Topping A50s Headphone Amplifier Review

[audio_tony]

Another thought on the low input impedance.

In the quest for low noise, a lot of opamp circuits use quite low values in the feedback loop.

So if you have a source with an opamp in the output stage, and this opamp has typical values of 2k (Rf) and 1k (Rin) in the feedback loop, that opamp is already driving 3k.

Connect that source to this headphone amp, and you have 3k in parallel with the 2.7k input impedance, and your source is now expected to drive a 1.4k load.

I know of many opamps (OPA2134 which is a popular one) that have rising THD when driving loads below 5k or so - therefore in such a situation using this headphone amp is likely to result in higher distortion from the source...

 

[solderdude]

Most opamps have no issues driving 600 Ohm loads. (there are exceptions but are rarely used as line-level output drivers)
With a 2.7k load this means Rf in the line driver into the amp can be as low as 800 Ohm.

 

[audio_tony]

Most opamps have no issues driving 600 Ohm loads. (there are exceptions but are rarely used as line-level output drivers)
With a 2.7k load this means Rf in the line driver into the amp can be as low as 800 Ohm.

I think Douglas Self would disagree with you. Look at the rising THD at HF as impedance drops.

The NE5532 and LM4562 would both be fine though. I chose the OPA2134 as I know it's a popular amp for use in DACs (my Audiolab Q DAC for example has 2x OPA134 - but they drive a class A output stage)

I realise this is using shunt feedback vs series feedback, but they doesn't make a huge amount of difference in this case.

 

[solderdude]

Most used modern opamps won't have much problems.
But yes, heavy loads usually lead to higher distortion.
It always makes me smile seeing 'headphone output stages' using regular opamps and C'Moy amps driving 32 Ohm headphones when they are barely suited to drive 600 Ohm loads. Yet most find it sounds fantastic.

The ones in the cheap Topping DACs don't seem to have much issues. At least I expect John tested this..
Would be daft to design an amp with low noise and distortion to have those advantages null and void when connected to sources.

 

[audio_tony]

Most used modern opamps won't have much problems.
The ones in the cheap Topping DACs don't seem to have much issues. At least I expect John tested this..
Would be daft to design an amp with low noise and distortion to have those advantages null and void when connected to sources.

Maybe most 'modern opamps' but there are a lot of designs out there still using older opamps for various reasons (cost, cult following etc. etc. etc.).

And even with a modern opamp that can happily drive super low impedance loads, it's likely that any new design will exploit this to the full by employing very low values for Rf and Rin anyway - which just moves the problem.

IMHO - an input impedance of 2.5k is ridiculously low and unnecessary, as it's still possible to achieve low noise without resorting to such a low impedance.

10k is about as low as any design should go, and in an ideal world I would suggest 22k is even better.

 

[solderdude]

My designs have a 20k input resistance but do not reach such low noise levels. It is essential to be this low when you want low noise levels with a the potmeter directly at the input.

Note that the output resistance of a source (between a few and 100 Ohm) is always in parallel to the input resistance of the amp.
A 20k potmeter at -6dB setting would increase that to 10k. In case of the 2.5k input R this would only be 1.2k.
That would lower the noise level which John is trying to get as low as possible.

 

[audio_tony]

Input impedance is stated as 2.5k - we don't know if this takes the pot into account (or even if the pot is 2.5k)

I appreciate he's trying to get the noise as low as possible, but raising the input impedance to 10k is not going to affect the noise that badly.

This is just 'specmanship'.

 

[solderdude]

there are a lot of designs out there still using older opamps for various reasons (cost, cult following etc. etc. etc.).

Yes, there sure are but would they be using an L30 or A50S or something exotic with tubes and 1M input R ?

I agree 2.5k is on the low side but low noise performance with high efficiency IEM's is a thing today.

Input impedance is stated as 2.5k - we don't know if this takes the pot into account (or even if the pot is 2.5k)

He did mention this is the case with the L30.
But sure.. one could put a 1x buffer before a low resistance pot and use a 10k input R. This would increase the noise level though.

 

[l1tmn]

Dear amirm,

Can you attach a capacitive load (100pF or 1nF) and watch the square wave response on an oscilloscope?

I feel that the ultimate low distortion amplifiers these days have very little phase margin due to the circuit configuration with very high open loop gain.

I'm sure they have special phase compensation like Neurochrome, but I'd like to see how well they hold up to capacitive loading by the square wave response.

I would like to see how well it works under capacitive load by the square wave response. I hope it will be included in the evaluation of the amplifier in the future.

 

[JohnYang1997]

Dear amirm,

Can you attach a capacitive load (100pF or 1nF) and watch the square wave response on an oscilloscope?

I feel that the ultimate low distortion amplifiers these days have very little phase margin due to the circuit configuration with very high open loop gain.

I'm sure they have special phase compensation like Neurochrome, but I'd like to see how well they hold up to capacitive loading by the square wave response.

I would like to see how well it works under capacitive load by the square wave response. I hope it will be included in the evaluation of the amplifier in the future.

80 degree + phase margin. At least is stable with 10nF.

 

[l1tmn]

80 degree + phase margin. At least is stable with 10nF.

I don't think this is the case if you have a capacitance load measure such as a network of inductors and resistors (capacitance isolator) in series with the output of the amplifier.

I would like to confirm that the product does not oscillate even with a capacitive load.

 

[Atanasi]

You can also have a ground on a 4-pin connector, namely the screen is the 5th (ground) connection.

Is this the expected wiring for 4-pin XLR? I think 3-pin XLR cables are not supposed to wire the chassis to the screen, they should rather use pin 1.

 

[Veri]

Is this the expected wiring for 4-pin XLR? I think 3-pin XLR cables are not supposed to wire the chassis to the screen, they should rather use pin 1.

3-pin XLR cables are dual mono... 6 pins. 4-pin XLR obviously only has 4 pins.

 

[Atanasi]

3-pin XLR cables are dual mono... 6 pins. 4-pin XLR obviously only has 4 pins.

The question is whether to connect the connector chassis to the screen, or leave the chassis floating like 3-pin XLR.

 

[solderdude]

When a 3 pin is properly connected the screen is also used but not connected to pin-1 (signal ground) but the chassis (also ground).

It makes sense to connect the screen to ground and not leave it floating.

 

[Mashcky]

What’s to stop someone from using a headphone amplifier as a speaker amp for high sensitivity or desktop speakers? Is it that they can’t produce enough current for low impedance (compared to headphones) speakers?

 

[solderdude]

Nothing can stop you from trying. The question is what would you do with 0.8W ?
In the end it is not good for the amp as it is not designed for this and can overheat (is protected for this though so will shut down)
I would recommend to use speaker amps for speakers and headphone amps for headphones.
See the data sheet ... minimum impedance 8 Ohms... headphones with 8 Ohm impedance that is.

 

[Pillars]

Seriously this... what good is a 4V in measurement if that's not something easily accessible to buyers? Yea I happen to have a Jensen 1:1 Transformer lying around but it's not 100% transparent sooo there's that... but I'm disappointed to not have any balanced inputs... if your running 4.4mm out why not also just run a pair of 3pin XLR ins? I mean P50 is quite LARGE so it doesn't seem like space was a concern when designing that... can't understand why this didn't at least have balanced inputs. Heck SP-200 has 3pin in even tho yea it's technically a single end amp.




Ehh there's Schiit Modius that's pretty outstanding...

If you've got a balanced DAC the Modius makes sense for sure, especially with the better quality pot and BAL pre outs. Otherwise, this A50s looks to perform pretty well based on measurements and the 4.4 jack is nice. I'd have hoped for better 1/4 power but it's not a showstopper.

 

[JohnYang1997]

If you've got a balanced DAC the Modius makes sense for sure, especially with the better quality pot and BAL pre outs. Otherwise, this A50s looks to perform pretty well based on measurements and the 4.4 jack is nice. I'd have hoped for better 1/4 power but it's not a showstopper.

Better quality pots don't mean they are better. In terms of power, A50s itself has the ability to drive to very high power even with 6.35mm. It's just that the gain is set to a low value. So feel free to turn it all the way up, it won't break a sweat.

 

[Pillars]

Better quality pots don't mean they are better.

You should get into politics.