values, rather than lower values. Why, I don't know for sure.

[Elso Kwak]

Originally posted by EC8010
You really don't want to use a FET input op-amp in this application. As a broad generalisation, FET op-amps are dominated by voltage noise, whereas bipolars are dominated by current noise. When you use a moving coil cartridge as the source, its low source resistance means that the current noise hardly develops any voltage, meaning that bipolars are quieter. AD797 might be a good choice.

Hi,
Funny thing thing is I started with bipolars (MAT02) but ended with FETs (2SK389). Now you wouldn't believe it. The circuit did NOT get more noisy! Yes I know it is against all textbook explanations. Horowitz has several pages about this noise problem. I simply cannot corrobate it.

[EC8010]

Originally posted by Elso Kwak
The circuit did NOT get more noisy! Yes I know it is against all textbook explanations.

I'm amazed! It took a lot of searching, but I finally found some data on the device, and see it has a transconductance of only 20mA/V, even when you thrash it. That sort of number certainly doesn't sound promising. Did you measure the noise when you compared against MAT-02, or was it by listening? Details, details?

[Elso Kwak]

Originally posted by EC8010


I'm amazed! It took a lot of searching, but I finally found some data on the device, and see it has a transconductance of only 20mA/V, even when you thrash it. That sort of number certainly doesn't sound promising. Did you measure the noise when you compared against MAT-02, or was it by listening? Details, details?

Hi,
Picture simply taken from the datasheet. The MC-cartridge is connected between +IN and -IN with 100 Ohm resistors to ground from the inputs. So the dartridge leads are totally floating with respect to ground. I replaced the opamp by an AD817. Later went totally discrete. I did not suspect the cross-quad would work with the FETs but it did!
I judged the noise form the same high volume setting, no measurements available.
The sound was much better than the circuit I built with the SSM2017 which again was better than the Mark Levinson JC1-AC, to my ears.

[EC8010]

Interesting stuff. Thanks.

[john curl]

Folks, this is the situation:
Noise has several sources. One is base resistivity or Rbb'. This can be 2-400 ohms. The very best NPN transistors have about 4 ohms.
Then the Voltage noise generator. For transistors, it is .5/Gm = 13 ohms at 1ma, 1.3 ohms at 10ma. For fets, it is .7/Gm or.7/.035 = 20 ohms at 10ma.
There is also a current noise generator, which will be neglected here.
Elso has a 33 ohm resistor in series with his input, so it pads any differences between the 2SK389 and and pair of transistors. This difference will be small in this case. Now what if you ran the transistors at 10 ma? Then the transistor would be best, BUT it would be made more noisy than necessary by 1/f noise and added current noise.
Also, Elso's circuit 'might' add some noise from the current sources that are not completely degenerated for noise gain. I hope that this helps.

[Elso Kwak]

John,
You mean by 33 Ohm the Rg resistor? Is this one in series with the signal?
I do not remember the current I am using through the FETs or the transistors MAT02. I cranked it up a bit above the value of fig.6 by having more LEDs in series as a replacement for the 3.9V zener diode. I started with two or three LEDs ended with five.
All in one the noise is low enough for me as the record surface is FAR more noisy than the electronics. I hear the stylus ploughing through the groove!

[peranders]

Originally posted by john curl
Folks, this is the situation:
Noise has several sources. One is base resistivity or Rbb'. This can be 2-400 ohms. The very best NPN transistors have about 4 ohms.

Rbe, Rbb, Bulk resistance, I'm not 100% sure about the difference, but MAT03 has 0.3 ohms Rbe. Isn't Rbb some resistance in the base itself and Rbe is the resistance from Rbb down to the emitter output?

[john curl]

Elso, are you use using another value feedback resistor?