PS Audio PMG Signature

[GXAlan]

These are the new products developed by Darren Myers. They talk up measurements, although I am not sure what voltage achieves the SINAD they are talking about.

Housed in a double chassis construction that shields against noise and interference, the PMG Signature achieves a SINAD of -150dB—so quiet it’s not there.

PMG Signature Preamplifier

Fully balanced high-performance preamplifier featuring zero-feedback Diamond buffers and a pure Class A analog signal path. The PMG Signature sets a new standard in clarity, control, and musical transparency.

www.psaudio.com

They did say they were going to start to publish measurements of their products. I assume it will be this generation.

Their DAC upsamples to DSD 22.4 MHz (in contrast to Marantz’s 11.2 MHz upsampling in the SA-10). Since DSD has ultrasonic noise, it never will measure as cleanly as a standard PCM DAC. This could be audiophoolery under pseudo blind testing, although I have found that my SA-10 to be more enjoyable than my Topping D90 when I had both although the difference was vanishingly small in comparison to the difference after redirecting my budget to speakers and DSP, so I sold both and just use a 90 dB SINAD DAC now.

 

[bryan0101]

I'm really curious how they measure the preamp to be -150db. It'd be a new bar for everyone if true.

 

[GXAlan]

“built on cutting edge measurements”

 

[GXAlan]

 

[wwenze]

Nothing legally stops a company from achieving high performance numbers, I'm sure it will be met with praise if it achieves what it states.

-150dB

0.0002% distortion plus -150dB noise is not 0.0003%, but that's none of my business. Maths is evil, amirite.

SINAD of -150dB

SINAD is literally in the table above.

The entire article is, well, interesting.

 

[JustJones]

There's 4 of these Signature Series components at $8500 each and they all appear to be sold out. Must be a hit.

Signature Series​

• 
  
  
  

  PMG Signature 512 DAC​

  
  
• 
  
  
  

  PMG Signature SACD Transport​

  
  
• 
  
  
  

  PMG Signature Preamplifier​

  
  
• 
  
  
  

  PMG Signature Phono Preamplifier​

 

[amirm]

Thd+n as specified translates to sinad of 110.

 

[GXAlan]

Thd+n as specified translates to sinad of 110.

Agreed, and it is specified it’s 1V input. I am thinking that the 150 dB has to be under some very specific condition.

If 1V output gave you 110 dB SINAD. Then residual noise is 0.003162 mV. Then if you added 150 dB, you would have to get all the way to 100V which doesn’t make sense.

Looking at it the other way. They say max gain of 12.6 dB. If you assume 4V in, that means it can output 17V.

If you wanted 150 dB SINAD at 17V, you would need residual noise of 0.0005376 mV.

The Node Icon was 0.0018962 mV residual noise.

That’s a difference of 10 dB.

Is it possible that the PS Audio has a 10 dB improvement in noise compared to the Node Icon and simultaneously can output 17V instead of the 3.7ishV?

 

[Matias]

So will they lend a unit for Amir to review? So we can compare in the standard measurements against the competition.

 

[unpluggged]

I thought this elderly scammer and charlatan had been banned on this forum along with his PoS Audio products?

 

[Zgrado1970]

The page for the transport states that your music collection contains “a treasure trove of music you’ve never really heard,” which of course the transport and dac will “unlock” for you. It also states that with an I2s connection, the transport will “unlock the raw DSD layer of SACD discs - delivering what was once hidden to external DACs.”

AFAIK, there are very few (none?) devices that allow one to extract the DSD layer of SACDs to play direct DSD with a capable external DAC. I recall reading some threads here about using a PC (or maybe an Oppo player) to do this.

So while I think direct DSD playback is silly, it is desired by the faith-based audiophile crowd. Is this device unique in providing that (for a mere $8500)?

 

[wwenze]

The typical technical discussion about I2S goes like:

Student: "If I2S invented by Philips is so good and is already a standard widely used internally, why does nobody use it for external and even Philips created the SPDIF format for external connection?"
Teacher: "Because I2S is meant for internal connection with 3 or 4 signal lines and the connection is not impedance-optimized for coaxial or whatever cable"
Student: "Then why use I2S over SPDIF for internal connections?"
Teacher: "Because it's cheap duh"

 

[Chris Brunhaver]

So, I'm the speaker guy for PS.

Paul significantly misspoke on the website and in his launch presentation and this will be corrected shortly. He's currently vacationing in Europe to see his oldest son and grandkids.

The table of specs are correct but were based on a prototype unit and there may be a couple of small tweaks as the production units measure a tad better.

We've spent a few years developing these as a company so I'm not sure how this happened but we'll make sure everyone knows the correct specs on this piece and we publish some AP measurements. It might not be quite as comprehensive as Amir's but I think we'll have plenty of additional info.

No the product doesn't have 150 db SINAD or even 150 dB noise floor. The discrete opamp that this preamp is based on has about 150 dB nosie floor in unity gain on it's own but in the design (with up to 12 dB of gain), 12 dB less feedback is used and the volume control circuit zero crossing detection (while a very good volume control) still dominates some of the performance. Noise floor is around -145, and overall performance is very good/excellent. As I recall, SINAD is ~115 dB range but I'll need to double check the exact spec. Channel separation is extremely high

Darren is no longer at PS (as of about 2 years ago) but has helped make sure his designs made it into product as intended.

The foundation of the preamp is a custom opamp design that's a unique current feedback design where the feedback path to the current node is buffered and the open loop performance is inherently very linear, so not a ton of feedback is used. Having flat distortion versus frequency was a goal of this preamp and this flat distortion vs frequency characteristic extends to 50 kHz. Low phase shift and wide bandwidth was a goal and there is ~500 kHz bandwidth.

We actually have a lot to talk about with the preamp tech and its a well through through and refined sounding design (and a huge step up from our previous tube "BHK" model), In my opinion.

Regarding the DAC, it's a somewhat similar approach to our previous DAC model (but the engineering was done by Doug Goldberg and Bob Stadtherr). It's a 1-bit DSD DAC where everything upsampled to DSD512 (though there are menu settings for 128fs and 256fs) and that does into a 64 tap FIR filter run through a small FPGA into an analog lowpass filter and eschews to use of the transfer approach that our previous DAC engineer Ted Smith used as part of the inductance in the lowpass filter).

The net result is a significantly improved objective performance versus our previous DAC with about 20 dB better SINAD at around or just over 100 dB (I need to look at the final measurements, as this isn't my project). We're using a Crystek ultra low phase noise femptoclock and advanced clock distribution chip and getting very good jitter performance. (around 2 picoseconds typical).

Do I think this will suddenly win over the ASR crowd? No. Though I do low the new analog and phono preamp in particular (which are Darren's work). They're a substantial step up for us as a company and the chassis are beautifully made and I hope people will really like them.

I'm sorry for the confusion and I'm going to work with Paul and our web team to get the mistakes corrected.

 

[amirm]

Do I think this will suddenly win over the ASR crowd? No.

Well, you have won me over with this post and the effort behind the product to significantly improve performance. All I ask for major brands is to be better than average. Sadly many have been well below average despite their higher prices. Getting to the top of the class gets extra points but is not as necessary as beating the average of all products tested.

FYI, I really like the look of the units and touch screen.

 

[kyuu]

Do I think this will suddenly win over the ASR crowd? No.

If the measurements are as good as is being made out, then it should at least not get roundly criticized. The chassis looks nice too. If there's something that's not going to win over the "ASR crowd" it's the prices, I think, as you obviously don't have to pay many thousands of dollars to get a good DAC.

 

[GXAlan]

So will they lend a unit for Amir to review? So we can compare in the standard measurements against the competition.

They may not as they have said in the past that they will release their own APx555 measurements for future products and this lineup is the future product they were talking about.

If you look at their DAC, the noise shaping and DSD conversion comes at the expense of SINAD but the implementation by Marantz was actually excellent and theoretically closer to the source than even the Topping D90.

Two reasonable DACs with a large PK Metric when tested with real music...

I've previously shown that my SA-11s2 and Panasonic UB9000 sound similarly and measure similarly with DeltaWave. In this post, I'm sharing some preliminary data showing a) a measurable PK Metric difference between two reasonable DACs with real music b) the DAC with a ~10 dB worse 1 kHz SINAD...

www.audiosciencereview.com

So, I'm the speaker guy for PS.

Paul significantly misspoke on the website and in his launch presentation and this will be corrected shortly.

The discrete opamp that this preamp is based on has about 150 dB nosie floor in unity gain on it's own but in the design (with up to 12 dB of gain), 12 dB less feedback is used and the volume control circuit zero crossing detection (while a very good volume control) still dominates some of the performance. Noise floor is around -145, and overall performance is very good/excellent.

That would also be spectacular if the noise floor is that low. That’s the one thing that’s still audible whereas distortion really isn’t audible.

 

[unpluggged]

Paul significantly misspoke

Nice attempt at damage management. But no.

the chassis are beautifully made and I hope people will really like them

That's all we need to know about your engineering.

 

[Chrispy]

Seems mostly PS just needs to overcome their history, and if they're doing that good for them, albeit perhaps a bit slow to do so.

 

[amirm]

Nice attempt at damage management. But no.

Please be kinder to industry people who despite some resentment from us, still come and give us information. I read Chris' post as genuine and no attempt at anything.

 

[Adis]

Agreed, and it is specified it’s 1V input. I am thinking that the 150 dB has to be under some very specific condition.

If 1V output gave you 110 dB SINAD. Then residual noise is 0.003162 mV. Then if you added 150 dB, you would have to get all the way to 100V which doesn’t make sense.

Looking at it the other way. They say max gain of 12.6 dB. If you assume 4V in, that means it can output 17V.

If you wanted 150 dB SINAD at 17V, you would need residual noise of 0.0005376 mV.

The Node Icon was 0.0018962 mV residual noise.

That’s a difference of 10 dB.

Is it possible that the PS Audio has a 10 dB improvement in noise compared to the Node Icon and simultaneously can output 17V instead of the 3.7ishV?

Maybe they use something similar to SPL's VOLTAiR technology?

https://spl.audio/en/spl-produkt/elector/