Review and Measurements of New JDS Labs Atom Headphone Amp

[Yviena]

It will not have issues with the HD800. I have used them as a pair extensively, its one of those few amps that pulls out the low end grunt of the HD800 while showcasing the headphone's expansive stage imaging with clarity.

On low-gain, the Atom's dial doesn't go beyond the 2 o'clock position. Usually its just about 1 o'clock only and I'm at the limit of what is comfortable for an extended listening session. That was mostly with my Khadas Tone Board at 100% digital volume. Sometimes its connected to my ES9038Pro based desktop amp at -5 dB output setting.

Hmm i see but i read somewhere that the HD800 has a impedance spike over 600 ohms under 100hz, could this be the reason i find bass to be a little bit more punchier in high gain even if i volume matched via player?

 

[maxxevv]

Are you referring to the Atom here or some other amp ?

The impedance spike is there indeed as documented by sites such as InnerFidelity. But I don't hear the issue with the Atom.

I have compared the HD800 via the build-in amp of the Anedio D2 and then separately output to the Atom from its DAC, the bass response on the Anedio's amp was a fair bit less compared to the Atom. It was a similar experience with my HPA-3B which sounds very, very similar to the Atom without level matching. So, on some amps such as the Anedio's build-in amp, the HD800 can sound somewhat "thin" comparatively. But on the Atom, it certainly doesn't that way for me.

 

[Yviena]

Are you referring to the Atom here or some other amp ?

The impedance spike is there indeed as documented by sites such as InnerFidelity. But I don't hear the issue with the Atom.

I have compared the HD800 via the build-in amp of the Anedio D2 and then separately output to the Atom from its DAC, the bass response on the Anedio's amp was a fair bit less compared to the Atom. It was a similar experience with my HPA-3B which sounds very, very similar to the Atom without level matching. So, on some amps such as the Anedio's build-in amp, the HD800 can sound somewhat "thin" comparatively. But on the Atom, it certainly doesn't that way for me.

I was refering to the ATOM, but i think i may have found why bass sounds a bit punchier, in high gain my DAC outputs 2.25vrms on RCA, i see that on high gain the ATOM can only accept up to 2.10vrms input before it starts distorting.

 

[maxxevv]

Ok. But still doesn't explain what you experienced in terms of bass response from the Atom (low gain) and the HD800 though.

I haven't had the need to run the Atom on high gain so far, so can't comment on that difference unfortunately.

 

[JohnYang1997]

Hmm i see but i read somewhere that the HD800 has a impedance spike over 600 ohms under 100hz, could this be the reason i find bass to be a little bit more punchier in high gain even if i volume matched via player?

Gain is set at the input section/gain section of atom. Which is a separate part from the output stage. There is no correlation in gain and power if the input signal level is ample. In terms of perception of increased bass, probably the volume isn't matched exactly.
There is no problem with the impedance spike. Neither the impedance spike affect atom which has minimum output impedance not the gain setting changes the output impedance of atom.
One thing that's remotely possible (hypothetical) is that when volume matched, the volume knob is at a lower position where the resistance of the it in the signal path is smaller which then cause less noise in the final output.
Another thing that's also hypothetical (which is a bit more evident). As I tested jds atom, I see power supply noise in the low frequency at 50hz/or 60hz in America. The power supply section of atom isn't beafy enough of clean out the mains noise. So as a result the main noise bleeds to the opamp output (gain stage). Increased gain resulting in lower psrr(power supply rejection ratio). So higher mains noise go through output stage and appear at output then your hd800.
One more possibility is the mains noise bleeds to the input via pcb design. The input traces running around will couple mains noise into the signal path. And it supports the last one.
How I managed to correct all these.
Use 1Kohm potentiometer instead of the 10k one. Decreased thermal noise at any potentiometer position. And it decreased the input impedance of the output stage.
Use 1kohm pull down resistor at input stage. As mains bleed through pcb is weak in current. A lower input impedance will eliminate the effect. P=UI. P being constant. Increased I will give lower noise voltage.
Use 4700uF power supply filtering capacitors. It will eliminate the mains noise resulting clean dc.

 

[JohnYang1997]

To think of these details require sophisticated experience and knowledge in analog circuits. Jds atom did a great job making a low distortion decent engineered amplifier. But it's not perfect.

 

[JohnYang1997]

Also you can use some other opamp with higher psrr for the second one. But it's better to prevent that from the happening.
To test if it's true, you can hook up a high sensitivity iem. Unplug the inputs. And listen if there is hum.
And do you hear white noise. Does the white noise change when you turning the volume. Maximum noise is at 50% position due resistance. And it should be at around 2-3 o'clock.
If the main hum is increased while turning up it supports my 2 and 3 guesses.

 

[maxxevv]

@solderdude shared with me this plot of the frequency response curve of the HD800 at different SPL settings.

I believe what you encountered with the higher gain has a lot to do with a higher SPL output compared to the low gain setting. as they are not level matched and you didn't realise it.

https://www.audiosciencereview.com/...headphone-measurement.7919/page-2#post-192005

 

[JohnYang1997]

To extend a bit more in the topic. Many people love class A or some linear power supply or discrete designed amplifier. They are all suspects to mains noise. Mains noise may enhance the perception of bass response (more body). And to think it's better is logically flawed. I went into this before when I didn't know all this.

 

[amirm]

I just ordered this based on this review and measurements data. Thanks for saving me from more expensive mistakes [I already made some in the past.] Currently I use my trusty O2,[massdrop] but for $99 I just couldn't pass this up. Thanks Amir! BTW I remember you from AVS forum when you were a MS "insider" there.

Wow, that takes me way back! Good to see you here.

 

[Yviena]

@solderdude shared with me this plot of the frequency response curve of the HD800 at different SPL settings.

I believe what you encountered with the higher gain has a lot to do with a higher SPL output compared to the low gain setting. as they are not level matched and you didn't realise it.

https://www.audiosciencereview.com/...headphone-measurement.7919/page-2#post-192005

I'm pretty sure i volume matched it to around 0.2-0.5db using the foobar volume control, as i could not tell a difference in volume but bass still somehow had more body, but i think the perception of more bass was due to more distortion.

 

[JohnYang1997]

I was refering to the ATOM, but i think i may have found why bass sounds a bit punchier, in high gain my DAC outputs 2.25vrms on RCA, i see that on high gain the ATOM can only accept up to 2.10vrms input before it starts distorting.

That's not an issue. JDS ATOM runs at +-15V (iirc). 2.25rms won't cause any distortion. NJM2068 can accept pretty high voltage before distorting(tho can have higher distortion at high gain but still). Punchier is directly related to being louder. The volume might not be matched exactly.

 

[JohnYang1997]

I'm pretty sure i volume matched it to around 0.2-0.5db using the foobar volume control, as i could not tell a difference in volume but bass still somehow had more body, but i think the perception of more bass was due to more distortion.

Distortion will cause brittle, thin, bright, grainy, stuffed sound depending on distortion frequency and distortion characteristics(phase, harmonic structure). More body is most likely the mains noise.

 

[Yviena]

Distortion will cause brittle, thin, bright, grainy, stuffed sound depending on distortion frequency and distortion characteristics(phase, harmonic structure). More body is most likely the mains noise.

Yeah i think it's maybe the mains noise then if EU has the same problems, I'm located in norway, i believe the mains are some kind of balanced type here though so shouldn't that help with the noise?

So replacing the AC caps with 4700uf total or each; should help then?

 

[maxxevv]

I'm pretty sure i volume matched it to around 0.2-0.5db using the foobar volume control, as i could not tell a difference in volume but bass still somehow had more body, but i think the perception of more bass was due to more distortion.

Its the Atom's low-gain versus its high-gain. Not the foobar part.

The funny thing about the Atom and any good linear low distortion amps paired with very "scalable" headphones such as the HD650 and the HD800 is that they don't sound like they are louder even when you crank up the volume significantly. Its something I had a good laugh/ scare over when I got my HPA-3B and then some on the Atom.

I usually listen to the headphones without looking at the volume dial and will adjust by feel, blind. There were a few times when the music was so good I had turned the volume up 2 notches higher than my normal listening volume without realising and I had been listening for a good 30~40 minutes on those settings when I looked at the dial !

 

[JohnYang1997]

Yeah i think it's maybe the mains noise then if EU has the same problems, I'm located in norway, i believe the mains are some kind of balanced type here though so shouldn't that help with the noise?

No. That's not relevant to ripple in dc and bleed in traces.
1, atom has 2 * 470uF per rail which is weak for ps filtering.
2, the input traces runs around the whole board due to two inputs and selection between the two.
These two potential issues in the design caused higher mains noise.

 

[JohnYang1997]

Its the Atom's low-gain versus its high-gain. Not the foobar part.

The funny thing about the Atom and any good linear low distortion amps paired with very "scalable" headphones such as the HD650 and the HD800 is that they don't sound like they are louder even when you crank up the volume significantly. Its something I had a good laugh/ scare over when I got my HPA-3B and then some on the Atom.

I usually listen to the headphones without looking at the volume dial and will adjust by feel, blind. There were a few times when the music was so good I had turned the volume up 2 notches higher than my normal listening volume without realising and I had been listening for a good 30~40 minutes on those settings when I looked at the dial !

I think he meant he used foobar to compensate for the gain in atom.
In this case the noise in the dac is amplified. It could also be the noise in the dac causing more body in the lows. That happens too.
It's better to use potentiometer to match volume.
One more thing that could be the case, reduced digital level will likely cause lower distortion in the dac. Most dacs has less distortion around -6db to -20db. Tho it trades off to the noise. It might case difference. Lower distortion = less bright. But it's unlikely tho distortion at most dac are already low

 

[Yviena]

I think he meant he used foobar to compensate for the gain in atom.
In this case the noise in the dac is amplified. It could also be the noise in the dac causing more body in the lows. That happens too.
It's better to use potentiometer to match volume.
One more thing that could be the case, reduced digital level will likely cause lower distortion in the dac. Most dacs has less distortion around -6db to -20db. Tho it trades off to the noise. It might case difference. Lower distortion = less bright. But it's unlikely tho distortion at most dac are already low

Yeah i compensated the gain in foobar, unsure if it can be the dac though, It's a Matrix sabre pro non-mqa so should be below the threshold of audibility.

When you did the mod on your atom what size did you use for the caps, 4700 each or total depending if they fit in the chassis, any recommended caps/resistors to use?

 

[maxxevv]

I think he meant he used foobar to compensate for the gain in atom.
In this case the noise in the dac is amplified. It could also be the noise in the dac causing more body in the lows. That happens too.

You technically can't do that unless you know the exact digital scaling ratio to analogue output levels of the DAC used. They are not linear both ways.

 

[solderdude]

Do you hear a low level hum on high gain ?
If not it isn't the power supply.

I see no reason why it would have more or 'tighter' bass in both settings.
The output stage is exactly the same in both cases, only the input gain differs.
This does not change tonal balance. It changes distortion levels (inaudibly low at both settings) and noise levels.

When you don't hear noise (unlikely as the HD800 isn't the most voltage sensitive headphone) nor hum in the high gain setting that cannot be it.

In my mind it could be 'very mild clipping' being on the verge of becoming audible due to the input opamp clipping.
This can be avoided by lowering the volume by 1 or 2 dB in the PC/laptop so can be avoided easily.
or there is some kind of loudness engaged in the player.
At high gain I don't suppose you are in channel imbalance territory yet.

This needs to be tested blind (as in not knowing what is playing) with someone else lowering the volume of the PC to avoid false positives.

Increasing the smoothing caps isn't a bad idea anyway but reckon JDS tested the ripple (reduction) of the supplied transformer at full load on both channels. The HD800 load is very low so I don't expect the voltage on the smoothing caps so sag much nor that the ripple becomes alarmingly high.