Parasound Zphono Phono Preamplifier Review

[amirm]

This is a review and detailed measurements of the Parasound Zphono phono stage. It is on kind loan from a member and costs US $200. There are also versions with USB and ADC (for ripping Vinyl) and higher-end version with more control.

The look is unmistakably Parasound which is to say somewhat industrial and plain:

The back panel shows the included, voltage selectable AC mains which I appreciate:

As you see the input gain can be changed from moving magnet (MM) to moving coil (MC).

Interesting to see an AC mains polarity switch. Not sure of the safety of that but I guess if you have a hum, it is worth having a switch like this to at least troubleshoot the problem.

Overall, the Zphono is a business-line phono amplifier with solid construction.

Phono Stage Audio Measurements
Let's start with our usual 1 kHz dashboard view with moving magnet setting:

As with all good phono stages, there is no visible distortion. So what sets SINAD is simply noise which in this case rises enough to set it to 75 dB. This puts the Zphono in the middle of the pack:

Moving Coil setting with input changed to 0.8 millivolts naturally degrades performance due to increased noise that goes with increased gain:

Since LP grove noise is likely to be higher than the preamp, the next test becomes more important which is the RIAA equalization:

We see nearly flat response which is what we want to have (i.e. no tonality imparted on behalf of the phono stage). A rumble filter would be nice but that is reserved for their higher end unit.

Let's sweep the input voltage and see where hard clipping occurs as this will impact how bad LP pops and clicks will sound:

This is better than a lot of budget phono preamps. But let's see if that is frequency dependend:

So no concern there.

We can see the same when we sweep the frequency fully:

Conclusions
The Parasound Zphono is not sexy but solidly delivers on basic functionality of a budget phono stage. Not much fault can be found in the measurements other than perhaps level of noise.

Overall I am happy to recommend the Parasound Zphono.


-----------
As always, questions, comments, recommendations, etc. are welcome.

Drove 150 miles today trying to find jars to continue canning our tomato harvest. Only found a few after visiting half a dozen stores. Apparently everyone had stayed home during the pandemic and gardened enough to need to can the surplus. Instead of stocking up weeks ago I stayed home and tested audio gear. So yes, it is all your fault and it is time to pay up by donating to the site using : https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/

 

[MadMan]

Why does the MM 1kHz tone go above 0dB ? And does that affect the results ?

 

[amirm]

I forgot to set the reference to that level. It is a graphing issue and doesn't change the result.

 

[restorer-john]

@amirm Do you have other choices for generator output impedances (you used 20R) closer to a typical MM and MC cartridge respectively?

 

[Hayabusa]

The RIAA curve is defined with two RC constants with a typical phase curve at the recording side in such a way that an analog filter with playback RC constants would correct the phase response back to 'flat' for the whole audio chain.
Now we see some phone preamps that have a DSP for RIAA.
Now the phase response could be quite different depending on how the RIAA is implemented digitaly.

So it would be interesting to see also a phase response.

 

[LTig]

The RIAA curve is defined with two RC constants ...

No, it has three constants.

 

[Rja4000]

@amirm
Hi Amir
About your ranking for phono preamps:
Shouldn't you compensate somehow for the gain difference ?

This one has 75dB SINAD for 46dB gain.
If you compare with the Cambridge Solo, that's 86dB SINAD for 39dB gain.

That means that, anyway, you'll have to push the Cambridge up by 7dB to achieve the same level. So another gain stage will be required, with its own noise contribution.
As SINAD is directly linked to noise, if I'm not wrong, the gap may actually be significantly lower than RAW SINAD indicates.
Is that right ?

 

[MC_RME]

I don't think so. SINAD without distortion (here...) is Signal to Noise. If you change the Parasound's circuit to have 7 dB less gain the noise floor will be 7 dB lower - but the signal too, so SNR (here SINAD) will not change at all. So the Cambridge still deserves the first place for being the most noiseless preamp tested so far.

 

[MC_RME]

@amirm Do you have other choices for generator output impedances (you used 20R) closer to a typical MM and MC cartridge respectively?

That's an interesting point. 20 Ohms basically short the phono preamps input, for max SNR. What impedance do typical MM and MC systems have - you have some numbers at hand?

 

[restorer-john]

That's an interesting point. 20 Ohms basically short the phono preamps input, for max SNR. What impedance do typical MM and MC systems have - you have some numbers at hand?

MC is very low, the loading (matching) on MC front ends usually is switched or suitable for 3 Ohms to 40 Ohms, whereas the loading for MM is typically 47K Ohms. The cartridge impedance itself with MCs and MMs* is around 1/10th-1/100th of the load value.

In my opinion, the source impedance for the MC is perhaps too high and for the MM is way too low. It will give artificially high numbers for the MM I would expect. @SIY may have some comments. It should be shorted or typical R IMO.

*I dug out a box of random average MMs to measure the DCR of them (both channels) and put in a spreadsheet. (the things we do)

 

[Deleted member 2348]

What impedance do typical MM and MC systems have - you have some numbers at hand?

MM: 500 … 1000 Ohm, 0.5 … 1H
MC: 3 … 30 Ohm, < 100µH

Testing SNR of a MM input with <100 Ohm source impedance doesn't make sense, because the Johnson noise of the MM cartridge and the input current noise of the amplifier (which is converted into voltage noise by the impedance of the cartridge) is not taken into account.

 

[restorer-john]

MM: 500 … 1000 Ohm, 0.5 … 1H

Plenty go lower than that. 30% of the ones I have tested so far.

 

[MC_RME]

MM: 500 … 1000 Ohm, 0.5 … 1H
MC: 3 … 30 Ohm, < 100µH

Testing SNR of a MM input with <100 Ohm source impedance doesn't make sense, because the Johnson noise of the MM cartridge and the input current noise of the amplifier (which is converted into voltage noise by the impedance of the cartridge) is not taken into account.

If you use a higher impedance you will raise the noise - but that only averages the result(s). It might be useful to use a lower impedance to better show the difference in excellence of engineering. You could also say lets measure with the needle dropped - more realistic again, but then noise would swamp away any difference...

For line inputs SNR is usually tested with shorted inputs. It is the task of the source to be as low impedance as possible. In real world using 150 Ohms instead of 0 Ohms does not make much difference with line. With high gain mic inputs the difference between 150 Ohms and 0 Ohms is about 3 dB.

For an MM input it is easy to test what happens as the AP can switch between 20 Ohm, 100 Ohm and 600 Ohm. The question is if such an impedance change just gives the expected small difference in noise, or some more artefacts like even more noise and THD due to specific properties of the circuitry under test.

Maybe SIY knows the answer already, and testing at those impedances won't change the outcome in a significant way.

 

[restorer-john]

Here are the results of 24 MM cartridges as pictured above. Left and Right channels.

at = Audio Technica (twin 90 degree magnets)
at vm = Audio Technica vertical magnet
piezo = The Japanese manufacturer named Piezo.

 

[milosz]

I'd love to see a "dashboard" of actual vinyl playback SINAD etc

 

[restorer-john]

I'd love to see a "dashboard" of actual vinyl playback SINAD

It would be a fight for the poor AP to give a stable number.

It's much easier to just listen to those records you love and marvel in the knowledge that a rock, dragged along a groove, and has its movements which induce a tiny voltage then amplified by 60dB, sounds so awesome.

 

[Deleted member 2348]

If you use a higher impedance you will raise the noise - but that only averages the result(s). It might be useful to use a lower impedance to better show the difference in excellence of engineering. You could also say lets measure with the needle dropped - more realistic again, but then noise would swamp away any difference...

Sorry, but this is wrong.

1.) For high frequencies the source impedance for the input approaches 47kOhm (input resistor). One way of "excellence of engineering" is to replace the 47k input resistor with a higher resistor (e.g. 1 Meg) which is not connected to ground but to an inverted and amplified "copy" of the input signal. Yamaha did this in the seventies in their top models. You can gain approx. 1.5dB of SNR with this "trick" with real life cartridges. This real life advantage would be hidden by the use of a low source impedance. See for example Self, Small Signal Audio Design, 2nd ed., p. 321ff.

2.) Input noise current is very real with bipolar transistor inputs. With a low source impedance you simply ignore the input noise current—with a real MM cartridge you can't. NE5534A with typ. 0.4pA/sqrt(Hz) is a good example for a very low input noise current with bipolar inputs, AD797 in contrast with 2pA is quite high. With 10k Source impedance (remember the inductance!) these values translate into 4nV/sqrt(Hz) (NE5534) and 20nV/Sqrt(Hz) (AD797). Compare this to the noise voltage densities of these OP amps (3.5nV/sqrt(Hz) for NE5534, 0.9nV/sqrt(Hz) for AD797).

 

[SIY]

MC is very low, the loading (matching) on MC front ends usually is switched or suitable for 3 Ohms to 40 Ohms, whereas the loading for MM is typically 47K Ohms. The cartridge impedance itself with MCs and MMs* is around 1/10th-1/100th of the load value.

In my opinion, the source impedance for the MC is perhaps too high and for the MM is way too low. It will give artificially high numbers for the MM I would expect. @SIY may have some comments. It should be shorted or typical R IMO.

The MC source impedance for units that are not transformer coupled is pretty noncritical as long as it's in the dozens of ohms or lower. Current noise is generally not enough of a factor to make much of a difference.

MM is a different matter- higher DCR and (most importantly) higher L mean that any input current noise will translate to a higher overall noise as the cartridge impedance rises with frequency (and thus shunts the 47k input resistance less).

Keep in mind that once you go beyond 70dB or so signal to noise, you've already hit the cartridge Johnson noise limit, so the preamp will have almost no noise contribution.

edit: @paschulke2 and I cross-posted.

 

[magicscreen]

It's much easier to just listen to those records you love and marvel in the knowledge that a rock, dragged along a groove, and has its movements which induce a tiny voltage then amplified by 60dB, sounds so awesome.

..and looks so awesome

 

[PuX]

seeing these measurements, I have a question:

could vinyl be better that cd audio in any way?

all I can think of is that for example there are vinyl rips with wider frequency range (with sound over 22khz).

are there any other vinyl benefits considering most phono preamps seem to have worse performance that an average DACs (in terms of SINAD at least)?