Geshelli Archel2.5 Pro Review (headphone amplifier)

[Nango]

@amirm: Why don't you measure the current the head phone amps do deliver? While voltage output is important for dynamic headphone current capacity is more important than voltage for planars (at least I thought it is). Would be good to know this figure as it is rarely disclosed by the mfr.

 

[solderdude]

It may help some but it is possible the internal design can't dissipate more power.

Indeed the max power output (and thus input) is limited by 2x 2W DCDC converters.
The 'knee' in low impedance headphones at around 300mW is caused by the 15V rails out of the converters 'sagging' under load.
At 32 Ohm and 300mW it puts out 3.1V RMS = 0.1A (0.14A peak, 4.4V peak) and when both channels are driven 0.3A peak @ 15V = 4.5W consumed peak power (3.2W nominal from 2W rated converter).
Depending on how large the reservoir caps are after the converter the actual distortion 'knee' (still inaudible low) will move upwards a bit as peak powers (short ones) will be drawn from those reservoir caps.

@amirm: Why don't you measure the current the head phone amps do deliver? While voltage output is important for dynamic headphone current capacity is more important than voltage for planars (at least I thought it is). Would be good to know this figure as it is rarely disclosed by the mfr.

See my reply to Amir.
available (peak) current in dynmaic conditions may be slightly higher than the nominal currents shown above.
It is basically limited by the DCDC converters. These DCDC converters are to be able to have +/- 15V rails, provide extra galvanic separation and limit currents in output devices.

Amir does post current limits but you have to calculate them from the output power in 32 Ohm plots.

 

[Anmol]

I have been listening through this amp for last few months with my 600 ohm beyerdynamic T1 2nd gen- subjectively this amp sounds absolutely superb. Works fine with my headphone on low gain. I find it better sounding to me than Atom, magni 3, beyer a20 and nx4.

 

[MKreroo]

Feeling a little down seeing this, I thought one of the biggest points for the 2.5 is the increase in max power, but it seems that at 300ohm it delivers less power than the 2 both on HG (assuming the gain is the same?).
As a 2.5 owner myself, the Channel Imbalanced is very noticeable at low volume (about 10%? of the full), especially with IEM (tried with SE215)

 

[JohnYang1997]

Thank you for the review Amir! One thing I don't understand is why the SINAD is only -116dB when distortion are around -130 and the signal to noise ratio is -124dB? I thought the SINAD should be noise limited, so must be -124dB.

SINAD is noise of the amp + the noise of AP. AP has higher input referred noise when the input is higher voltages. So in short it has to be much much better than 124db SNR to have equally good SINAD.
You can look at previous Archel 2 or L30 measurements for more insight.

 

[bigjacko]

it has to be much much better than 124db SNR to have equally good SINAD.

Is SINAD limited to AP's input referred noise? For example if the noise is -125dB, even the amp is way lower than that, we still cannot get SINAD lower than -125dB. Also does noise of AP change with input voltage because it has to change something inside AP?

 

[JohnYang1997]

Is SINAD limited to AP's input referred noise? For example if the noise is -125dB, even the amp is way lower than that, we still cannot get SINAD lower than -125dB. Also does noise of AP change with input voltage because it has to change something inside AP?

Sqrt(A^2+B^2). If you completely eliminate A(device under test) you get B(AP). Other wise you get something in between. But if you want to get A accurately, B needs to be much smaller than A as well. In between you get the noise of the device under test plus the noise of AP. You will see difference(smaller noise, better SINAD) in this region but the absolute result will not be accurate anymore(worse SINAD than SNR). Even in SNR, the limitation can be AP.

 

[YSC]

it looked nice for the max output SINAD, but to me it seems not a good option for headphones, since most if not all headphone suitable for this amp would want to use the volume knob in the mid to low range of the pot, where the tracking is poor from the channel balance department should make the sound stage and image start to go off quite quickly in the attenuation range, which seems more audible than 120db SINAD vs say 90 SINAD?

 

[Matias]

Channel imbalance could be better. They should test pots individually (binning) to ensure better tracking, maybe at a price premium, say +20 usd for guaranteed -40dB attenuation withing +/- 0.5 dB tracking?

 

[restorer-john]

They should test pots individually (binning) to ensure better tracking, maybe at a price premium, say +20 usd for guaranteed -40dB attenuation withing +/- 0.5 dB tracking?

Or just buy/specify better pots in the first place. The ones they use are bottom of the barrel.

It's just another case of an otherwise excellent performing product, let down by cost cutting with the part that gets used the most (the volume pot). When will these people learn?

And, excellent review @amirm

 

[PeteL]

@amirm: Why don't you measure the current the head phone amps do deliver? While voltage output is important for dynamic headphone current capacity is more important than voltage for planars (at least I thought it is). Would be good to know this figure as it is rarely disclosed by the mfr.

I=sqrt(P/R)

 

[YSC]

Or just buy/specify better pots in the first place. The ones they use are bottom of the barrel.

It's just another case of an otherwise excellent performing product, let down by cost cutting with the part that gets used the most (the volume pot). When will these people learn?

And, excellent review @amirm

And possibly the part affecting user impression most, I am pretty sure 1db channel imbalance makes sound worse than 80db SINAD..

 

[anmpr1]

A real engineer will not make plexiglass chassis.

The box or chassis appears to be metal (or plastic looking like metal?). The front/back panel is plexiglass (or clear plastic looking like plexiglass?). At this price point, is this really an issue for anyone?

 

[okok]

snr much more important than thd+n
edit: the 50mv snr

 

[PeteL]

The box or chassis appears to be metal (or plastic looking like metal?). The front/back panel is plexiglass (or clear plastic looking like plexiglass?). At this price point, is this really an issue for anyone?

Well, these Headphone amps reviews are getting quite predictable, they almost all measure very well, so yes it's totally justified to buy a desktop amp based on look.

 

[restorer-john]

Actually, I suspect there is an entire class of nerd engineer that thinks plexiglass is cool.

The same nerds that just can't get enough of RGB LEDs in their gaming pooders. My 13yo son just bought RGB RAM and an RGB PSU for his his new gaming pooder build, to go with his RGB mouse and Keyboard...

The Geshelli products are just hilarious. No mature grown adult would put them in a HiFi system. Maybe on Billy-Bob's 16yo computer desk, in Mom's dank basement, perhaps. Nasty cheap extrusion, perspex end caps and arguably the most putrid implementation of controls I have ever seen. Just horrible in every respect, and a very, very, marginal step above the CNC perspex boxes all over eBay for "makers" and "hobbyists".

Anyone can put a quality performing line stage in a horrible box. It's not a big ask. Why people would clamour to buy such a hideous thing is beyond me. Buy the Topping, any Topping, for goodness sake. At least they tried.

 

[anmpr1]

The Geshelli products are just hilarious. No mature grown adult would put them in a HiFi system. Anyone can put a quality performing line stage in a horrible box. It's not a big ask. Why people would clamour to buy such a hideous thing is beyond me.

In the context of a twenty thousand dollar music streamer named after a pink goat-man, this little Geshelli is looking pretty good! LOL

https://www.stereophile.com/content/pink-faun-216x-music-streamer

 

[phoenixdogfan]

Never been a fan of plexiglass either. Then again, my instinct is to put everything in a BDI cabinet and run all the cables through the walls and under the floorboards.

 

[Billy Budapest]

SINAD calculation looks wrong, just eying the noise floor and distortion levels on the graph. Weird.

 

[Billy Budapest]

The same nerds that just can't get enough of RGB LEDs in their gaming pooders. My 13yo son just bought RGB RAM and an RGB PSU for his his new gaming pooder build, to go with his RGB mouse and Keyboard...

The Geshelli products are just hilarious. No mature grown adult would put them in a HiFi system. Maybe on Billy-Bob's 16yo computer desk, in Mom's dank basement, perhaps. Nasty cheap extrusion, perspex end caps and arguably the most putrid implementation of controls I have ever seen. Just horrible in every respect, and a very, very, marginal step above the CNC perspex boxes all over eBay for "makers" and "hobbyists".

Anyone can put a quality performing line stage in a horrible box. It's not a big ask. Why people would clamour to buy such a hideous thing is beyond me. Buy the Topping, any Topping, for goodness sake. At least they tried.

Huh. I think the Geshelli products look great, especially in the smoked plexiglass and black aluminum.