Sabaj A20h Balanced Headphone Amp Review

[Roland68]

Glad to see the performance has been confirmed in the field again.
The reality is that this device has a very low price; so likely some corner had to be cut to hit this price target.

The biggest compromise is likely the external power supply, as it's purchased separately. The prices for larger quantities are so low that most experts cannot even imagine how it can be produced for this.

 

[Jimster480]

The biggest compromise is likely the external power supply, as it's purchased separately. The prices for larger quantities are so low that most experts cannot even imagine how it can be produced for this.

That is the case with lots of these "simpler" products these days. The prices don't make any sense.

 

[jkim]

I bet that's power supply leakage current. When leaving the amplifier ground floating, this may generate enough common-mode signal to tax the input stage's CMRR. Which is why the two ends of a balanced connection are supposed to be kept at near enough the same potential.

(2) is still worth looking into either way.

I actually was waiting for you to reply---as you are one of the few who may answer my questions concerning this issue!

Right, it is power supply leakage. In fact, ALL DACs/amps I have tested so far, which were connected to the mains directly or indirectly (e.g., through computer power), exhibited power supply leakage to some degree WHEN their ground was NOT connected to the ground of my ADC for testing.

My ultimate question is, what if headphones are connected to these balanced outputs? Headphones are essentially a device with no ground by nature. This means that this power supply leakage, if present, must be fed into headphones. Am I thinking correctly?

 

[AnalogSteph]

My ultimate question is, what if headphones are connected to these balanced outputs? Headphones are essentially a device with no ground by nature. This means that this power supply leakage, if present, must be fed into headphones. Am I thinking correctly?

I have, in fact, encountered a hum when IEMs were plugged into a laptop with a 2-prong mains power supply before. Touch the device and it would go away, as capacitive coupling to your hand would then provide the more attractive path to ground as opposed to taking the circuitous route via your ears first. That was with a traditional 3-pin plug though. This phenomenon should theoretically not be happening with a balanced headphone connection (one of the few times that has a real advantage), as leakage would be appearing as a common-mode signal instead of on the ground returns only. Likewise, full-size headphones tend to exhibit much less capacitive coupling and are a lot less sensitive.

A PC headset to telephone handset converter / switch thing I'm using supplied with an universal (transformer) wall wart has similar issues... place your hand over it and the hum goes down. That has some issues with poor shielding I think, it's also very sensitive to RFI which leads me to suspect that there are some problems with subpar ground layout as well.

 

[jkim]

Picked one up at a very low price to use it with a USB dongle. Quickly measured its 1 kHz sine tone SINAD:

View attachment 437082

This is equivalent to Amir's AP result (SINAD 118 dB) considering mine must have added slight harmonic products from my DAC/ADC setup (E1DA 9039S / Cosmos ADCiso).

However, as some members observed in this thread, this amp seems to be very susceptible to AC and RF noise. Across my multiple measurement trials, I noticed some variability of its noise floor. For example, as Amir mentioned in his Sabaj A10h review, this amp's noise floor also increases quite a bit when its output ground is made "floating." See below:
View attachment 437088

As a result, SINAD worsened by 10 dB. Of course, amps and DACs can suffer from AC/RF noise when their ground is not connected to test gear. But the adverse effect is not to this extent. Wolf at L7audio also observed the same phenomenon in his A10H review.

But then why? Two potential causes: (1) quality of the included power supply; or (2) the circuit GND is not grounded to the case.

I will look into both. Any suggestion?

Replying to my own post, after some experiments, I concluded that the noise-prone character of the amp is due to its design. Change of its power supply (I have some 15V DC adapters laying around) and case grounding does not much change such a characteristic.

I forgot to take a picture of its interior. But if you open it, you will find a number of long, vertical metal pins connect the bottom and top PCBs. They are just like many antennas! Even during rolling averaging in REW, the noise floor is constantly fluctuating. My room is an ideal place for RFI noise tests---it is next to a room with a bunch of routers for the entire building For example, below is a shot when the noise is rising:

Sure, it may still be below audibility. But I am not comfortable. Sometimes, the noise level becomes even greater. Some member mentioned he was hearing noise when his smartphone is near the amp. Now I believe his experience is real.

In comparison, my Topping L70 is absolutely quiet in exactly the same setup.

You may say you have not noticed any problem. I believe it, too, because every environment is different. But I can say this device is very finicky in dealing with RFI, which is built into its design. I guess I will let it go...

 

[AudioFans]

I have, in fact, encountered a hum when IEMs were plugged into a laptop with a 2-prong mains power supply before. Touch the device and it would go away, as capacitive coupling to your hand would then provide the more attractive path to ground as opposed to taking the circuitous route via your ears first. That was with a traditional 3-pin plug though. This phenomenon should theoretically not be happening with a balanced headphone connection (one of the few times that has a real advantage), as leakage would be appearing as a common-mode signal instead of on the ground returns only. Likewise, full-size headphones tend to exhibit much less capacitive coupling and are a lot less sensitive.

A PC headset to telephone handset converter / switch thing I'm using supplied with an universal (transformer) wall wart has similar issues... place your hand over it and the hum goes down. That has some issues with poor shielding I think, it's also very sensitive to RFI which leads me to suspect that there are some problems with subpar ground layout as well.

I find the same problem. When I use the he400se or edxs, it has hmmm noise;while i use hd600 ,this problem disappear. I use the 3 feet power can occer this problem a bit