Parasound Zamp V.3 Amplifier Review

[preload]

Just because I was curious, I decided to measure crosstalk on a Zamp v3 to see how it compares to the measurements posted here. I generated sine waves using REW on my laptop. The signal was played through the same, single channel of a Meizu USB-C DAC and fed individually to L and R inputs of the Zamp. I then used a Fluke 87 to measure voltage at the speaker outputs for the driven channel vs the non-driven channel. I did not attach a speaker load during measurement. I used an online calculator to convert voltage ratios to dB, and averaged the L->R and R->L crosstalk. Results are below. Voltage output was about 11-12V, which comes out to roughly 15W into 8 ohms.

In the unit I measured, I found that crosstalk was about 9-10 dB better at 1, 10, and 20kHz than those reported in ASR. This makes me suspect that either: the units were measured were substantially different (10dB is a major difference), there was a difference in test conditions, or there was a measurement error. Perhaps Amir or someone else can help reconcile why there would be such a large difference. My guess is that perhaps crosstalk performance varies with output voltage or the unit tested here was defective.

Channel Separation

 

[amirm]

I used an online calculator to convert voltage ratios to dB, and averaged the L->R and R->L crosstalk.

I don't do any averaging. What were the numbers for each channel separately?

FYI, while Fluke does rate the 87 up to 20 kHz, response is not flat:

It is already down (or up?) 3 dB at just 1 kHz. Hard to figure out the impact in this scenario but in general, it is not the right tool for the job.

Overall, what you show is poor performance anyway.

 

[preload]

I don't do any averaging. What were the numbers for each channel separately?

FYI, while Fluke does rate the 87 up to 20 kHz, response is not flat:

View attachment 66346

It is already down (or up?) 3 dB at just 1 kHz. Hard to figure out the impact in this scenario but in general, it is not the right tool for the job.

Overall, what you show is poor performance anyway.

Hi Amir, I agree with you the channel separation performance isn't great either way. At the same time, we all want to know we're looking at representative measurements, and a 10dB difference is enough to give anyone pause.

Here are the raw data you asked for:

The crosswalk is consistently worse in the R->L direction, and consistently by about 4dB compared to L->R.

Regarding the Fluke, it's not the most elegant tool for this, of course. At the same time the 87 is an industrial grade DMM - it really should be up to the task of measuring a simple AC voltage consisting of a sine wave. Moreover, since we're interested in the ratio of measured voltages (i.e. decibels) as opposed to the absolute voltage, any error/bias in the absolute voltage measurements should theoretically cancel out. But to be sure, let's say I dropped my Fluke 87 too many times, and the measurement of 11.13V was actually 10% higher (12.243V) and for some reason, the Fluke was measuring 10% lower in the millivolt range (i.e. 0.00297V instead of 0.0033V). The crosstalk would then calculate out to 72.3dB instead of 70.6dB (a 1.7dB diff)- and that's assuming a totally broken DMM with measurement biases on opposite directions! So my opinion is that the Fluke 87 should be accurate enough for the purpose of measuring crosstalk.

To clarify the issue of frequency response of the Fluke 87, I had a similar concern (particularly about measuring @ 20kHz). However, looking at the measured speaker-out voltages obtained at 10kHz and at 20khz, the response is only -0.3 dB at 10kHz and -0.8 dB at 20kHz relative to 1kHz. I didn't change the input level between tested frequencies, so the measured V drop is a function of Fluke + Zamp + DAC. But unless the Zamp and DAC have a combined rising freq response curve (unheard of), the Fluke's frequency response should be better than 0.3dB down at 10kHz and 0.7dB down at 20khz relative to 1kHz. So, my opinion is that frequency response of the Fluke shouldn't be an issue here either.

Perhaps there's another reason to explain the difference in crosstalk measurements.

 

[amirm]

The crosswalk is consistently worse in the R->L direction, and consistently about 4dB compared to L->R. I thought it was be easier to compare by providing averages.

The direction many times makes a difference. So if you and I used the reverse one, then 4 dB of that differential goes away.

 

[Angsty]

I should add that the ZPhono USB may be completely different under the hood, as it had no venting grill. The arrangement of ports on the back was completely different from what I see here, too. So I may just be confusing things. But I'd love to see a review of it because the hums told me there was something really wrong with it. Admittedly, the hums were faint, and could only be heard at higher volumes, but none of my other phono amps had that problem. A $100 ProJect USB box that had different issues had no hum issues whatsoever.

I have a Zphono USB just laying around if @amirm wanted to test it. It’s been officially discontinued, but Audio Advisor still has a stock of them to sell. It might be good just to see how effective the AD converter is.

 

[tvrgeek]

I have one of these little puppies. I to had asymmetric distortion levels and quite a bit of hum. I keep it because it fits in my case with the rest of my speaker lab.

 

[tvrgeek]

I have one I got for my workbench. Used fortunately. I think the review here is being overly kind. I too plan on digging into it and see if I can dix it some. Amazing they bring decent money on E-bay. I have the schematic. Very strait forward and conventional design. It should be a decent performer. Why someone changed the servo op amp to some esoteric hybrid I don't know. Saw a chip and figured it was bad? For fun, I might toss it into LTSpice and see what it says. IPS CM has more degen than usual, IPS and VAS differential. A lot of "protection" stuff layered on. Only 8000uF on each rail so that is a bit light. I am sure there is something amiss with the layout/grounding. A few other datils. No reason it is not over 100dB SNR and THD in the .00x range. Fun project maybe.

 

[tvrgeek]

https://www.head-fi.org/threads/parasound-and-burson-adventure.830082/


Resized and optimized

View attachment 52810

I was about to toss mine ( use for speaker testing) in the get rid of pile, but wondering. I can add some heat sink, better op amp ( not as extreme) and maybe a Mu-metal band around the transformer. Wonder it would creep up to acceptable for my desktop? THough people seem to pay good money for them, maybe best to dump and go Chinese class D.

 

[NyquistRate]

Could we get a second copy tested?

 

[tvrgeek]

Op amp is the DC servo. Hard to believe a transformation by this.
Basic circuit is pretty conventional and no glaring faults.
Going t put it into LTSpice and see if any mistake was made. Just looking at it, it should do 5W into 8 well under .01.

Got much of it in. Having trouble getting the model stable with the servo connected. All the protection circuitry may be part of the performance issues. There is more circuitry for the protection, turn on control and thermal monitoring than there is in amplifier! No current mirror on the input stage surprises me. Has to guess at some generic transistors.

I can see if I can do some simple tests with RightMark on mine. It does hum. It does get warm. It should be better than the above test though.

 

[tvrgeek]

Spice simulation, granted I made some assumptions on devices and had to adjust some currents for them, but, basic amplifier SIMULATES @ .008% @ 40W, 8 Ohms, 1KHz. But, add in the output current sense and it jumps to .112% . Input muting, bridge components, volume controls and output relays all yet to be added to the simulation.

Bias is not optimized and I did no simulation of the thermal compensation of the bias. Just showing why an otherwise very conventional amplifier measures so poorly.

Servo problem was not simulating a FET input op-amp. Critical with the impedances. When I drag mine in I will peek inside and see if I can identify some of the transistors. This got me curious if it is a candidate for POOGE. Though a bigger case and heat sink might be needed. If the transformer was pulled to the side, there would be a lot more room.

 

[gosammot]

Hello all,
I am a new member although I have been following the website and Amir's content for a while.
Had a chance to test a slightly different version of this amp at work a couple of days ago, and the distortion figures I'm getting are much better than what Amir reports (5W loaded with 4R). Not sure what might be causing this discrepancy. Curious to hear your thoughts.
Have a nice weekend,
Gosammot

 

[dalbert02]

Hello all,
I am a new member although I have been following the website and Amir's content for a while.
Had a chance to test a slightly different version of this amp at work a couple of days ago, and the distortion figures I'm getting are much better than what Amir reports (5W loaded with 4R). Not sure what might be causing this discrepancy. Curious to hear your thoughts.
Have a nice weekend,
Gosammot

I have 5 of these amps. My plan was to use a mini DSP and make some tri-amped speakers. (90W+90W+45W) I never got around to it and they are just sitting on the shelf. If Amir is interested, I'll ship them out. He graciously tested my ZDAC which performed poorly and that took the wind outta my sails to build the speakers. I wanted to use ZDACs and ZAmps exclusively as I found them inexpensive on eBay.

 

[cdc]

Interesting that Amir measured -60dB @1kHz crosstalk compared to Parasound specs -80dB at 1kHz compared to Audio Critic

Audio Critic
"The maximum separation of 66dB is a little bit on the scant side but still guaranteed inaudible"

RE Channels:
Amir: Well, this ain't a pretty sight. One channel is much worse than the other. No amount of playing with the grounding, or messing with the switch in the back helped with any of the mains/power supply related tones either.

Audio Critic Fig. 5 at 40 watts shows identical results as 1 Watt. I can;'t import the graph but if you go to the article, they look much better than what Amir measured.

I wonder what is going on here and why such differences in measurements.

 

[Bds3151]

I have one of these and ran tests - I think you can consider the one Amrim got as broken. Mine doesn't get more than lukewarm and tested much higher than this one. In excess of 90db S/N for example as I don't remember the exact number..

 

[AltoVariago]

I have one of these and ran tests - I think you can consider the one Amrim got as broken. Mine doesn't get more than lukewarm and tested much higher than this one. In excess of 90db S/N for example as I don't remember the exact number..

I have the same feeling. I have two that have been powering for years 4 Bose outdoor speaker and they never got more than warm even after hours at remarkable volumes

 

[maxswede]

@amirm can we get this re-tested and put it to bed once and for all?

 

[milosz]

Even though this is a somewhat flawed product - and it looks like the one Amir tested was defective - I think it's important to remember that when this thing came out, there weren't the profusion of small cool-running high quality class-D amps that we see today. The amp fulfilled a need, it's a niche product that has now more or less been made obsolete by advances in class-D technology.

 

[alvigg]

I had 3 of these. Sold them and bought Fosi V3's thanks to this website. A good upgrade and I made money.

 

[bitonio]

I actually did the opposite journey. I got the revision 2 of the z.amp v3 to replace a more than decent Aiyama A07, as my powered speakers had a bad hum I can totally hear in my new office room, much quieter than the old one.

Not an end game, just a fun project to compare different tech with my ears.
I agree with the recent Class D crop is really, really good for the money, yet the big power block is a big no no if you take cable management seriously and have already some challenge with this I had to use RCA split cable with the Aiyama, plus the powerbrick, gross. The volume knob on the Aiyama is very light, toy, ready to fall-off sort of.

The Zamp in the other hand they though of everything RCA loop out so you can connect the sub, gain knob, properly spaced out speaker terminals, trigger in and out, and NO powerbrick! Okay it takes more room on the desk but nothing crazy on my 60" wide desk. The finish of the case is a different league from chifi, as you would expect.

I concur, the Parasound little amp is luckwarm after 6 hours of playback and mine has my headphone stack on it.