Vintage Mic Transformer for MC step-up - gain-phase and impedance measurements

[Cqwet Dbdfte]

The phono pre-amps, with open inputs, have 3mV rms output noise and have 49.5dB gain at 1KHz, so ~69.5dB gain at 20Hz, to include passive RIAA network. If the pre-amps did not generate any noise, 1uV would the produce 3mV rms out, hence the susceptibility at the cartridge level.


On the turntable with 60Hz motor running, needle up, and 27dB transformers in line, I'm getting ~1.8mV and ~2.8mV of random noise, mostly low freq.

With this gain, the phono signal needs to be attenuated before hitting the power amps, with 15.5dB (voltage) gain.

At 20Hz the system would have (27+69.5+15.5 dB) 112dB gain, so attenuation is needed, which will also attenuate the noise.


I do not have a vinyl test disc ATM, so I cannot relate the noise to a cartridge signal level, other than noise in the speakers, where it s small enough not to be heard in a silent room, at "high" volume setting, unless I put my ear into the speaker element, not exactly scientific, but I'll get there.


The source of the 4dB extra noise in one channel I have to dig into, I suspect a noisier voltage ref in one of four current sources.

 

[Gold]

On the turntable with 60Hz motor running, needle up, and 27dB transformers in line, I'm getting ~1.8mV and ~2.8mV of random noise, mostly low frequency

2.8 mV is -48.28dBu. That isn’t very good.

 

[Cqwet Dbdfte]

Not a problem if you attenuate it 30dB.
Signal-to-noise is more interesting, as various preamps have different gain.

 

[Gold]

Not a problem if you attenuate it 30dB.
Signal-to-noise is more interesting, as various preamps have different gain.

I’m interested in the noise at the output of the phono pre. You said that is 2.8mV. That’s-48.28 dBu. A really terrible number. 30dB down is -78.28dBu. That’s good number but why apply all the gain on the front end to just throw it away. Just use 30dB less gain at the front end. It makes no sense. If you attenuate at the output of the line stage you are throwing away 30dB of SNR. Still makes no sense.

The line out of my phono pre is -70dBu. You should be able to match it. Perhaps the power supply diodes are bad. That should get you another 20dB SNR.

If you want to relate numbers to listening level then you need to state it in xx dB spl @ x meters.

 

[Cqwet Dbdfte]

With 70dB of gain I can toss 30dB and still have too much signal.
I checked the gain number, the Power amp looked a bit low, should be 23.75dB.

 

[Gold]

With 70dB of gain I can toss 30dB and still have too much signal.
I checked the gain number, the Power amp looked a bit low, should be 23.75dB.

What is your listening level in dB spl. I’m interested in real numbers. Not what you feel is sufficient. My monitor level is calibrated. I can tell you exactly how much gain is in the monitor amp to provide xxdB spl at 1M.

 

[Gold]

I hope you realize the that 30dB of gain you are thowing away has signal and noise. So you may have changed the noise floor but you haven’t changed the SNR which is still a dismal -48.28dB

 

[Cqwet Dbdfte]

Until I know what "signal" is I cannot make a number for SNR.
I will have a test disc monday. Meanwhile I put on a music disk, which sounded "pretty loud", connected my peak hold RMS DMM, and got a max hold of 4.173Vrms on the phono stage output.
So with a ~ 1.8Vrms noise this would be 67.3dB.
The power amps need 130mV for a 1W (2V@4Ohm).

 

[Gold]

You’re measuring p to p and not RMS? Thats really weird but I could adjust my numbers to that.

I predicted the numbers you would see before you saw them. i’d say the -55dBu channel is pretty close. The -48dBu is probably about 6dB off.

 

[Cqwet Dbdfte]

...no, rms peak hold on a slow DMM, like 3 samples per second.
The disc is from 1962, Sergio Mendes, "The swinger from Rio", stereo issue, first or second track has some loud saxophone passages.
Those levels on this this disc would suggest a 67dB range, with max +14.6dBu

 

[Gold]

Lets stick to test signals and concern ourselves with reference level to noise floor. Headroom or clip point isn’t a concern. Total dynamic range is. If we were splitting hairs okay. I’m getting -70dBu. You’re getting -55dBu.

 

[Cqwet Dbdfte]

...mmm, if I pad with 15dB is it not-67dBu?
I have yet not seen a reason to tweak the phono stage to lower the noise a few dB.
Without any further tweaks the output clipping level maxes out at ~ 140V p-p @1KHz (+36dBu)

 

[Gold]

Ok, my clip point is +27dBu. Adjust accordingly.

 

[MisterCMRR]

IMHO, all these numbers are "bogus" because no reference level is specified. A standard test is at a stylus tip velocity of 5 cm/sec, which is generally provided on test records as a 1 kHz tone. Unless you know the level of a standard reference level, any absolute noise level is meaningless. What matters is how far the noise level is below the reference signal. A meaningful spec would be something like "70 dB below 5 cm/sec output at 1 kHz, A-weighted."

This touches on one of my biggest pet peeves about marketing BS - unqualified test or performance numbers. They're worse than useless because they're misleading!

 

[Cqwet Dbdfte]

IMHO, all these numbers are "bogus" because no reference level is specified. A standard test is at a stylus tip velocity of 5 cm/sec, which is generally provided on test records as a 1 kHz tone. Unless you know the level of a standard reference level, any absolute noise level is meaningless. What matters is how far the noise level is below the reference signal. A meaningful spec would be something like "70 dB below 5 cm/sec output at 1 kHz, A-weighted."

This touches on one of my biggest pet peeves about marketing BS - unqualified test or performance numbers. They're worse than useless because they're misleading!

Well said, without a "signal" level, you cannot state the SNR.
I'm not "marketing" my phono preamp.

 

[Gold]

Why do think I have harped on having Standard Reference Level. I mentioned it at least three times before it sunk it that it was necessary. I also insisted no measurements are valid without a Standard Reference Level? At least three times.

II have specified my calibration at least three times. 1K @ 7cm/sec lateral = +4dBu. My noise floor is 74dB below reference or -70dBu

BTW 5cm/sec is 3dB down from 7cm/sec. If you have a consumer test record with 1k @ 5cm/sec lateral add 3dB.

 

[JohnRoberts]

IMHO, all these numbers are "bogus" because no reference level is specified. A standard test is at a stylus tip velocity of 5 cm/sec, which is generally provided on test records as a 1 kHz tone. Unless you know the level of a standard reference level, any absolute noise level is meaningless. What matters is how far the noise level is below the reference signal. A meaningful spec would be something like "70 dB below 5 cm/sec output at 1 kHz, A-weighted."

This touches on one of my biggest pet peeves about marketing BS - unqualified test or performance numbers. They're worse than useless because they're misleading!

Noise weighting for RIAA preamps is another variable. The RIAA equalization is effectively a LPF with the dominant pole at roughly 2kHz. This LPF besides rolling off HF noise, also tends to understate HF (harmonic) distortion since the higher harmonics would get attenuated by the EQ.

I found two tone IMD testing using 19kHz:20kHz @ 1:1 for a stimulus very revealing in my phono preamp design work. The 1kHz IM distortion product from that was not only NOT attenuated, it was boosted by the RIAA EQ another 20dB.

The old school THD+N specification could hide lots of HF intermodulation nonlinearity, creating SKUs that measured good but sounded bad.

JR

 

[Gold]

We are talking about a 15dB difference. When it’s closer we can start splitting hair.

If you really want to get into the weeds cm/sec is a peak to peak measurement. Not RMS. It has something to do with the math being easier in polar or cartesian. Way above my pay grade.

 

[Gold]

Well Mr. CMRR,

I am claiming that a balanced interface is beating an unbalanced interface by 15dB. The circuit I’m using is ‘cohen’ type front end. The exact circuit is at KA Electonics. I don’t know what the CMRR is. I’d think there is a way to see if my numbers add up, so to speak.

 

[JohnRoberts]

Well Mr. CMRR,

I am claiming that a balanced interface is beating an unbalanced interface by 15dB. The circuit I’m using is ‘cohen’ type front end. The exact circuit is at KA Electonics. I don’t know what the CMRR is. I’d think there is a way to see if my numbers add up, so to speak.

That sounds a little like my old P-10 phono preamp topology.

Flat Phono Preamp Based on John's P10 and 2SK389 - Pro Audio Design Forum

Here is link to a variant Wayne did over a decade later...

JR