ifi Zen CAN Review (Headphone Amp)

[Leiker535]

On full gain(+18) with the hd600 or Ananda 8 o'clock is loud, 9 is too loud with pain starting.

What are you feeding the amp? My zen can sig is unbearably loud even with 2V at max on unity gain with the 650s. I measured 120 dB at the 18dB gain setting, and 110 dB at 0db, both with a 1khz pure tone while using a dbmeter (A weighted, though it wouldnt matter). Alas, I don't have a 4V DAC at hand, but it would be even louder still. This amp, while I had the Zen DAC sig v1 at fixed through the 4V 4.4mm out, was too loud for the 6xx even at around 9 hours at the 0 gain setting, so much so that I prefered feeding it SE or halving the output using variable.

 

[Leiker535]

4 Volt Balanced input and 6V when switched to variable on my Zen-Dac.

I wouldn't feed the 6V from variable if you have the Zen DAC signature, as it adds a bit of HD to the mix, and unnecessarily so as the amp is already plenty powerful with 4V. Stereophile states that they fixed it on the V2, but I can't attest to that as I don't have one. Here is my measurement while using a Focuriste (disregard the dBC, just mind the THD):

While oversaturating the rca out at variable:

While keeping it at fixed:

 

[pioterd]

My zen can sig is unbearably loud even with 2V at max on unity gain with the 650s

Similar experience here, with DT150 and K712. If I feed ZenCan with 2V RCA, unity gain is absolutely all I need, sometimes even I’d like it to be quieter with volume knob at around 9’ o’clock, during the silence of the night.

halving the output using variable.

Is there any negative consequence in sound quality when switching from fixed to variable?

 

[Leiker535]

Is there any negative consequence in sound quality when switching from fixed to variable?

Depends on how good your ears are at finding harmonic distortion, or even if you like it. With variable, the V1 only has this behaviour when you push it past 2V for the RCAs and 4V for the 4.4mm out, this would be near 3 a'clock on the pot from what I could test. Below that it's completely normal.

 

[LaL]

Funny, this looks like the original 1.0 PCB with the U8 8-Core XMOS. So that's basically my original design.

Funny how it got to 2.1... From 1.0...

Thor

The first few for sale after the first DROP had a sticker over the original Logo.

 

[pioterd]

With variable, the V1 only has this behaviour when you push it past 2V

Thanks. And you can do this only if you turn on power match (gain), without it, V1 will not even provide 2V in Variable mode, but something around 1.5V. Did you notice distortion when powered with just usb cable? Or with power supply at least 1A/5V as well?

 

[Thorsten Loesch]

I had this problem with a cheap 4.4mm interconnect cable between the Dac and the Can amp, I still use this cheap cable but I have to bend it a certain way and it's fine if I don't move it

I'm embarrassed to admit to using a prototype iFi cable. It certainly makes perfect contact and the braiding and silver plated copper may improve sound further than just a reliable connection.

What are you feeding the amp? My zen can sig is unbearably loud even with 2V at max on unity gain with the 650s. I measured 120 dB at the 18dB gain setting, and 110 dB at 0db, both with a 1khz

This suggests at 18dB the Amplifier is clipping or the headphone is severely overloading (or both) as you should be measuring 128dB.

Alas, I don't have a 4V DAC at hand, but it would be even louder still.

It would not. The way the Amplifier is designed, each input has different gain.

So the Balanced Input expects 4V BAL, has 22kOhm input Impedance and then converts the input to 2V SE.

The RCA Input expects 2V SE, has 1MOhm input impedance and just operates as unity gain buffer, outputting 2V SE.

The 3.5mm input expects 1V SE, has 1MOhm input impedance and 6dB Gain, outputting 2V SE.

So each input ultimately gives 2V on the volume control if the nominal input is applied.

If the Zen Can is set to Unity Gain and the volume control is set to max, then

4V BAL IN = 4V BAL OUT
2V RCA IN = 2V SE OUT (4V BAL OUT)
1V 3.5 IN = 2V SE OUT (4V BAL OUT)

This amp, while I had the Zen DAC sig v1 at fixed through the 4V 4.4mm out, was too loud for the 6xx even at around 9 hours at the 0 gain setting, so much so that I prefered feeding it SE or halving the output using variable.

That sounds wrong.

With the volume control at ~ 12 O'Clock the output should ~200mV SE / ~ 400mV BAL for a Zen DAC (Bal or SE) with 0dBFS.

Worst compressed music is around -12dBFS average, or ~50mV SE / ~ 100mV BAL in this setup for average output (not peak)

The HD6XX is 103dB/1V so on the Balanced output average SPL with heavily compressed music should be 83dB and 77dB with SE Out.

I wouldn't feed the 6V from variable if you have the Zen DAC signature, as it adds a bit of HD to the mix, and unnecessarily so as the amp is already plenty powerful with 4V. Stereophile states that they fixed it on the V2, but I can't attest to that as I don't have one. Here is my measurement while using a Focuriste (disregard the dBC, just mind the THD):

While oversaturating the rca out at variable:

I don't know about that, it should not be according to design and prototypes. Perhaps a manufacturing error not picked up by QC.

When I was still working there, I approved the Standard Operating Procedures (SOP) for Quality Control prepared by chinese "engineers". Invariably several rounds of corrections and resubmission were needed, as it was obvious that these "engineers" skill set was at it's limit with anything beyond brainless copy/paste.

This led to a tension, as I was told to just approve these SOP's so production could go ahead when in fact these SOP's would not be able to to pick up problems. Pointing out that providing inadequate SOP's for approval on Friday evening for approval with production starting monday morning was not reasonable got brushed off.

The input stage includes analogue switches that should run on +/-5.1V and handle +/-5.6V signals (4V RMS on SE, 8V RMS on BAL). Of course, if someone messed up in the production this may be incorrect. It's just how the Zen Can was designed.

Is there any negative consequence in sound quality when switching from fixed to variable?

Yes. The variable output on the Zen DAC passes through the Headphone Amplifier, so there will be more noise and distortion on the Vairable output and IMHO the sound quality is a little worse than fixed which is directly from the DAC.

Thanks. And you can do this only if you turn on power match (gain), without it, V1 will not even provide 2V in Variable mode, but something around 1.5V.

Correct. Whoever called this "power match" is a marketing futzy.

It is a gain switch. with the "low gain" setting, the headphone amplifier gain (and thus variable line out gain) is unity and the DAC output is attenuated by around 6dB before the volume control. So the low gain headphone & line output should deliver around 1V RMS max. SE & 2V BAL.

This is mainly intended for high sensitivity Headphones, but may also be useful with active speakers designed for 3.5mm Input at 1V.

The "high gain" setting removes the attenuation before the volume control and switches the Headphone Amplifier to around 5dB Gain giving the maximum around 3.1V RMS max SE & 6.2V BAL.

This is mainly intended for low sensitivity headphones.

In case someone paid attention, as the "signature" version is based on the standard PCB and has a variable output, yup, all the headphone amplifier and even the switching for gain and bass boost are present in the "signature" and just switched to bass boost off and high gain. So yes, you pay more for the removal of the headphone Jack's and push-buttons.

Naturally I argued for a version with fixed output only and all the headphone section really removed, or a version with the headphone amplifier retained. Instead it ended up as is.

Now AGAIN, all switching is CPU control and it is possible that there is a programming mistake and the Zen DAC does not work as designed and it is also possible that this was never picked up, as I did not do final testing.

I did test prototypes and picked a range of hardware and software bug's and documented the fixes. Maybe this was mislaid and never fixed...

With the volume control set to max and high gain the variable output will approach clipping, since it is designed to be "Volume Controlled" output, not as some kind of "signal booster".

If the device following the Zen DAC has a volume control, use the fixed output

Did you notice distortion when powered with just usb cable? Or with power supply at least 1A/5V as well?

Should not make any difference, if the USB cable and USB port provide 5V.

Thor

 

[pioterd]

around 6dB before the volume control

Thank you Mr Loesch once again! Isn’t it more like +10db gain (power match)? Some calculations seem to say that from around 1.4V to 3.3V, 10dB gain is needed.

 

[staticV3]

Some calculations seem to say that from around 1.4V to 3.3V, 10dB gain is needed.

Just do the calculation yourself:
20*log10(3.3/1.4)=7.4477dB

 

[Thorsten Loesch]

Thank you Mr Loesch once again! Isn’t it more like +10db gain (power match)? Some calculations seem to say that from around 1.4V to 3.3V, 10dB gain is needed.

Remember, there is both attenuation (0dB or -6dB) before the Volume control and gain in the headphone amplifier (0dB or ~ 4.6dB).

Of course, the chinese "engineers" may have changed values from my design or not correctly implemented the software control of all switches.

So what I write is about design intention, not produced design, which without having a unit and testing I cannot say much..

It would for example seem that to get 1.4V output the HP Amp is never switched over to unity gain, because with 6dB attenuation and 0dB Gain, the maximum output should be 1V, no more than that. This would produce more noise than needed.

Typical problem, chinese "engineers" generally lack the ability to analyse circuits to determine how they should work and the ability to read notes and comments on schematics that clearly state the relevant details (they could double-click and copy paste into any on line translator, but even notes actually in chinese tend to be ignored). Obviously the gain button did something and thus was considered working, even if it was not working as intended.

Thor

 

[Leiker535]

If the Zen Can is set to Unity Gain and the volume control is set to max, then

4V BAL IN = 4V BAL OUT
2V RCA IN = 2V SE OUT (4V BAL OUT)

1V 3.5 IN = 2V SE OUT (4V BAL OUT)

This very useful information to have. With most amplifiers I tried, you had to feed it bal for it to deliver the max power ratings. It would mean that it really doesn't matter if the source is bal or not, aside from generally better Snr, right?

This suggests at 18dB the Amplifier is clipping or the headphone is severely overloading (or both) as you should be measuring 128dB.

I may have indeed, as I was feeding it through the 3.5 input while using a 2v dac. I wouldn't put too much faith in my uncertified dbmeter precision for such high figures also.

That sounds wrong.

With the volume control at ~ 12 O'Clock the output should ~200mV SE / ~ 400mV BAL for a Zen DAC (Bal or SE) with 0dBFS.

To clarify, I was referring the bal out of the zen can, not the zen dac/amp.

 

[pioterd]

2V RCA IN = 2V SE OUT (4V BAL OUT)

Maybe I misunderstood somewhere, but if it goes that way, and we still have +18dB gain (8x) then the maximum voltage output would be 16V SE and 32V BAL? Specification gives 8V SE/16V BAL values as a maximum for Can.

Of course, the chinese "engineers" may have changed values from my design or not correctly implemented the software control of all switches.

So what I write is about design intention, not produced design, which without having a unit and testing I cannot say much..

Reading something like this, Im surprised that the final product may not follow main engineer (constructor) expectations? Quality control/verification of the final result should be part of the whole process, don't you think so? After putting so much effort to develop good product, final stage failure would be disappointing :-/

 

[JaMaSt]

Correct. Whoever called this "power match" is a marketing futzy.

It is a gain switch. with the "low gain" setting, the headphone amplifier gain (and thus variable line out gain) is unity and the DAC output is attenuated by around 6dB before the volume control. So the low gain headphone & line output should deliver around 1V RMS max. SE & 2V BAL.

This is mainly intended for high sensitivity Headphones, but may also be useful with active speakers designed for 3.5mm Input at 1V.

The "high gain" setting removes the attenuation before the volume control and switches the Headphone Amplifier to around 5dB Gain giving the maximum around 3.1V RMS max SE & 6.2V BAL.

This is mainly intended for low sensitivity headphones.

My own experience with the Zen DAC v2, using my Unique Melody Mest Mk2 IEM (112dB, 12 Ohm), is that the 4.4 headphone output of the Zen sounds much better with an added 20 Ohm impedance adaptor (bringing the IEM up to ~32 Ohm). Without the adaptor, as you turn up the volume, you get the typical, shouty, "too high gain" sound. It's not bad per se, it's just that the increase in volume without the adaptor is not uniform across the entire spectrum. The bass only gets a little louder why the uppers get exponentially louder.

 

[Thorsten Loesch]

This very useful information to have. With most amplifiers I tried, you had to feed it bal for it to deliver the max power ratings.

Not sure about that. If you look at competently designed BAL & SE input Headphone amplifiers, such as the Neve RNHP or the Bryston BHA-1 they work the same way as Zen CAN.

In fact, the Zen CAN was in part designed against the Neve RNHP in terms of functionality (it uses a very different circuitry) and with reference to products like the Bryston BHA-1, SPL Phonitor mini and Violectric's Balanced designs.

Yes, it was designed for a 100 USD retail price-point, but it was designed to be competitive in terms of features and performance to compete against most high end balanced headphone amplifiers.

It would mean that it really doesn't matter if the source is bal or not, aside from generally better Snr, right?

Correct. And do NOT assume that balanced connections offer inherently better SNR.

For example, the input circuitry of the Zen Can offers 126dB SNR for a 2V SE input but only 118dB for 4V BAL input. Actually the 3.5mm SE input with 1V input has the same SNR as the BAL input with 4V!

There are several reasons for that, but the key one is that balanced input circuitry uses differential amplifiers which, compared to the same Op-Amp used as Buffer has 6dB more noise. Add some extra noise from resistors and we get the 8dB difference.

This is common for many balanced inputs, including those using expensive "balanced receiver IC's" from TI or THAT.

Also, balanced Headphone Outputs will, on the same headphone and same SPL at least 3dB or 9dB noisier than the same design in SE.

Honestly, Balanced for headphones and home audio is a stupid fad if you ask me. It actually delivers poorer performance for more money.

To clarify, I was referring the bal out of the zen can, not the zen dac/amp.

Ok, the Zen CAN will have at best 0dB gain, not -6dB. If you then use the 3.5mm input with 2V you will have in effect 6dB gain as minimum.

Maybe I misunderstood somewhere, but if it goes that way, and we still have +18dB gain (8x) then the maximum voltage output would be 16V SE and 32V BAL? Specification gives 8V SE/16V BAL values as a maximum for Can.

8V/16V is the maximum output from the actual amplifier. So there is some excess gain available for weak sources or quiet recordings. This is common design practice.

Reading something like this, Im surprised that the final product may not follow main engineer (constructor) expectations? Quality control/verification of the final result should be part of the whole process, don't you think so? After putting so much effort to develop good product, final stage failure would be disappointing :-/

Well, if you manufacture in China, under chinese management and quality control and with a chinese engineering department and you insist on having very large numbers of very different products because your sales & marketing department is incompetent to sell a really good product using "push" and thus to increase sales they always need more and newer products, then this is a direct and unavoidable result.

This state together with other factors led to my departure. The Zen range was actually my last throw of the dice to change the way things worked. While I did succeed in pushing the products through, in the end sales and marketing again intervened making what would have been absolute market leaders at 99 USD per pcs (Zen DAC, Zen CAN, Zen PHONO, the signatures were intended to cost 149 USD) and would have sold in much bigger numbers. As is, at double that the product is kinda meh and of course does not sell as well. I guess the profit margins look better.

Thor

 

[Thorsten Loesch]

My own experience with the Zen DAC v2, using my Unique Melody Mest Mk2 IEM (112dB, 12 Ohm), is that the 4.4 headphone output of the Zen sounds much better with an added 20 Ohm impedance adaptor (bringing the IEM up to ~32 Ohm).

I would not recommend the balanced out for low impedance headphones.

It uses the MAX97220 Headphone Amplifier IC, 1pcs per channel. The IC is rated for 100mW into 16 Ohm and 125mW into 32 Ohm. I would suggest that 32 Ohm is the lowest impedance that should be used on the BAL out and 16 Ohm on the SE out.

These "Impedance Adaptors" change the frequency response of IEM's and also transient behavior. Results tend to be unpredictable.

Try your IEM on the SE out without "impedance adapter".

Without the adaptor, as you turn up the volume, you get the typical, shouty, "too high gain" sound. It's not bad per se, it's just that the increase in volume without the adaptor is not uniform across the entire spectrum. The bass only gets a little louder why the uppers get exponentially louder.

Well, that is perception. I think what you experience is the "Aphex" effect. Overloading the output stage gives a lot of relatively low order and lowish level distortion, that tends to make mid's sound louder.

I like to use the "Aphex" effect together with MaxxBass when setting up PA/Club Systems as they give the impression of more and deeper bass plus a louder playback level, without actually outputting extra SPL. Using these effects wisely literally can cut the size and cost of a PA system in half or give extra headroom for the same money and size.

But yes, you are operating the HP Amplifier outside the design envelope. I would suggest to not do that.

Thor

 

[Veri]

These "Impedance Adaptors" change the frequency response of IEM's and also transient behavior. Results tend to be unpredictable.

But yes, you are operating the HP Amplifier outside the design envelope. I would suggest to not do that.

Thor

Indeed. Was thinking the same

 

[JaMaSt]

I like to use the "Aphex" effect together with MaxxBass when setting up PA/Club Systems as they give the impression of more and deeper bass plus a louder playback level, without actually outputting extra SPL. Using these effects wisely literally can cut the size and cost of a PA system in half or give extra headroom for the same money and size.

But yes, you are operating the HP Amplifier outside the design envelope. I would suggest to not do that.

Thanks for the input. The "Aphex" description you give is a much more concise explanation of what I was hearing.

When I made the post, I'd already had the 4.4mm iFi iEmatch in transit. Having received it, I hear less "bloom" in the low and mids and more detail than with the impedance adaptor. I wanted the Zen DAC to watch YouTube videos and do some gaming. It's really a nice product. My main gear for my ripped flac files and wifi streaming is the iBasso dx312, but the audio/video latency is a no-go when using it with USB from my PC. I'm pleased with the product. It's heft, low-profile an finish are really nice.

 

[pioterd]

So there is some excess gain available for weak sources

Naturally I argued for a version with fixed output only and all the headphone section really removed, or a version with the headphone amplifier retained. Instead it ended up as is.

Thank you. ZenDac V1 or V2 seems to be very nice option still, sounds great, it’s driving abilities are nice for not very demanding headphones…voltage is enough but if it provided more current would be even more competent Dac/amp device.
Indeed, with power match off, ZenDac outputs essentialy very similar power as ZenCan+0.5V 3.5mm input (at unity gain).

 

[Thorsten Loesch]

Thank you. ZenDac V1 or V2 seems to be very nice option still, sounds great, it’s driving abilities are nice for not very demanding headphones…

That was the design brief.

Originally it was intended to sell at 99 USD, it was important to not make it "too good" to allow space for a dedicated HP Amp and to not make internal competition for higher price products.

Indeed, with power match off, ZenDac outputs essentialy very similar power as ZenCan+0.5V 3.5mm input (at unity gain).

Yup, BUT Zen CAN has a few more gears and goes not just to 11, it goes to 18!

Thor

 

[pioterd]

Yup, BUT Zen CAN has a few more gears

yup and that’s why, with whole respect to yours involvement in ZenDac, I prefer such combination (low output good quality dongle + ZenCan) rather than just ZenDac for desktop use.