This is a review and detailed measurements of the NAD M28 Multichannel amplifier based on Purifi Eigentakt technology. It was kindly purchased by a member from a local dealer (Gig Harbor Audio) and sent to me for testing. The M28 costs US $4999 through its dealers.
As with the M27 amplifier I reviewed, the M28 comes in the same gorgeous industrial design:
At nearly 50 pounds, this is a heavy amplifier likely due to the weight of the enclosure as the class D amplifiers and it switching power supply are light.
The one thing I am not a fan of is the touch on/off button on top. It is not responsive. I wish the NAD logo was a hard switch.
The back panel shows how you do a first class job of termination:
I have kind permission from the owner for a teardown so look for that soon.
EDIT: Tear down is now here: https://www.audiosciencereview.com/...28-7-channel-purifi-amplifier-teardown.15981/
Amplifier Measurements
The Purifi 1ET400A that the NAD is based on scored a SINAD (sum of distortion and noise) of 104 dB. Let's see we get that from M28:
Oh no!!! Distortion has risen by 15 dB but thankfully it is below threshold of audibility. Noise however, has degraded fair bit, sinking SINAD by the tune of 6 dB. There is also good bit of variability with the SINAD measurement constantly changing 1 dB or more. As it is, the M28 slips in performance even relative to last year's M27:
For a clue of what may have happened we can look at the amplifier gain which is now at 29 dB versus reference design's 27 dB. Given the anemic output of AV products from their pre-outs, I can see why NAD opted to do this but whatever the reason, noise performance has suffered. We can clearly see this in signal to noise ratio:
As noted, we have dropped 5 dB in noise performance which more or less explains the similar drop in SINAD. This is a shame as noise can be audible unlike low levels of distortion. Fortunately we still clear my hurdle of 96 dB for 16 bit content even at 5 watts so practical consequence will be small.
Before someone asks, here is the dashboard with RCA:
Thankfully performance is similar.
Frequency response is good and matches the reference design:
Crosstalk performance was superb:
You buy a power amplifier for "wattage" so let's see how the M28 performs into 4 ohm load:
Gosh, not only do we have higher noise level but also much higher (in relative terms) distortion as power output increases relative to reference design. I did not want to believe this so to confirm my setup, I pulled the reference design and measured it again and its performance was unchanged (green). So this is definitely a degradation in M28 design.
We can see similar issue with IMD intermodulation test:
Fortunately as noted, we do get more power than the reference design which was likely limited by its power supply.
Letting distortion climb up to 1% produces more power naturally:
Interesting to see that burst power is actually a bit lower! The test attempts to change output power until it achieves 1% THD+N and it must have reached that at that lower power. I wonder if the sudden burst on the input increases distortion a bit. It is an engineering curiosity than anything else.
Switching to 8 ohm load we get:
We again have lower performance compared to reference but the gap is lower.
With AVRs, we see power reducing as we increase the number of channels. Not here:
This is excellent and shows the value of high efficiency amplification and power supply.
Sweeping at different frequencies shows some small oddities but nothing of major concern:
Note that our normal testing for power per industry standard uses 1 kHz but the most need is at lower frequencies. For that reason, I have started to annotate power at 20 Hz as seen above. All amplifiers I have tested reduce their power at low frequencies.
EDIT: per request here is multitone test:
Conclusions
Without context, the NAD M28 would get very high praise from me. Alas, taking a reference design and reducing its performance is not going to put me in that mood. This is the cost of manufacturers not providing detailed specifications/measurements like I am providing. If they had done that, their marketing department would immediately see that they have taken a step back and push engineering to do better. In absence of that, the "Eigentakt" label is plastered in the marketing material and celebration begins.
What NAD has done is taken a superior amplifier design and brought its performance below that of older Hypex technology. And charge $600 more for the privilege. Not right.
On the glass half full side of the argument, this is a beautiful amplifier, superbly crafted with excellent performance and tons of power. I looked and could not find any M27 for sale anyway so M28 it is.
I am going to of course put the M28 on my recommended list. Just no celebration. NAD, you could have done better.
------------
As always, questions, comments, recommendations, etc. are welcome.
A wind pattern is pulling all the smoke from wildfires on the entire west coast of US toward us. Woke up with combination of fog and smoke. What does that have to do with me pleading for money from you all? Nothing. Sometimes I like to act like the newsman. Since you did not have to read another lame joke, how about using this occasion to donate using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
As with the M27 amplifier I reviewed, the M28 comes in the same gorgeous industrial design:
At nearly 50 pounds, this is a heavy amplifier likely due to the weight of the enclosure as the class D amplifiers and it switching power supply are light.
The one thing I am not a fan of is the touch on/off button on top. It is not responsive. I wish the NAD logo was a hard switch.
The back panel shows how you do a first class job of termination:
I have kind permission from the owner for a teardown so look for that soon.
EDIT: Tear down is now here: https://www.audiosciencereview.com/...28-7-channel-purifi-amplifier-teardown.15981/
Amplifier Measurements
The Purifi 1ET400A that the NAD is based on scored a SINAD (sum of distortion and noise) of 104 dB. Let's see we get that from M28:
Oh no!!! Distortion has risen by 15 dB but thankfully it is below threshold of audibility. Noise however, has degraded fair bit, sinking SINAD by the tune of 6 dB. There is also good bit of variability with the SINAD measurement constantly changing 1 dB or more. As it is, the M28 slips in performance even relative to last year's M27:
For a clue of what may have happened we can look at the amplifier gain which is now at 29 dB versus reference design's 27 dB. Given the anemic output of AV products from their pre-outs, I can see why NAD opted to do this but whatever the reason, noise performance has suffered. We can clearly see this in signal to noise ratio:
As noted, we have dropped 5 dB in noise performance which more or less explains the similar drop in SINAD. This is a shame as noise can be audible unlike low levels of distortion. Fortunately we still clear my hurdle of 96 dB for 16 bit content even at 5 watts so practical consequence will be small.
Before someone asks, here is the dashboard with RCA:
Thankfully performance is similar.
Frequency response is good and matches the reference design:
Crosstalk performance was superb:
You buy a power amplifier for "wattage" so let's see how the M28 performs into 4 ohm load:
Gosh, not only do we have higher noise level but also much higher (in relative terms) distortion as power output increases relative to reference design. I did not want to believe this so to confirm my setup, I pulled the reference design and measured it again and its performance was unchanged (green). So this is definitely a degradation in M28 design.
We can see similar issue with IMD intermodulation test:
Fortunately as noted, we do get more power than the reference design which was likely limited by its power supply.
Letting distortion climb up to 1% produces more power naturally:
Interesting to see that burst power is actually a bit lower! The test attempts to change output power until it achieves 1% THD+N and it must have reached that at that lower power. I wonder if the sudden burst on the input increases distortion a bit. It is an engineering curiosity than anything else.
Switching to 8 ohm load we get:
We again have lower performance compared to reference but the gap is lower.
With AVRs, we see power reducing as we increase the number of channels. Not here:
This is excellent and shows the value of high efficiency amplification and power supply.
Sweeping at different frequencies shows some small oddities but nothing of major concern:
Note that our normal testing for power per industry standard uses 1 kHz but the most need is at lower frequencies. For that reason, I have started to annotate power at 20 Hz as seen above. All amplifiers I have tested reduce their power at low frequencies.
EDIT: per request here is multitone test:
Conclusions
Without context, the NAD M28 would get very high praise from me. Alas, taking a reference design and reducing its performance is not going to put me in that mood. This is the cost of manufacturers not providing detailed specifications/measurements like I am providing. If they had done that, their marketing department would immediately see that they have taken a step back and push engineering to do better. In absence of that, the "Eigentakt" label is plastered in the marketing material and celebration begins.
What NAD has done is taken a superior amplifier design and brought its performance below that of older Hypex technology. And charge $600 more for the privilege. Not right.
On the glass half full side of the argument, this is a beautiful amplifier, superbly crafted with excellent performance and tons of power. I looked and could not find any M27 for sale anyway so M28 it is.
I am going to of course put the M28 on my recommended list. Just no celebration. NAD, you could have done better.
------------
As always, questions, comments, recommendations, etc. are welcome.
A wind pattern is pulling all the smoke from wildfires on the entire west coast of US toward us. Woke up with combination of fog and smoke. What does that have to do with me pleading for money from you all? Nothing. Sometimes I like to act like the newsman. Since you did not have to read another lame joke, how about using this occasion to donate using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
Last edited:
