Image shamelessly stolen from Amir
@JPJ kindly sent me his Puffin stage for some further measurements following the brief review by Amir. After an unconscionable delay, here we go.
First, the basics. First, I flashed the ROM with the latest firmware version (1.20), which got rid of some of the anomalies I'd seen in earlier measurements. This is done via a USB cable to a module on the circuit board. I set the gain to 40dB (since that's right for most MMs I have on hand), load to 47k (ditto). Input capacitance measured slightly over the 50pF spec, which is suitable, accounting for typical cable capacitances, for MMs like the Audio Technica I use. For other cartridges, either use higher capacitance cables or make use of the extensive EQ functions built into this unit.
Gain measured 40.1 dB, almost exactly as advertised.
Note that I used a 100:1 attenuator at the input to allow the generator to run closer to its sweet spot and to feed the unit from a 1k source resistance. So any place you see a Generator Voltage, divide it by 100. There’s probably a way to do this in the APx software, but I was too lazy to figure it out. Next time. The “A” reference in the graphs (as in “dBRA”) is the output with 5mV in at 1kHz (500 mV).
First up, the noise spectrum, generated with a Stanton 881 as the source:
As expected, it roughly corresponds to the RIAA curve. The 60Hz spike might be something I could get rid of with some fiddling with grounding, so don’t worry about it too much.
Next, RIAA conformance:
Note the expanded scale- the channel to channel match is superb, and the overall deviation above the bass region is less than 0.5dB. The low cut filter was off, so the 2dB bass rolloff at 20 Hz is a bit odd.
With 5mV in, the 1kHz distortion spectrum looks like this:
Nice and low, much better than any cartridge will give you. Now although I limited the measurements because of the literally hundreds of setting options on this unit, curiosity got the best of me, and I repeated this measurement with the “Tube FX” on (I displaced the frequencies slightly for better visibility of the changes):
Not my cup of tea, and if I built a tube stage that bad, I’d probably slit my wrists in anguish.
Overload was measured at four frequencies (and remember that the generator voltage needs to be divided by 100):
That’s not inspiring performance. I’d want to see things an order of magnitude better, but I’m spoiled by my current gear…
THD+N versus frequency at 5mV/1kHz and RIAA pre-equalized:
This appears to be noise-dominated, and again, lower than any actual cartridge could provide..
Impulse response gives a nice insight into how the digital processing works:
Finally, a pre-equalized multitone:
All in all, other than the mediocre overload performance, this seems like a very well-engineered unit for the money. The feature set can be overwhelming, and is probably worthy of an entire article rather than this brief measurement suite.