But there must be some inverting going on to get the negated OUTxB. Maybe it is not a fully ballanced design from input to output in the MA?
Se that if one should choose to keep c62 and c60 for balanced input, R16 has to go. It connects c60 to ground as it is only 10ohm
Is the chip really symmetrical from input to output itself? I wonder since it can be driven both balanced and unbalanced.
That is if in0A to out0A don't have any shared circutry with in0B to out0B
Last edited:
When you drive only half of the bridge, the other half is just acting as a current source/sink biased at half supply voltage. There will be no voltage gain in that half. You end up with twice the noise and half the output.
Someone can correct me if I'm wrong. I usually am. 😆
Someone can correct me if I'm wrong. I usually am. 😆
All I know it needs both caps and MKS 2.5 mm do a good job there. If one of the caps could be omitted they would not be there. The "tricks" one sees are for instance L7. Planned to be either a coil or a ferrite bead but it is a 0 Ohm resistor in reality. Many devices with MA12070 suffer from cost cutting with a sales price of below 30 Euro. Except some design imperfections costing a few resistors too many one rarely sees abundance of parts. In cheap devices one practically never sees adequate output filtering.
Hindsight is OK but as the ICs are discontinued one better finds a relatively well designed device with it. The Aiyima A8 is still one of the better ones.
Hindsight is OK but as the ICs are discontinued one better finds a relatively well designed device with it. The Aiyima A8 is still one of the better ones.
Last edited:
For BTL the datasheet show a case where only inxA is driven. As it is no output DC cap on output both bridges must be driven somehow. So there must be some balancing or inverting inside the MA
The board I'm looking at has fine components. Not high end, but not so bad that the chip spec is compromised by the passive components. Often it can be seen at LF and high input voltage if there has been too much saving going on. An example is a cheap multichannel DAC board like this. The distortion raises under 100 Hz. Then it is better to build your own output section as the chip and layout is good enogh.
One interresting point in my opinion is the balancing and lower end caps. My measurements on the chaep DAC board showed that by removing the bad caps and balancing two output, second harmonics and LF noise is greatly reduced. Wonder if the differential input of the MA has the same effect or if the common mode rejection is 0 on the input stage of the MA?
Wow. I tend to agree that driving this chip i differential mode is a good thing 🤓
Think I have to try to connect the multidac chip without output filter. One channel to in0A and one inverted chanel to in0B. Just DC blocking and maybe some propperly sized resistors and a HF cap. The MA datasheet says HF filtering on input. Maybe it is good enough so DAC HF filtering is not needed?
Think I have to try to connect the multidac chip without output filter. One channel to in0A and one inverted chanel to in0B. Just DC blocking and maybe some propperly sized resistors and a HF cap. The MA datasheet says HF filtering on input. Maybe it is good enough so DAC HF filtering is not needed?
Last edited:
I've tried to connect differential outputs from AK4493S directly (without external filters on op amps) but with simply RC filter to diff inputs of MA12070 with DC blocking caps, it works without oscillations or any problems
Both halves of the bridge are biased to VCC/2. It has nothing to do with how the inputs are driven.
OK, maybe I misunderstood.
I was thinking of the way 3 phase motors are driven. One bridge for each phase. And one input for each phase
This is 2 phase. But it seems like the inputs are not direcly connected to the phase.
We have working amps where there are 1 input and both phases is driven. one in phase with input. The other output 180 degrees out of phase with the input
That implies that if only one halv of the bridge is driven there will be a DC with value Vcc/2
I was thinking of the way 3 phase motors are driven. One bridge for each phase. And one input for each phase
This is 2 phase. But it seems like the inputs are not direcly connected to the phase.
We have working amps where there are 1 input and both phases is driven. one in phase with input. The other output 180 degrees out of phase with the input
That implies that if only one halv of the bridge is driven there will be a DC with value Vcc/2
OK, maybe I see what you mean, looking at the drawing above. Is the case that there is only signal on left side and not on right side. Then same DC is on both side of speaker so the speaker see no DC. (PVDD on left = PVDD on right)
Then this could be checked by measuring if there are signal on both sides. One 180degrees out of phase of the other and same magnitude
That's your opinion....
95 euros ???Single chip??? and asymmetrical.... mods on a board A 95 euros .....With an ugly case... No thanks.
The potential cost and improvement is to be thought about .....🙄
@theAnonymous1 Made a measurement with 1k input. scope on output. 4 order 20k software filter on channels also.
Beware of the aliasing. Only because of digital scope and do not exist in reality on amp output
No load, only scope
PBTL
single ended input. Signal to in0A. In0Bgrounded via DC blocking cap and 10ohm resistor
19V PVDD
Green: OUT0A - ground
Blue: OUT0B-ground
+ filtered out
Both bridges has signal😅
Beware of the aliasing. Only because of digital scope and do not exist in reality on amp output
No load, only scope
PBTL
single ended input. Signal to in0A. In0Bgrounded via DC blocking cap and 10ohm resistor
19V PVDD
Green: OUT0A - ground
Blue: OUT0B-ground
+ filtered out
Both bridges has signal😅
Last edited:
50 Euro, good/moddable output filter as an exception, input stage can easily be changed to simple line inputs and they are active to both input pins AFAIK. I had a few but this one performed quite good. A10a also was nice.
What do 2 chips bring?
After a change to a balanced DAC and balanced MA5332MS I would not bother anymore but the second revision A8 was worth the money.
Last edited:
The MA chip in single can be a little light on speakers that are difficult to drive... And 2 is better... (we are made that way). 😂
On speakers with low efficiency or chaotic impedance curve, and it would seem that we gain a few Watts before the 10% distortion. To be confirmed, I have not made any measurements... And yes often in normal listening 5 good Watts is enough
To choose between a car of 70 HP AND 140.... I prefer to take a safety margin.