Parks Audio Puffin Review (Phono Stage)

[amirm]

This is a review and detailed measurements of the Parks Audio Puffin Phono Pre-amplifier. This is a special device in that it digitizes the cartridge output and then is able to perform a wealth of signal processing on the signal from pop and tick removal to RIAA equalization. The standard one with analog in and out costs US $489 including Prime shipping on Amazon. The version the member purchased with Toslink digital out costs a lot more at $639 including shipping.

While compact, this is an attractive phono stage:

User interface is easy to navigate with countless option (22?). As you can see, there is a volume control as well so you could drive a power amp directly instead of going through a pre-amplifier (assuming you don't have other inputs).

The Toslink option is brought out on one side:

I tested and compared the unit using both RCA out and Digital.

Parks Audio Puffin Measurements
Let's start with our using moving magnet dashboard of 5 mv in with the device configured to 40 dB:

As you see, I had an easy time dialing out mains noise and hum. On the other hand, it is disappointing to see such high level of distortion.

Switching to digital out eliminates all but the second harmonic:

We don't gain a whole lot there and performance is still middling:

Measuring distortion alone we get:

So basically most of the distortion is baked in the input buffer and ADC side.

Where the Puffin shines is in its perfection implementation of RIAA equalization:

This gets even better when using digital out:

The manual says the internal sample rate is 96 kHz but per above, I am not seeing it clear much beyond 20 kHz which indicates a sample rate of 44.1 kHz. But maybe sample rate is high but internal processing is done at lower rate. Or there is an option to change the same rate -- I did not see it in my quick read.

A tricky part of this design is what to do with wide dynamic range of the input with respect to pops and clicks. So let's run our overload test:

This is not much of a headroom. Fortunately there is some pop removal processing so if the ADC doesn't overload, maybe that is a good workaround.

Conclusions
I have a soft spot for the Puffin as it is made locally and looks kind of cute. Its RIAA equalization is the most perfect I have seen in any phono stage which is the most audible thing. All the other filters and processing far outweigh anything that is out there and would make for a bunch of fun tweaking. Alas, not a fan of the high price with digital output. As far as I am concerned, that should be standard in the base price let alone incur a $200 penalty. Headroom is low as well and distortion is higher than pure analog phono stages. Then again, the LP medium has so much distortion that the latter point doesn't make any difference in reality.

Overall, I am going to put the Parks Audio Puffin on my recommended list but you may not.

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

Appreciate any donations using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/

 

[sam_adams]

Is the device powered by an external wall wart or other type of PS?

 

[RickSanchez]

I won't lie: the name "Puffin" for this device makes me think of this guy ...

 

[amirm]

Is the device powered by an external wall wart or other type of PS?

A small switching power supply gives it juice (about 2X the size of a phone charger).

 

[Doodski]

I like the carbon fiber sort of look.

 

[restorer-john]

@amirm Are we really seeing a 1kHz overload at just 20mV for a typical MM gain set to 40dB? That's unbelievably poor for any phono stage, regardless of price. Even that little Realistic 40 year old horror-show had an 80mV overload.

Here's a block diagram I found in another review:

(audioXpress)

And the RIAA deviation is shocking above 14kHz where it drops like a stone to end up nearly 3dB down at 20kHz. Is there an issue with feeding the front end with such a low source (AP ~20R)? Are you sure there was no filters or rolloffs set into it?

 

[DDF]

@amirm Are we really seeing a 1kHz overload at just 20mV for a typical MM gain set to 40dB? That's unbelievably poor

I agree, 20 mV is too low for MM IME. A phono input should ideally be able to handle 50cm/sec modulation from the record
http://www.pspatialaudio.com/max_velo.htm

A 12" 45 single can support a maximum recorded velocity of up to 68cm/s, so 68 cm/s is even better.

My Stanton 881 MkIIs (common) sensitivity is 1mV/cm/sec and will easily blow right through 20mV even without pops and clicks.

I had one 45 rpm with at least 50 cm/s (estimated) and set up my rig to handle 70 mV from the cartridge before clipping, to avoid distortion

 

[restorer-john]

I agree, 20 mV is too low for MM IME. A phono input should ideally be able to handle 50cm/sec modulation from the record
http://www.pspatialaudio.com/max_velo.htm
View attachment 108458
A 12" 45 single can support a maximum recorded velocity of up to 68cm/s, so 68 cm/s is even better.

My Stanton 881 MkIIs (common) sensitivity is 1mV/cm/sec and will easily blow right through 20mV even without pops and clicks.

I had one 45 rpm with at least 50 cm/s (estimated) and set up my rig to handle 70 mV from the cartridge before clipping, to avoid distortion

I agree. I have some really hot 12" 45s that back in the day, with a high-ish output MM would hit 80mV. That's with no surface irregularites.

If it's overloading at 20mV with ~40dB gain, what hope does it have at MC gains of ~60dB? It's going to be overloading under 2mV.

I am hoping there is something not set correctly that Amir may have missed (it is a complicated little thing) that is causing this, otherwise it is just broken by design. Disappointing.

 

[amirm]

@amirm Are we really seeing a 1kHz overload at just 20mV for a typical MM gain set to 40dB?

That's it. I think his ADC is overflowing, not any kind of analog issue. If you look at the digital dashboard, at 5 mv he only has 12 dB of headroom before digital clipping. That is a ratio of 4X. 5m times 4 = 20mv and that is what we are seeing as the clipping point.

 

[amirm]

And the RIAA deviation is shocking above 14kHz where it drops like a stone to end up nearly 3dB down at 20kHz.

It has a bunch of filters so maybe one of them was set for high frequency. Didn't think to check that. I overrode the low frequency one but the high frequency one may still be on I just packed the thing so hate to get it back out.

 

[restorer-john]

I just packed the thing so hate to get it back out.

Thanks anyway. Leave it packed up.

Maybe the designer may weigh in if he sees this review with some comments.

 

[amirm]

Thanks anyway. Maybe the designer may weigh in if he sees this review with some comments.

Actually he gave a podcast to Chris at AmpHour a few months back. It is there that I remembered him saying he was challenged with dynamic range of the ADC and such. It is a longwinded and disorganized chat though:

https://hwcdn.libsyn.com/p/4/f/7/4f...11646010&hwt=759ba22884a99e156260bb591c869f7b

 

[restorer-john]

It is there that I remembered him saying he was challenged with dynamic range of the ADC and such.

I wonder why he simply didn't cascade switched 20dB and 40dB dedicated MC and MM flat stages along with an attenuated and switched line input for the strange "CD" option. He could have trimmed the levels with digitally switched attenuators and/or fixed gain stages like the AP and other test gear does for range changing to keep the signal in the sweet spot of the A/D and not hit the wall.

I guess it's the price you pay for flat amplifying un-equalized signal straight from a cartridge and doing RIAA in DSP. There's 40dB chopped out of the A/Ds range just for RIAA REC/EQ and then the music dynamic range on top of that. I can see why it runs out of DR and clips.

I'm glad you got to review this Puffin. It has intrigued me for a long time.

 

[Newman]

Special ratings ceramic for this device:

 

[Frank Dernie]

The manual says the internal sample rate is 96 kHz but per above, I am not seeing it clear much beyond 20 kHz which indicates a sample rate of 44.1 kHz.

It would make sense to either sample at 44 or filter the HF rubbish before the ADC since the RIAA record curve hugely amplifies the HF before the record is cut so the cartridge output potentially contains a lot of inaudible HF signal which, if digitised first and RIAA correction in the digital domain, will maybe limit the dynamic range achieveable. Even just the audio band needs 40dB of dynamic range over and above the music to digitise it all pre-correction.

 

[pavuol]

So with the additional 24/96 outputs vinyl lovers are now able to squeze the hell out of their platters make "hi-res" digital backups?

 

[Frank Dernie]

I haven't looked into it but presumably the Puffin has and input gain adjustment to match the cartridge.
The phono input on the Devialets I have owned do, and it is a critical adjustment to maximise dynamic range and avoid overload.

 

[Francis Vaughan]

the cartridge output potentially contains a lot of inaudible HF signal which, if digitised first

In principle this should never happen. Minimally the anti-alias filter should brick wall everything above the Nyquist frequency. The problem is as you note, since the media inherently contains lots of energy up there, the anti-alias filter may still leak stuff. At which point things turn quite nasty.

The manual says it uses a TI PCM1808 ADC, which is 24 bits wide and can run up to 96kHz. It is an ADC that is targeted at "cost sensitive" consumer applications. The ADC includes a low pass filter, but an optional additional one is suggested.

One of the really neat things this pre provides is support for vintage vinyl enthusiasts. So not just RIAA but a slew of other pre-RIAA equalisation curves. Plus useful mono support. It is a one stop box, turntable, pre, power amp. Even has room bass equalisation.

One suspects that internally the DAC ADC are run as a master slave pair with the DSP in between. This has the advantage that you are immune from worrying about anyone elses sample clock. Run the whole thing at the rate that suits you. The downside is that if you want to create a digital output, a sample rate converter with its own clock will be needed. So there is additional hardware and grief. OTOH, if someone wants digital out, there is no need for the DAC. So arguably a different product is the answer.

 

[Soniclife]

OTOH, if someone wants digital out, there is no need for the DAC. So arguably a different product is the answer.

Sadly the market is not awash with options, which is why there is so much interest in this product. A few competing products would be great.

Regarding the headroom, for the people that own this does it seem to be a problem? My Devialet which may have the same problem is if anything better with surface noise than analogue devices.

 

[Frank Dernie]

In principle this should never happen. Minimally the anti-alias filter should brick wall everything above the Nyquist frequency. The problem is as you note, since the media inherently contains lots of energy up there, the anti-alias filter may still leak stuff. At which point things turn quite nasty.

Quite, but this was not my point.
The RIAA curve gives around 40dB cut/boost between 20Hz and 20kHz. If the weighting continues up into the inaudible range between 22.05 kHz and 48kHz, as it would if the sampling (and hence anti aliasing filter) is 96kHz, then the boost would continue for another 40dB or so between these pointlessly high frequencies.
Now yes, they probably won't be at a level potentially causing overload but still, having the anti-aliasing filter at 22.05 kHz rather than 48kHz is IMO much better, or, if for marketing reasons 96/24 is considered essential in the literature, a filter much lower than that would be advisable anyway before the normal anti aliasing filter, just to be sure to avoid pointless overload.
Sorry my earlier post wasn't clear enough on my meaning.