A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

Where we discuss new analog design ideas for Pro Audio and modern spins on vintage ones.
Hans
Posts: 62
Joined: Fri Jun 24, 2016 8:48 am

Re: A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

Post by Hans » Tue Sep 18, 2018 5:06 am

Here some additional info taken from the circuit.
The only deviation from Wayne's design is the 1uF cap before the servo, because a 0.47uF in combination with the 100nF in the voltage divider caused a bump in the FR between 10Hz and 200Hz.

Hans

Wayne Circuit_2.jpg

User avatar
mediatechnology
Posts: 3275
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

Post by mediatechnology » Tue Sep 18, 2018 5:18 am

Thank you for doing those sims Hans.

What happens when Rg (R1) is made 1Ω with an Rspurce of 1Ω?

I hadn't gone looking yet for the resonance bump with 100 nF/0.47µF.
How big was the bump?

Hans
Posts: 62
Joined: Fri Jun 24, 2016 8:48 am

Re: A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

Post by Hans » Tue Sep 18, 2018 6:32 am

With 0.47uF the bump is close to 0.2dB@50Hz and with 1uF below 0.1dB@50Hz, not really a problem, but less elegant.

With Rg as well as Rsource being 1 Ohm, Noise from 20 Hz to 20 Khz is 49,3 nV, as opposed to 52.6 nV with Rsource = 1 Ohm / Rg = 2 Ohm
A single 1 Ohm resistor produces 18.2 nV from 20 Hz to 20 Khz.

In case of Rg = 2 Ohm, the circuit is 9.2 dB noisier as just Rsource,
and in case of Rg = 1 Ohm the difference is 8.7 dB, an improvement of only 0.5 dB.

Hans

User avatar
mediatechnology
Posts: 3275
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

Post by mediatechnology » Tue Sep 18, 2018 7:41 am

Since this is a model and you have the luxury of perfect devices that we don't have in the real-world what happens if you model it with the ZTX851 base resistance of 0Ω?

Notwithstanding your concerns about the differential resistance of the OPA1612 how does it model at 1Ω Rsource in the first socket?

Also what does the simulator tell us for 3.3Ω?
Self gets a 6.4 dB NF at 3.3Ω with his 3X 2N4401 MC preamp.

We get a simulated 9.2 dB NF at 1Ω with 2X ZTX851 differentially.

Hans
Posts: 62
Joined: Fri Jun 24, 2016 8:48 am

Re: A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

Post by Hans » Tue Sep 18, 2018 9:10 am

Since this is a model and you have the luxury of perfect devices that we don't have in the real-world what happens if you model it with the ZTX851 base resistance of 0Ω?
Remember my Sim right at the beginning of this thread? Your measured results were even better.
My experience with this kind of circuits is that Simulations are very close to the real world indeed.

Could you explain in some detail what sense it could make to change the base resistance of the ZTX into 0 Ohm?
I do not object in trying this, but I see no real purpose.
Also what does the simulator tell us for 3.3Ω?
Self gets a 6.4 dB NF at 3.3Ω with his 3X 2N4401 MC preamp.
Which one should be 3.3 Ohm, Rg or Rsource, and what should be the value for the other one ?
We get a simulated 9.2 dB NF at 1Ω with 2X ZTX851 differentially.
That is what I mentioned for Rg=2 Ohm. For Rg=1 Ohm it is 8.7 dB.
Notwithstanding your concerns about the differential resistance of the OPA1612 how does it model at 1Ω Rsource in the first socket?
Rg + Rsource 1 Ohm: 46.7 nV
Rg 2 Ohm / Rsource 1 Ohm: 50.2 nV

However, because Vref from the voltage divider has been lowered by the load of the Opa1612's input to 4.4 Volt, collector current went up from 5.5 mA to 6.6 mA, resulting in less noise from the ZTX, so it's not a fair compare.

When adjusting the top two resistors from 10K to 6K8 to restore the 5.5 Volt, noise figures are:
Rg + Rsource 1 Ohm: 49.7 nV (versus 49.3 nV for the Opa1642)
Rg 2 Ohm / Rsource 1 Ohm: 53.0 nV (versus 52.6 for the Opa1642)

Hans

Hans
Posts: 62
Joined: Fri Jun 24, 2016 8:48 am

Re: A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

Post by Hans » Tue Sep 18, 2018 9:20 am

Wayne

You may be surprised that the Opa1612 is producing a tad more noise, despite its lower Vnoise of 1.1nV/rtHz.
The reason however is of course the added current noise of 1,7 pA/rtHz injecting into the 1K collector resistor and also into the Vref divider.

Hans

User avatar
JR.
Posts: 2418
Joined: Sat May 24, 2008 7:21 pm

Re: A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

Post by JR. » Tue Sep 18, 2018 10:06 am

Hans wrote:
Tue Sep 18, 2018 9:10 am
Since this is a model and you have the luxury of perfect devices that we don't have in the real-world what happens if you model it with the ZTX851 base resistance of 0Ω?
Remember my Sim right at the beginning of this thread? Your measured results were even better.
My experience with this kind of circuits is that Simulations are very close to the real world indeed.

Could you explain in some detail what sense it could make to change the base resistance of the ZTX into 0 Ohm?
I do not object in trying this, but I see no real purpose.
In circuit design zeroing out a term allows you to better understand the significance of different noise sources, by how much the total changes. Generally we do not have the option to experiment inside active devices.

JR
Also what does the simulator tell us for 3.3Ω?
Self gets a 6.4 dB NF at 3.3Ω with his 3X 2N4401 MC preamp.
Which one should be 3.3 Ohm, Rg or Rsource, and what should be the value for the other one ?
We get a simulated 9.2 dB NF at 1Ω with 2X ZTX851 differentially.
That is what I mentioned for Rg=2 Ohm. For Rg=1 Ohm it is 8.7 dB.
Notwithstanding your concerns about the differential resistance of the OPA1612 how does it model at 1Ω Rsource in the first socket?
Rg + Rsource 1 Ohm: 46.7 nV
Rg 2 Ohm / Rsource 1 Ohm: 50.2 nV

However, because Vref from the voltage divider has been lowered by the load of the Opa1612's input to 4.4 Volt, collector current went up from 5.5 mA to 6.6 mA, resulting in less noise from the ZTX, so it's not a fair compare.

When adjusting the top two resistors from 10K to 6K8 to restore the 5.5 Volt, noise figures are:
Rg + Rsource 1 Ohm: 49.7 nV (versus 49.3 nV for the Opa1642)
Rg 2 Ohm / Rsource 1 Ohm: 53.0 nV (versus 52.6 for the Opa1642)

Hans

User avatar
mediatechnology
Posts: 3275
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

Post by mediatechnology » Tue Sep 18, 2018 10:30 am

Could you explain in some detail what sense it could make to change the base resistance of the ZTX into 0 Ohm?
I do not object in trying this, but I see no real purpose.
Because it tells us what all our other noise sources add up to and establishes the ultimate minimum.

If we were to parallel additional ZTX851 it couldn't get any better than the noise with the base resistance at 0Ω.
0Ω base resistance establishes our floor.
By scaling the base resistance by 1/2 in the model you could quickly see how paralleled devices work.

I was curious what NF a 3.3Ω Rsource would have (with Rg=2Ω) in the current model.
That way we have sims at 1, 3.3 and 10 Ohms with Rg at 2Ω.

I just looked at the Protoboard to see what I was using for the input op amp when I made those last measurements: It's a 5532.
IIRC a NJM2068 measures the quietest but requires a Cc of about 300 pF.

BTW I don't see the effect of differential input impedance loading the bias network. I think there's an error in the model. In the actual circuit, the op amp is going to hold the inputs in DC null. If the non-inverting input is at +5.5V, the inverting input follows and is also at 5.5V. (Neglecting offset errors.) The 20KΩ differential resistance of the op amp is bootstrapped - there's no (significant) current flow and it doesn't load the network. This is where sim separates from reality. FWIW I just fired up the Protoboard and verified DC reality.

Measure the DC voltages at the inverting and non-inverting inputs and tell me if they're the same in your model. They would almost have to be. If so, where's the current flow in the 20KΩ Rdiff?

You have to understand Hans that the last set of tests I did were after adding the servo but before I set this thing aside for a few months.
I wasn't focused on noise at that point other than nulling out all the induced measurement hum.

Just by varying the loop area of Rsource (an untrimmed resistor) I could make drastic changes in the hum signature.

My true lack of interest in noise performance during that last round of tests is clearly evidenced by a 5532 I just found on the Protoboard which isn't that bad but not as good as an OPA1612 or NJM2068.

Hans
Posts: 62
Joined: Fri Jun 24, 2016 8:48 am

Re: A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

Post by Hans » Tue Sep 18, 2018 11:08 am

I will comeback to your questions tomorrow.
By now I will have to join my rowing team.

Hans

Hans
Posts: 62
Joined: Fri Jun 24, 2016 8:48 am

Re: A Low Noise Balanced Input Moving Coil Preamp Using the ZTX851

Post by Hans » Wed Sep 19, 2018 3:21 am

Here is the answer to your first question concerning the ZTX851, noise specified from 20Hz to 20 Khz:

Rsource 1 Ohm/Rg 1 Ohm:
Noise Free Transistor: 48.9 nV
2 ZTX851 : 49.1 nV
1 ZTX851 : 49.3 nV

Rsource 1 Ohm/Rg 2 Ohm:
Noise Free Transistor: 52.3 nV
2 ZTX851 : 52.5 nV
1 ZTX851 : 52.6 nV

Hans

Post Reply