Measurements of Hifidiy forum Anniversary TDA-1541 My DIY Replica

[WolfX-700]

In 2014, the hifidiy.net forum released a series of DIY projects to mark the tenth anniversary of the forum.
This 1541 PCB is based on the 1541 DAC project.

Basically the TDA1315H input, SAA7220+TDA1541 architecture supports RCA and XLR outputs.

Some of the above PCB replicas were made last year. And I made this DAC based on this PCB.

In production, I use JRC5534DD for I/V & OPA1611for LPF.

 

[solderdude]

Looks nice.

Some constructive comments:

The Linearity test from Amir goes down to -120dB, yours to -80, perhaps use the same scales in order to get comparative plots.

The phase plot appears to have a considerable (audible) phase shift until one looks at the scale.. 0.4 degrees is nothing. Perhaps use a +/- 90 degrees scale ?

100dB distance in multitone is quite good for an ancient partly R2R design.

 

[WolfX-700]

Looks nice.

Some constructive comments:

The Linearity test from Amir goes down to -120dB, yours to -80, perhaps use the same scales in order to get comparative plots.

The phase plot appears to have a considerable (audible) phase shift until one looks at the scale.. 0.4 degrees is nothing. Perhaps use a +/- 90 degrees scale ?

100dB distance in multitone is quite good for an ancient partly R2R design.

Thanks for your comments

The 1541 is a 16-bit DAC that is impossible to process - below 96dbfs (-144for24bit) so I used -80.

I am trying to highlight the change in phase. If you use +/- 90 degrees scale , then we can see nothing.

 

[solderdude]

The 1541 is a 16-bit DAC that is impossible to process - below 96dbfs (-144for24bit) so I used -80.

True but why not go down to -100dB

I am trying to highlight the change in phase. If you use +/- 90 degrees scale , then we can see nothing.

That is also true but on the other hand 0.4 degrees phase shift is nothing consequential either so seeing a flat line would be appropriate.
It is always the fine balance between seeing what can be measured vs showing the relevance of the measurement.

 

[WolfX-700]

True but why not go down to -100dB



That is also true but on the other hand 0.4 degrees phase shift is nothing consequential either so seeing a flat line would be appropriate.
It is always the fine balance between seeing what can be measured vs showing the relevance of the measurement.

My decision is to try to use linearity starting at -96dbfs on the 16bitDAC next time.

I also thought about the +/- 90 degrees scale. It is indeed a more "standardized" choice. I will use it later.

Thankyou

 

[Rja4000]

Hi
Well done !
To be more useful, it would be great to describe shortly the device (in text) as a header.
Otherwise, we have to guess.

 

[bunkbail]

Looks great for an ancient DAC chip. Can we see the innards of the DAC?

 

[WolfX-700]

Hi
Well done !
To be more useful, it would be great to describe shortly the device (in text) as a header.
Otherwise, we have to guess.

Thankyou. It's done.

 

[WolfX-700]

Looks great for an ancient DAC chip. Can we see the innards of the DAC?

 

[Veri]

Pretty good for a very old 16-bit DAC. Multitone is not bad at all I own an older Audial DIY Dac with same chip and it sounds just, good.

The 10th anniversary design also shows up on aliexpress, on eBay even completely pre-built. Is this the same one you used?
https://aliexpress.com/item/32978376662.html
https://ebay.com/itm/323820031544

 

[solderdude]

Fortunately the DAC is properly driven by SAA7220 so has proper filtering unlike some of the NOS designs.

 

[WolfX-700]

Pretty good for a very old 16-bit DAC. Multitone is not bad at all I own an older Audial DIY Dac with same chip and it sounds just, good.

The 10th anniversary design also shows up on aliexpress, on eBay even completely pre-built. Is this the same one you used?
https://aliexpress.com/item/32978376662.html
https://ebay.com/itm/323820031544

Yep I got PCB from Taobao

 

[WolfX-700]

 

[maty]

I agree, 0.4º is almost nothing.

The problem, with amplifiers, is when the phase shift at 20 kHz is > 6º. With class A and class AB is not a problem, that is why there are hardly any measurements available on the net.

1º at 20 kHz is a meritorious result in amplification.

 

[alec226]

RCA output 5Vrms? Would that be a little too high?

Do you have a Ming9002 1541 OS3? Would like to see the measurements.

 

[Veri]

RCA output 5Vrms? Would that be a little too high?

It's due to the voltage output of TDA1541A. A proper implementation should have high V output.

 

[alec226]

It's due to the voltage output of TDA1541A. A proper implementation should have high V output.

Got it, thanks. Is there an ideal output level for 1541A?

 

[Veri]

Got it, thanks. Is there an ideal output level for 1541A?

Good question, full output will give highest SNR if your upstream gear works with hot input. 4V/5V.
I would avoid lowering digital volume considering you're doing the volume control in 16-bit, not that accurate.

 

[veeceem]

@WolfX-700 since you've been actively measuring stuffs and post here. Why don't you just make a thread with links to all gears that you measured? A ranking graphics would also be nice. Make it easier to follow. Thanks for your works, btw

 

[Veri]

@WolfX-700 since you've been actively measuring stuffs and post here. Why don't you just make a thread with links to all gears that you measured? A ranking graphics would also be nice. Make it easier to follow. Thanks for your works, btw

There's one on his website Check it out:
https://l7audiolab.com/dac-sinad-charts