FET / NE5534 Questions

[Joe Viau]

Hello:

I redrew a schematic for a preamp that I saw, as the schematic was fuzzy in areas.  I have made some observations and have a few questions.





It is known that the circuit is dealing with a low impedance signal.  The two FETs are labeled as one component, which leads me
to believe that the FETs are a matched dual FET device of some type--2n5515, NTE461 or something like that.

Questions:

- What are the two FETs doing? Is it a differential amp?

- If I were to match FETs myself, could I use RG's FET Matcher? I can understand how it can match up FETs for more uniform phasing in a phase shifter, but in this application the FETs are (maybe?) used as amplifiers.  If the matcher could not be used, how would one match them?

- What low-noise FET would lend itself to this application? It seems to me that low noise trumps other considerations, at least to my limited knowledge.  If not so, I would be happy to be corrected.

- I notices on the 5534 that there is a 1M resistor, used for compensation(?)--one leg goes to V-.  Does the other leg go to Balance, Compensation, or Balance / Compensation? This configuration is used in other areas of the circuit and all of the other 5534's have this resistor as well.

Thanks in advance for any and all replies.

-- Joe

[Joe Viau]

+1 bump

[Joe Viau]

Bump...

[Nasse]

A JFET can be made to act as a current source by tying its gate to its source. The current then flowing is the IDSS of the FET

I really dont know much just googled it

[PRR]

> +1 bump
> Bump...

It's a partial circuit, a re-trace of something fuzzy, of unknown origin, goals, design strategy.

I would at least want to know where the output went, what comes to the "+ in", and why "+" not "-".

Unless the missing part off to the right is very odd, it probably isn't going to work. The 5534 needs a compensation capacitor for most useful gains.

> resistor, used for compensation(?)

No. Skim the datasheet figure 15 and figure 4. Compensation is a capacitor pin 5 to pin 8. For high closed-loop gain it may be omitted, at unity gain you may use 22pFd.... but in any music system you should use more like 47pFd.

I have no idea what the 1 Meg is supposed to be. It isn't offset adjustment because that must fade from pin 1 to pin 8. If it is a DE-offset trim, why the fancy-pants with matched FETs?

> It is known that the circuit is dealing with a low impedance signal.

Then I wonder why the FETs. The 5534 is an excellent low-noise amp for low and medium impedances. Most FETs will degrade the noise voltage.

> What are the two FETs doing?

It's a fancy Cathode Follower. Yeah, on an FET we call it a Source. Unity gain, high input Z, lowish output Z.

A simple CF uses a cathode resistor and has a small but uncertain DC offset. If you replace the resistor with an identical FET biased zero-volt, the top FET operates zero-volt and the offset is zero. Well, kinda. There ARE better ways to do this, at least for audio systems. And in stage-amp work, who cares what the DC offset is? We can always cap-couple it away.

> match FETs myself

Tie gate to source. Put FET, 9V batt, and current meter in series. Pick two with the same current (if you can).

If not well matched, then the Source is at maybe +0.1V or -0.3V when the Gate is at zero. In a DC amplifier, this may matter. In audio it rarely matters.

> low noise trumps other considerations

Rarely. I'd suggest that easy-to-understand (and debug!) is THE design goal, until you reach a point that "everything is easy".

[JDoyle]

The JFETs serve as a buffer - if the JFETs are matched, then the output from the JFETs will bias to ground and subsequently bias the 5534 to ground as well. Using a current source as a load for a source follower allows the designer to attain a gain closer to the ideal of 1.

I don't believe that low-noise was the main design goal of the circuit because the op amp is used in an inverting configuration which has 14dB more of noise than the noninverting configuration.

The 1 Meg resistor is a mystery to me as well.

Regards,

Jay Doyle

[Joe Viau]

Thanks for the replies, they are helping me.

The circuit was drawn with op amps.  I found that they were NE5534 but the schematic does not show which input is inverting or non-inverting... in other words, the op amp is drawn as a triangle pointing right, with two inputs.  Is there a convention used when you don't indicate which input is which?

To let you know, this is part of the circuit for the on-board electronics used in Alembic instruments.  The input is from a low (to middle) impedance pick-up, with the output going to a resistor network that has a gain and hum balance control.  There is a neck and bridge pickup, along with a dummy coil that also has a pre-amp.  the neck and bridge pre-amps each go to what I think is a Sallen-Key low pass filter.  Alembic says that the rolloff is 15db with a frequency range of 350 Hz to 6 KHz.  The 2.2M resistor at the start of the circuit was in the schematic was the value was not indicated.  I got the value from a Dutch article written about Alembic basses.  They attempted to reverse engineer the bass but some of the stuff is wrong.  The inverting input that the article shows could well be wrong.

I am hoping to dump the schematic into something like ExpressSCH to clarify its workings.  I do not have experience with SPICE but maybe now wouldn't be a bad time to start learning this as well.  Any recommendations about tools would be welcome as well.  Eagle would be another possibility.  I like ExpressSCH for simple circuits.

Thanks again.

[JDoyle]

The circuit was drawn with op amps.  I found that they were NE5534 but the schematic does not show which input is inverting or non-inverting... in other words, the op amp is drawn as a triangle pointing right, with two inputs.  Is there a convention used when you don't indicate which input is which?

There isn't a convention but looking at the overall circuit design, I'm 99.9% certain that the inputs should be switched.