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.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

THD "Measurements"

Post by mediatechnology »

I promised to post some THD "measurements."
"Measurements" are in quotes - it's more like THD interpretation.


The measurement chain is:

Soundcard Out > THAT1646 > -60dB U-Pad > ZTX851 Balanced Preamp > PCM4222EVM (96k/24b) >AudioTester
The DUT gain is 62 dB; PCM4222EVM loading pads the output by just under 0.7dB
Because there is 6 dB gain in the THAT1646, the soundcard output is set 6 dB lower.

Rsource is 10Ω; the front-end op amp is an NE5532.

For the FFTs which show the "Generator Monitor" the Right Channel Flow is Soundcard Out > PCM4222EVM.

There is a 6 dB loss through this path due to the high 560Ω output impedance of the soundcard and the same (about) 560Ω termination of the PCM4222EVM.

When the generator output level is reduced by -6dB along with the DUT channel so that they level match, there is a 12 dB offset in overall level seen at the PCM4222EVM inputs.

The gain offset, though not intentional or ideal, does let us see, at different levels, the overall THD signature of the D/A and A/D.

Though not the best, I use what I have and I'm sure the explanation makes everything about as clear as mud.

Having said all of that here are the actual measurements. The first two are the overall THD of the preamp and measurement system.

Image
ZTX851 Moving Coil Preamp THD 62dB Gain

Note that no effort was made to null the induced hum radiating into the ProtoBoard.

Image
ZTX851 Moving Coil Preamp THD 62dB Gain 5 kHz

The next images show the THD signature of the D/A and A/D loop-through in the Right (red) channel.

Image
ZTX851 Moving Coil Preamp THD 62dB Gain 5 kHz vs GenMon

The THD signature at -10 dB.

Image
ZTX851 Moving Coil Preamp THD 62dB Gain 5 kHz vs GenMon at -10dB

The PCM4222 THD starts to head north out of the noise floor around -20 dBFS which corresponds to about -5 dBu.
Its also noteworthy to point out that if something measures lower than the source it's contributing THD by subtracting distortion from the source.
It's one of those cases where "less equals more."

What I don't know is if the slight rise in THD that I'm seeing is the preamp or converter.
The reality is is that its the sum of everything.
That's why this is a THD "interpretation."

What I confirmed is that the preamp's THD contribution, based on what I've seen, is low enough to justify laying out a board.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

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

Post by JR. »

Sweet... I'm pretty sure I can't hear that.

JR
Cancel the "cancel culture", do not support mob hatred.
KMN
Posts: 53
Joined: Sat Jan 20, 2018 2:49 pm

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

Post by KMN »

First, really nice job. Whether by wank, luck or unstated genius, that thing is quiet.

From the schematic symmetry, this design looks to be a candidate for a "humbucker" board layout. My new word from another thread here on proaudiodesignforum.

Please bare with my attempt to describe what I'm talking about, and I'll apologize ahead of time since I don't know how to describe this clearly...

If the circuitry for the plus and minus halves of the balanced signal were shaped identically but for example plus circuitry was on top of the board while minus circuitry on the bottom such that everything lines up so that mirrored components and traces from the plus and minus were physically stacked directly above and below from one another...incident fields would produce only common mode signal.

The idea is to try and reduce total loop area swept by the balanced circuitry and it's ability to receive differential noise.

Does anybody here have any experience whether doing this sort of layout would introduce any real world performance advantages for the effort in a high gain, possibly high impedance, balanced front end?
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

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

Post by JR. »

KMN wrote: Tue Mar 26, 2019 12:11 pm First, really nice job. Whether by wank, luck or unstated genius, that thing is quiet.

From the schematic symmetry, this design looks to be a candidate for a "humbucker" board layout. My new word from another thread here on proaudiodesignforum.

Please bare with my attempt to describe what I'm talking about, and I'll apologize ahead of time since I don't know how to describe this clearly...

If the circuitry for the plus and minus halves of the balanced signal were shaped identically but for example plus circuitry was on top of the board while minus circuitry on the bottom such that everything lines up so that mirrored components and traces from the plus and minus were physically stacked directly above and below from one another...incident fields would produce only common mode signal.

The idea is to try and reduce total loop area swept by the balanced circuitry and it's ability to receive differential noise.

Does anybody here have any experience whether doing this sort of layout would introduce any real world performance advantages for the effort in a high gain, possibly high impedance, balanced front end?
I had a design engineer who worked for me last century who did exactly that to improve hum rejection in critical circuit traces inside top box powered mixers that could have large magnetic fields (big transformers in small packages). With single sided PCB it could be a struggle involving jumpers and other gyrations. FWIW it worked, he had a name for it that I don't recall.

JR

PS: Star quad shielded wire uses a similar approach where a two tightly wound wire pairs maintain very good symmetry and small loop area, the alternating pairs cancel fields and appear to exhibit zero loop area. This also works.
Cancel the "cancel culture", do not support mob hatred.
KMN
Posts: 53
Joined: Sat Jan 20, 2018 2:49 pm

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

Post by KMN »

Thank you for replying, JR. I figured this is one of the few places where someone might have an opinion or first hand knowledge of such things as they relate to audio.
User avatar
terkio
Posts: 304
Joined: Thu Feb 07, 2019 3:17 am
Contact:

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

Post by terkio »

To minimize loop area this is an interesting way, however there are other objectives.
Some components better be on the same side of the PCB for better thermal matching. It is likely there is less temperature gradient for components close on the same side rather than closer but on opposite side.
Practical considerations too.
_Some components must be on the same side, here the two transistors.
_Some components are dual, cannot be split.
_The foot print of components like op amps, don't mirror when on opposite sides, they rotate.
_Through hole components cannot be exactly on top of the other, there must be some offset.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Out-of-band and Flux Loop Tests

Post by mediatechnology »

Minimizing loop area is key.
At the end of the day my hunch is normal differential mode hum induced into the cart will be the greatest source.

I have a layout started on my desktop but have been waiting on some rainy weather to start it.
Rather than do layout I mowed about 1/3rd of an acre with a push mower and re-cut maybe 1500-2000' feet of trail through the remaining forested acre with a much-cheaper push mower.
After a recovery day I ran errands.
Slowly, as the weather grew more dreary and cloudy I returned to the preamp.

What I did do was re-examine some previous conclusions about "oscillation" with an inductive source. See: https://proaudiodesignforum.com/forum/p ... 73&#p11373

I also looked at out-of-band response to RF.

The first test I did was to use a Grid Dip Oscillator and see if I could inject RF into the circuit and see peaking or DC shift in the servo indicating rectification.

The source impedance was a 10Ω resistor.
I was looking to see if radiation into any of the circuit nodes made it take off.
It didn't.
The Grid Dip Oscillator only goes down to 1.6 MHZ where I did see a rising frequency response.

I decided to see if I could make it oscillate again with an inductive source.

I chose 20 µH + 10 Ω.
I couldn't make it oscillate under any conditions.
What I did find, and what made me lead to the wrong conclusions, is that the input becomes resonant.
With certain combinations the input resonated with "hash," at multiples of Fsample, with the PCM4222EVM being used to measure the DUT's in-band performance.

All I was doing was shifting the input resonant frequency which, as it accentuated Fs as it was tuned, made me believe it was oscillating.

It wasn't.

Disconnecting the A/D allowed me to see what was really going on.

I used a function generator and flux loop to inject differential signal into the input inductors and looked for resonance.

I found resonance around 500 kHz.

The obvious thing is that termination capacitance at the values one would expect with a MC input (1 nF) makes no real difference.
It may have value to suppress RF, but it doesn't significantly affect tuning or Q.

Termination resistance makes a huge difference in out-of-band resonance.

Source = 20µH + 10Ω.

Rterm = 499Ω X2: Fo 567 kHz Q=4.75

Rterm = 200Ω||998Ω: Fo 526 kHz Q=1.4

Rterm = 100Ω||998Ω: Fo 470 kHz Q=0.9

Making Rsource 0Ω + 20 µH raises the Q with 998Ω loading to about "5."

For a low inductance MC cartridge (and therefor low output) an unterminated input seems like a very, very bad idea because it can resonate out-of-band.
You can make a 103 run shrill & peaky by loading it with 2-3k ohms resistive, or 80-160ohms on a transformer. Similarly loading a 103R with 20-45ohms resistive or about 9 on a transformer will roll it off FAST. https://www.head-fi.org/threads/denon-d ... st-6016153
I can believe that. I can make a MM shrill and peaky the same way by under-loading it with C.

I'll leave a 5 mm space on the PCB for a 1 nF film C for problematic RFI situations but its not required for stability.
User avatar
terkio
Posts: 304
Joined: Thu Feb 07, 2019 3:17 am
Contact:

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

Post by terkio »

by mediatechnology » Sun Mar 24, 2019 8:07 am
The measurement chain is:

Soundcard Out > THAT1646 > -60dB U-Pad > ZTX851 Balanced Preamp > PCM4222EVM (96k/24b) >AudioTester
Thanks for your detailed presentation of the THD measurements.
How do you connect the PCM4222EVM (96k/24b) to the PC ?
How does the Audio Tester software see the PCM4222EVM, does it need a driver ?
billshurv
Posts: 436
Joined: Thu Feb 01, 2018 6:07 am

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

Post by billshurv »

It's SPDIF/AES out so you need a soundcard with that in. I have on the bench waiting for a round tuit* an maudio microtrackII that needs a battery replacement. The idea with that is that I can do battery based logging of stuff if needed then process offline. Also less of a pain than balancing the laptop on the piano!

* English joke, explained in attached image.
Attachments
a-round-tuit.jpg
billshurv
Posts: 436
Joined: Thu Feb 01, 2018 6:07 am

Re: Out-of-band and Flux Loop Tests

Post by billshurv »

mediatechnology wrote: Mon Apr 01, 2019 9:41 am

I have a layout started on my desktop but have been waiting on some rainy weather to start it.
Rather than do layout I mowed about 1/3rd of an acre with a push mower and re-cut maybe 1500-2000' feet of trail through the remaining forested acre with a much-cheaper push mower.
After a recovery day I ran errands.
Slowly, as the weather grew more dreary and cloudy I returned to the preamp.
Sounds like you need some goats!
Post Reply