Geshelli Archel2 Headphone AMP Review: another champ?

[Jave]

Will @amirm or @WolfX-700 measure the new 2.5 version? (Including multitone and THD+N vs freq)

 

[solderdude]

I have no idea what this shit mean. What's the secret sauce and why isn't everybody doing it to achieve almost distortion-free results?

Secret source for low distortion amps:
Choosing low distortion components (the right opamps).
Using the correct circuit and above all layout of the feedback path and decoupling.
High enough voltage rails.
Filtering those voltage rails.
Ensuring enough current can be available (selection of output devices and power supply)
Galvanic separation of audiopath, power supply and control circuit could also help.

All of the above is applied in this amp.

I don't know if they improved the low frequency noise in this version but could also be achieved by component choices but would have financial consequences.
I would use 10kA potmeter (RK271) but would add another $ 20.

It's just proper engineering. No special 'audio components' need to be used (other than suitable opamps) to get a good performing amp.
Obviously we don't know whether or not the magic unknown to us (yet) signal also passes through this device.
It would need to be tested by special ears for this.
Alas, according to those people, these folks are not among ASR members so would have to wait to hear from the 'better hearing' forums for that aspect. It is not expensive enough though...

 

[Jave]

...
I would use 10kA potmeter (RK271)...

Why 10k? Seems like they use 100k.

 

[solderdude]

lower noise, lower immunity to pick up stray RF and the RK271 choice is obvious.

 

[netcipher]

Instead of the built-in attenuator, you can use this one:
https://www.ebay.com/itm/DACT-Type-...ontrol-100K-Replace-Accessories-/274146774142

 

[mkawa]

lower noise, lower immunity to pick up stray RF and the RK271 choice is obvious.

i agree but i literally don't think a blue velvet would fit on the board. the rk16 they used is nothing to sneeze at, and it's nice to see something other than an rk09.

 

[trl]

Instead of the built-in attenuator, you can use this one:
https://www.ebay.com/itm/DACT-Type-...ontrol-100K-Replace-Accessories-/274146774142

A bit pricier than the original pot., only 21 volume levels is not enough, the inside 21 contacts will break faster than the carbon ribbon of a regular pot.

 

[netcipher]

A bit pricier than the original pot., only 21 volume levels is not enough, the inside 21 contacts will break faster than the carbon ribbon of a regular pot.

If we consider a universal attenuator for a headphone amplifier, then you can use this:
https://pl.aliexpress.com/item/32831375230.html
Supports both balance and unbalanced connections. Does not require soldering a headphone amplifier. Of course, you need to check at the interference from the transformer and other switching electronics
But 256 steps of regulation should be enough in excess

 

[JohnYang1997]

Secret source for low distortion amps:
Choosing low distortion components (the right opamps).
Using the correct circuit and above all layout of the feedback path and decoupling.
High enough voltage rails.
Filtering those voltage rails.
Ensuring enough current can be available (selection of output devices and power supply)
Galvanic separation of audiopath, power supply and control circuit could also help.

All of the above is applied in this amp.

I don't know if they improved the low frequency noise in this version but could also be achieved by component choices but would have financial consequences.
I would use 10kA potmeter (RK271) but would add another $ 20.

It's just proper engineering. No special 'audio components' need to be used (other than suitable opamps) to get a good performing amp.
Obviously we don't know whether or not the magic unknown to us (yet) signal also passes through this device.
It would need to be tested by special ears for this.
Alas, according to those people, these folks are not among ASR members so would have to wait to hear from the 'better hearing' forums for that aspect. It is not expensive enough though...

That depends on how low you were talking about. Atom level or THX level. It's truly different. There are a lot more to consider if you want -130db harmonics with high gain with low impedance load with unbalanced input and perhaps with balanced output (which is more difficult to achieve same performance than unbalanced out).
Capacitors, resistors in feedback, no star grounding, power supply choice and placement, just to name a few. These can affect another 12-20db performance from already decent engineering (eg no x7r in feedback path, star-grounding, copper pour etc).

 

[solderdude]

Doesn't this all fit under choosing the right components, circuit and PCB layout

 

[EVOLVIST]

Secret source for low distortion amps:
Choosing low distortion components (the right opamps).
Using the correct circuit and above all layout of the feedback path and decoupling.
High enough voltage rails.
Filtering those voltage rails.
Ensuring enough current can be available (selection of output devices and power supply)
Galvanic separation of audiopath, power supply and control circuit could also help.

All of the above is applied in this amp.

I don't know if they improved the low frequency noise in this version but could also be achieved by component choices but would have financial consequences.
I would use 10kA potmeter (RK271) but would add another $ 20.

It's just proper engineering. No special 'audio components' need to be used (other than suitable opamps) to get a good performing amp.
Obviously we don't know whether or not the magic unknown to us (yet) signal also passes through this device.
It would need to be tested by special ears for this.
Alas, according to those people, these folks are not among ASR members so would have to wait to hear from the 'better hearing' forums for that aspect. It is not expensive enough though...

Then how come everybody isn't making amps like this? The idea of getting exactly what comes out of a DAC should be the call to all, i.e. "A wire with gain." I mean, I'm just asking, because I don't know. You make it sound so easy. I have the 2.5 and I've never heard such clarity in a headphone rig since I had the SPL Auditor. Maybe not even then.

Can folks make a speaker amp with this kind of technology and low distortion that's not THX?

 

[JohnYang1997]

Then how come everybody isn't making amps like this? The idea of getting exactly what comes out of a DAC should be the call to all, i.e. "A wire with gain." I mean, I'm just asking, because I don't know. You make it sound so easy. I have the 2.5 and I've never heard such clarity in a headphone rig since I had the SPL Auditor. Maybe not even then.

Can folks make a speaker amp with this kind of technology and low distortion that's not THX?

Issue on this was already discussed on this thread before. I'll just briefly talk about it here.
Firstly, very rarely good engineers are doing audio nowadays. For two reasons, they don't value hifi much and audio electronics are very well researched, it's easy to make good audio products.
Then, it's the issue with how go it is.
If just able to resolve CD, most EE students can make it work in a product in a few weeks maybe months.
If it's over ppm(0.0001% or -120db), it's harder but still fairly straightforward and can be done fairly easily by good EE students with good source materials.
If it's sub ppm or maybe even -130db or -140db range of distortion and under 100nV supply noise and under 200nV noise across audible range, it requires more than a year for good EE students to accomplish with luck of encountering good reading materials and lots of experiments. This generally required much deeper understanding of circuits, components, and interrelation between practice and what's being taught including interfacing devices, interfacing actual environments.
But for most people this is seriously overdone and no one really need to do this. It's way better to spend the time and energy on something newer to us.

 

[solderdude]

Then how come everybody isn't making amps like this? The idea of getting exactly what comes out of a DAC should be the call to all, i.e. "A wire with gain." I mean, I'm just asking, because I don't know. You make it sound so easy. I have the 2.5 and I've never heard such clarity in a headphone rig since I had the SPL Auditor. Maybe not even then.

Can folks make a speaker amp with this kind of technology and low distortion that's not THX?

Even Schiit is making amps like these for even less... just not galvanically separated and not balanced input but more output power.

Speaker amps require different engineering because of different voltage and current levels.

A lot of manufacturers purposely make higher distortion amps which is more difficult to do with a specific type of distortion than just slapping a bunch of opamps on a 2 layer PCB with proper wire routing.
They often do this to separate them from others and sometimes the design is just crappy because the designer doesn't know any better or liked the sound better that way.
There are also manufacturers that actually try and succeed but usually cost more.

Schiit, JDS labs and Geshelli try this for as little money as possible. Others like to build something more substantial. Bigger, heavier, more power, more options, better longevity/parts, nicer looking.

I have no idea why you never experienced such clarity before.

 

[netcipher]

That depends on how low you were talking about. Atom level or THX level. It's truly different. There are a lot more to consider if you want -130db harmonics with high gain with low impedance load with unbalanced input and perhaps with balanced output (which is more difficult to achieve same performance than unbalanced out).
Capacitors, resistors in feedback, no star grounding, power supply choice and placement, just to name a few. These can affect another 12-20db performance from already decent engineering (eg no x7r in feedback path, star-grounding, copper pour etc).

It was possible not to raise the topic of an external relay attenuator, but other discrete controls did not fit inside the amplifier. It is clear that to obtain an additional 10.20 dB to the characteristics, other design approaches are required. We can say this is one of the approaches to abandon the shortcomings of digital attenuators. If you use a passive relay attenuator with memory position contacts, then turn off the power after setting the volume level.
Archel 2.5 is far from THX amplifiers in quality of implementation, but in conjunction with Anedio D2 gave very good characteristics and sound.

 

[majingotan]

liked the sound better that way.

It's mostly this. Them who don't give a crap about measurements give them leeway to tweak/color the sound to the way they like. They should think like those engineers at Mola Mola who know how to make their products sound great (for the price they better be) while providing absolute transparency.

 

[EVOLVIST]

I have no idea why you never experienced such clarity before.

It probably is a combo of hearing my Linn KDS/3 for the first time via cans. Plus the amp is damn fine. Couple that with my new cans and I think I've reached a serendipitous symbiosis with my HP rig now.

 

[Blur]

Quick question for you all: If there are XLR inputs on the back of this amp, but no balanced out for the headphone, does the output still roughly double using XLR in? It looks like it did for Amirm's testing, but unless the innards are balanced out to SE I don't see how it is possible.

 

[JohnYang1997]

Quick question for you all: If there are XLR inputs on the back of this amp, but no balanced out for the headphone, does the output still roughly double using XLR in? It looks like it did for Amirm's testing, but unless the innards are balanced out to SE I don't see how it is possible.

It can, with the most common configuration of differential amplifier.

 

[Blur]

So it's differential to single-ended? If that is truly the case then I'm even more stoked as I'll end up with more power for SE output.

 

[Deleted member 12179]

It's a beautiful industrial design. Everybody say it's ugly, no better proof than that about what a beautiful design it is. It looks total punk, total outsider, looks ghetto, sounds like angels. The Basquiat of the amplifiers. Matches really well the aesthetics of a khadas tone board with clear case.



My only concern, just my personal case needs, that it has no pass-through output for RCA to speaker monitors. I wonder how Topping A50 will compare to this and JDS Atom. Thanks for the awesome reviews!!