Tear down of Massdrop THX AAA 789 Headphone Amp

[KunkyAudio7861]

Hey guys! I am hoping to find some help online to instruct me on how I can fix my THX AAA 789 myself. I'm very new to all of this, but really hoping to take a crack at this. Long story short, I plugged in a faulty XLR to my THX and now when I power it up, it stays on the red indication light. I think I shorted something?
Can those of you with experience help me? Is it possible to diagnose the issue with a multi-meter?
I've taken the board out of the enclosure, I don't see in visible damage. Any help is much appreciated! Thank you!

 

[amirm]

Even someone experienced likely can't fix it given the surface mount devices and lack of schematic. I recommend sending it for repair/replacement given its cost and great performance.

 

[trl]

@KunkyAudio7861, is there a reason why not sending the unit back for repair?

I'm sure that between the output stage and the XLR plug there are few-hundred-Ohms resistors to protect the output opamps from faulty cables. Are you sure you haven't injected AC or DC voltage from an external source (DAC perhaps) into the XLR-output plugs from the THX headamp?

What kind of fault were the XLR cables had, please? Maybe shorted to outer ground of the plug (outer shield)?

Thank you!

 

[amirm]

I am assuming by XLR cable he means the balanced headphone jack....

 

[trl]

Oh...

So something around TI L239 (https://www.audiosciencereview.com/...-aaa-789-teardown-main-pcb-reduced-jpg.17796/)?

I do hope this nice headamp can be returned for servicing.

 

[solderdude]

Looks like the noise levels of the OPA564 aren't a real life problem.
When looking at the measurements I don't see any alarming noise levels.

 

[JohnYang1997]

Looks like the noise levels of the OPA564 aren't a real life problem.
When looking at the measurements I don't see any alarming noise levels.

Because it bas extra error correction. If not the performance won't pass 100db.

 

[JohnYang1997]

100nV @ 1kHz is high in my opinion. (and the test is usually done at unity gain)
A good opamp is less than 2nV @ 1khz.
The LME49600 buffer is 3nV, for example.
Two of the LME49600 in bridge mode will probably drive just about any headphone out there. I made a headphone amp that way once.
So, it seems to me that they are playing games with specs, since I doubt anyone needs 6W to drive headphones.
I simulated their patented circuit in Spice, and what they are doing is separating the DC feedback from the AC, as well as providing a feedback loop around the buffer itself. I don't see any magic in this device. The fact that the noise from it is "not a problem in real life" probably says a lot more about human perception and tolerance for noise than anything else. Electronics are so good that they can still be just OK and perform well.

I don't think this one is composite amplifier. It's known as feedforward error correction. (but it might be marketing bullshit).
Feedback essentially changes the input signal to hold the output to the original input tightly. Feedforward should be feeding forward after the output with two signal path independent to each other. Resulting more effective error correction and higher stability.

 

[JohnYang1997]

As I understand it, error correction doesn't have an effect on noise.

The OPA564 doesn't have access pins for applying the feedback shown in the patent, so it has to be a composite of the OPA1602 and OPA564. Otherwise, they are not using the patent in this amplifier. I looked at the patent drawings and, while I am not an expert on feedforward, all I see are feedback loops.

Yes it does effect on noise of course. It will be reduced by the amount of feedback.
As parent go, I was too confused about what these really are. Also the performance shown on massdrop was never replicated 3rd party. Who knows what happened. Cost down for mass production? Or just complete bullshit. Idk.

 

[solderdude]

100nV @ 1kHz is high in my opinion. (and the test is usually done at unity gain)
A good opamp is less than 2nV @ 1khz.
The LME49600 buffer is 3nV, for example.
Two of the LME49600 in bridge mode will probably drive just about any headphone out there. I made a headphone amp that way once.
So, it seems to me that they are playing games with specs, since I doubt anyone needs 6W to drive headphones.
I simulated their patented circuit in Spice, and what they are doing is separating the DC feedback from the AC, as well as providing a feedback loop around the buffer itself. I don't see any magic in this device. The fact that the noise from it is "not a problem in real life" probably says a lot more about human perception and tolerance for noise than anything else. Electronics are so good that they can still be just OK and perform well.

So the amps spec: Noise (A-wt): 1.2 uV, potentiometer at nil... is about human tolerance for noise ?
Amir's measurements that show a low noise floor (and low distortion) is higher than the measurements show ?

Let's say I use a really sensitive headphone of say 110dB/V
1.2uV = -118dB so this high sensitive headphone produces -8dBA and you feel this is not a problem because of human perception and tolerance for noise and not because it is way below the audible threshold even with a very sensitive headphone ?
The specifications of SNR: 136 dB, 300 ohms, <1% THD (=200mW = 7.7V) are just something that's made up and they are playing games ?

I do agree that there is no magic in this device... there isn't any magic in whatever device unless magic consists of electrons.

 

[Music1969]

Also the performance shown on massdrop was never replicated 3rd party. Who knows what happened. Cost down for mass production? Or just complete bullshit. Idk.

But didn't @amirm's measurements show the THX 789 amp measures state of the art? High power output at ultra low distortions and very low output impedance?

 

[JohnYang1997]

But didn't @amirm's show the THX 789 amp measurements are state of the art ?

Sure it's great but not the same as on massdrop. And the neurochrome hp-1 has far lower distortion harmonics than this. Well, we aren't talking about audibility but pure electrical performance.

 

[JohnYang1997]

Yes, the noise of the entire circuit is affected by the gain thereof, which is function of feedback. The noise of an opamp circuit is proportional to the inherent unity gain noise quoted in the data sheet times the gain of the circuit. The noise is not independently controlled by feedback, is what I was trying to say.

No the composite amplifier has more feedback from the gain of first stage. 1602 has very low voltage noise and very high gain thus the entire circuit will be limited by resistor noise mainly.

 

[JohnYang1997]

The thermal noise of an opamp circuit is a function of several characteristics, including the absolute values of the feedback resistors, the noise gain of the circuit, the source impedance, and the current and voltage input noise for the opamp. The noise gain is the same as the signal gain for non-inverting amplifiers, but the noise is increased further by the absolute values of the resistors times the current input noise. So, without knowing what they used for these values, it's hard to see which noise source is dominant. The 1602 has high current input noise, so they would have had to use low value feedback resistors to reduce noise, which they can do since the power amps can drive a low impedance feedback loop(s). They would also need low source impedance to get maximum performance out of the 1602.

They only have around 1.2uV at output stage (volume nil). So idk why you are always suspecting things that don't comply with reality.

 

[JohnYang1997]

Since it was mentioned, is it possible replace the current potentiometer with an Alps blue velvet?


QUOTE="Jimmy, post: 148369, member: 2373"]I think that if this is just a licensed THX's design it would be beneficial to distribute it under another brand, maybe adding an internal connector for a DAC board like f.e. the Moon 430, Jotunheim, ..., and maybe include a better potentiometer too (Alps blue velvet, Vishay pro, ...), selling it as they are doing now is seriously limiting its potential user base.

[/QUOTE]
Don't use high resistance pot. They are very very noisy that degrades performance hugely. I would rather replace to a bourn 1kohm pot. All amplifier that uses big pots are shit in my book.

 

[Dirk Wright]

You can't just use a 1k pot because then the input impedance usually becomes only 1k and many source op amps can't drive that. There is a way to do this, but you can't just replace a 10k with a 1k and expect the circuit to perform well.

 

[JohnYang1997]

You can't just use a 1k pot because then the input impedance usually becomes only 1k and many source op amps can't drive that. There is a way to do this, but you can't just replace a 10k with a 1k and expect the circuit to perform well.

Correction: Many opamp can drive that.

 

[trl]

"2-8 INPUT CIRCUIT? The higher the input impedance the more stray noise pickup you get if the amp is connected to an un-terminated cable or gear that is powered off. And much less than 10K can excessively load the outputs of some DACs, preamps, etc. So 10K was chosen as optimal. An RC filter provides RF protection with a cutoff around 3 Mhz. This is low enough to filter out most RF energy while not creating phase shift in the audio band with even higher impedance sources. Cell phones operate at 800+ Mhz and are the most common source of RF problems. A good input circuit should also have some series resistance to help limit current into the op amp if it’s overloaded and provide greater ESD protection. But the larger the series resistor, the worse the noise performance of the amp so it’s a trade-off." - source: http://nwavguy.blogspot.com/2011/07/o2-design-process.html.

http://www.ti.com/lit/ds/symlink/opa1612.pdf - you can see that on 600 Ohms this opamp struggles on the THD+N graph, while on 2 KOhms output source impedance it performs better on both unity gain and 10X gain too.

While http://www.ti.com/lit/ds/symlink/lme49720.pdf seems to drive much better 600 Ohms source.

Personally, I think that 1 KOhm potentiometer might not be considered the best "upgrade" for a 10 KOhms potentiometer in the input stage, unless we can actually measure linearity, THD+N and output voltage of the connected DAC/source while connected to the amp )before vs. after the "upgrade"). Of course, heat on the output opamps could be also measured before vs. after.

 

[JohnYang1997]

"2-8 INPUT CIRCUIT? The higher the input impedance the more stray noise pickup you get if the amp is connected to an un-terminated cable or gear that is powered off. And much less than 10K can excessively load the outputs of some DACs, preamps, etc. So 10K was chosen as optimal. An RC filter provides RF protection with a cutoff around 3 Mhz. This is low enough to filter out most RF energy while not creating phase shift in the audio band with even higher impedance sources. Cell phones operate at 800+ Mhz and are the most common source of RF problems. A good input circuit should also have some series resistance to help limit current into the op amp if it’s overloaded and provide greater ESD protection. But the larger the series resistor, the worse the noise performance of the amp so it’s a trade-off." - source: http://nwavguy.blogspot.com/2011/07/o2-design-process.html.

http://www.ti.com/lit/ds/symlink/opa1612.pdf - you can see that on 600 Ohms this opamp struggles on the THD+N graph, while on 2 KOhms output source impedance it performs better on both unity gain and 10X gain too.

While http://www.ti.com/lit/ds/symlink/lme49720.pdf seems to drive much better 600 Ohms source.

Personally, I think that 1 KOhm potentiometer might not be considered the best "upgrade" for a 10 KOhms potentiometer in the input stage, unless we can actually measure linearity, THD+N and output voltage of the connected DAC/source while connected to the amp )before vs. after the "upgrade"). Of course, heat on the output opamps could be also measured before vs. after.

0.00004% at 10khz isn't struggling, is it? Plus 1KOhm is much higher than 600Ohm. As long as it doesn't pull the performance of the entire system down. It's worth it.

Btw the original question is to change to a 100k alps pot, which is way more noisy the 10k pot. 10k is a middle ground for me. 1k is a bit aggressive but still acceptable. And it's less noisy. Maximum of 250 ohm vs 2.5k ohm equivalent noise resistance.

 

[Dirk Wright]

You can gain up to10dB of s/n, depending on the input opamp (FET vs. bipolar) by going to a lower value pot for the volume control (10k to 1k in this case). However, as has been stated, 1k ohm is generally considered too low by the industry as a whole for an input impedance. If you use a buffer with high input impedance to drive the 1k pot, then it works great. So, no, I don't think it's a good idea to just swap the pot for a much lower value. Some kind of low noise, low distortion buffer has to be added to drive the low Z pot. Now the issue becomes one of designing an awesome buffer that can handle the full output from a DAC or CD player, which is another story.