Chasing SINAD: 33 ohm at 5mW

[boXem]

All data are taken from Amir's potentiometer left right unbalanced graph. He will slowly rotate from max to min and capture using AP.

For low impedance HP, the lower level of the pot is where we normally listen to. So when Amir is turning the knob, how much time does it take to travel from -20dB to -40dB. If the time is more than 5 seconds, it means I have better resolution at the low level. If it's less than 5, IEM users better don't get this.

In summary, if @amirm is nervous, test is failed, if he is sleeping or the volume control is irritatingly slow, test is passed.
The HPA-4 has steps of 0.5 dB. It is accurate. And it seems fast too.

Solutions:
- calibrate @amirm
- buy @amirm a motor that will actuate the volume pot at a fixed speed
- find curve volume versus pot angle

 

[wasnotwasnotwas]

I used Final Audio E3000 IEMs with the Atom without problems.

What kind of IEMs do you have?

These were some obscure (to me at least) German brand, horluchs. 115 dB sensitivity I think

 

[pozz]

These were some obscure (to me at least) German brand, horluchs. 115 dB sensitivity I think

Mine are advertised at 100dB. I couldn't find a measurement confirming that, but still significantly lower than yours.

 

[wasnotwasnotwas]

Mine are advertised at 100dB. I couldn't find a measurement confirming that, but still significantly lower than yours.

Im going to do a test when I get chance- plug the atom into a whole new socket away from anything and see how high I can get it with those IEM before noise becomes a problem- with nothing connected. Then with a DAC connected.

These in fact- 20 ohm and 120dB/mw (I paid nowhere near that price btw- amazon warehouse deal) https://hoerluchs-unlimited.com/en/product/hl4400/

 

[jasonhanjk]

In summary, if @amirm is nervous, test is failed, if he is sleeping or the volume control is irritatingly slow, test is passed.
The HPA-4 has steps of 0.5 dB. It is accurate. And it seems fast too.

Solutions:
- calibrate @amirm
- buy @amirm a motor that will actuate the volume pot at a fixed speed
- find curve volume versus pot angle

I've relook into HPA4 and I agree it should be back in the rank. Will update again when there's another amp review.
The best way is to have a volume vs pot angle, which means data have to re-collected.
Actually can be easily done by user with a multimeter with dBV function.

As of current data, that's the best I have.

 

[boXem]

As of current data, that's the best I have.

Yep. Isn't it better to say that we have nothing on the subject?
Not saying it's not important, just that it is not measured.

 

[jasonhanjk]

Interestingly, I got a second hand Atom, it does exactly as expected. Except for when I plug in sensitive IEMs- I hear the noise audibly. Not so on the L30 in lowest gain.

So I am not sure how to square this "5 second" metric with my real world experience . Unless the noise comes from elsewhere- but I doubt it- SMSL SU8 DAC

Have you tried 1X gain, unplug all the input, volume pot to min, not touching the amp, put a box between the amp and the table, etc?

I get that the time taken to get from -20 to -40 indicates how much "room" you have on the pot - but I dont get how I have more resolution at , say -30 (as a fixed point on any knob) on amp A vs amp B just because it takes longer to get there. But I'm not great at understanding such things

First there must be an assumption; that the tester will rotate the pot at the rate of degrees over time remains same for all the amp.
Now the test between amp A and amp B, both of their pot are 270 degrees from min to max.
The tester will move 27 deg every 2 seconds, after 20 seconds the test would be complete.

Amp A uses 4 seconds to travel from -20dB to -40dB.
Amp B uses 8 seconds instead.
For that 20dB range. Amp A have 54 deg of resolution, whereas Amp B have 108 deg.

 

[jasonhanjk]

Yep. Isn't it better to say that we have nothing on the subject?
Not saying it's not important, just that it is not measured.

Not really. There's something there.
Why some amps are able to exceed 5 seconds while others don't.

I'll let the amp manufacturer go back and think. Those who knows the answer please don't answer it here.

 

[wasnotwasnotwas]

Have you tried 1X gain, unplug all the input, volume pot to min, not touching the amp, put a box between the amp and the table, etc?




First there must be an assumption; that the tester will rotate the pot at the rate of degrees over time remains same for all the amp.
Now the test between amp A and amp B, both of their pot are 270 degrees from min to max.
The tester will move 27 deg every 2 seconds, after 20 seconds the test would be complete.

Amp A uses 4 seconds to travel from -20dB to -40dB.
Amp B uses 8 seconds instead.
For that 20dB range. Amp A have 54 deg of resolution, whereas Amp B have 108 deg.

I get the amount of turn- I just fail to see how that equates to one amp being better for sensitive IEM at a given output level. But I shall stick with the thread and try and glean it.

Re my atom experience, agreed https://www.audiosciencereview.com/...chasing-sinad-33-ohm-at-5mw.17527/post-570342

 

[boXem]

Not really. There's something there.
Why some amps are able to exceed 5 seconds while others don't.

I'll let the amp manufacturer go back and think. Those who knows the answer please don't answer it here.

Sorry, but there is nothing. The fact that one single sample (HPA4) gets the method failing shows that the method is not the right one.
What if the measurement procedure was asking to reach precise points? Then the shorter time the better the accuracy.
The amp manufacturer has 20 years of engineering behind him. The main thing he learnt during these 20 years is humility. But sometime, not often, he doesn't need to go back and think.

 

[pozz]

I'll echo the comments above: duration is not a good measure for ranking linearity. It would be better to list the corresponding attentuation (e.g., -60dB) where the channel imbalance reaches 1dB/0.5dB/0.1dB (take your pick).

 

[wasnotwasnotwas]

Im going to do a test when I get chance- plug the atom into a whole new socket away from anything and see how high I can get it with those IEM before noise becomes a problem- with nothing connected. Then with a DAC connected.

These in fact- 20 ohm and 120dB/mw (I paid nowhere near that price btw- amazon warehouse deal) https://hoerluchs-unlimited.com/en/product/hl4400/

EDIT- plugged it all back in and......silence- need to go to high gain with nothing playing way past 12 o clock to get any semblance of hiss.

Oh well!

 

[wasnotwasnotwas]

After a bit of thought, the table before the pot test was useful. After the pot test, not so much. As long as I dont hit significant imbalance at the lower end of the volume scale, Im pretty unfussed about how much useable turn I have on the pot. Plus I can attenuate digitally before the amp to better set that.

The L30 and I suspect quite a few others should definitely still be "in".

 

[Balle Clorin]

They are all good, unless you can hear a mouse fart next to a drag racer at 100yards

 

[jasonhanjk]

Sorry, but there is nothing. The fact that one single sample (HPA4) gets the method failing shows that the method is not the right one.
What if the measurement procedure was asking to reach precise points? Then the shorter time the better the accuracy.
The amp manufacturer has 20 years of engineering behind him. The main thing he learnt during these 20 years is humility. But sometime, not often, he doesn't need to go back and think.

I'm not referring to Benchmark. Since it will be back at next update.

 

[jasonhanjk]

After a bit of thought, the table before the pot test was useful. After the pot test, not so much. As long as I dont hit significant imbalance at the lower end of the volume scale, Im pretty unfussed about how much useable turn I have on the pot. Plus I can attenuate digitally before the amp to better set that.

The L30 and I suspect quite a few others should definitely still be "in".

It would be in as long as there's data to shows it's good.

Here's a data from a "branded" pot. If it's as good as this (or close to) then it's in.

 

[wasnotwasnotwas]

I'll echo the comments above: duration is not a good measure for ranking linearity. It would be better to list the corresponding attentuation (e.g., -60dB) where the channel imbalance reaches 1dB/0.5dB/0.1dB (take your pick).

Im not sure the OP is ranking or factoring in linearity at all. He is corresponding the amount of turn on the pot you have between -20 and -40v to "resolution" (his word)- which I dont think it does either. He was better leaving at a table of SINAD at 5mw/ 33ohm . Now combining that with amps that may lose linearity at that level would be useful, like you suggest.

 

[okok]

snr relates to (iem) noisy or not, thd+n not that important

and you misread 5mv and 5mw

sorry