RME changes DAC-chip

[3125b]

RME provides measurements in the manual: Bedienungsanleitung (rme-audio.de) (from page 57)
They don't differentiate between the two FS versions other than filter settings, so it's pretty safe to assume that ESS vs. AKM just doesn't make a difference.

 

[Duffy Moon]

RME provides measurements in the manual: Bedienungsanleitung (rme-audio.de) (from page 57)
They don't differentiate between the two FS versions other than filter settings, so it's pretty safe to assume that ESS vs. AKM just doesn't make a difference.

Given that there probably isn't a difference other than those filters, I wonder which version forum members would choose and why. Is the brickwall filter better in some way/s?

 

[Skeptischism]

RME provides measurements in the manual: Bedienungsanleitung (rme-audio.de) (from page 57)
They don't differentiate between the two FS versions other than filter settings, so it's pretty safe to assume that ESS vs. AKM just doesn't make a difference.

Well, it will have made a pretty large difference to the designer and the entire layout of the DAC and ADC sections. the ADI-2 used the AK4493. it most certainly makes a difference. the AK4493 is a voltage out DAC chip, so does not require an IV stage, just buffering and filtering. the AK is a much easier chip to get right than the ESS and if there is no difference in the performance, something isnt right.

 

[Willem]

something isnt right

Why? Doesn't it just mean that RME got it right?

 

[Frgirard]

Well, it will have made a pretty large difference to the designer and the entire layout of the DAC and ADC sections. the ADI-2 used the AK4493. it most certainly makes a difference. the AK4493 is a voltage out DAC chip, so does not require an IV stage, just buffering and filtering. the AK is a much easier chip to get right than the ESS and if there is no difference in the performance, something isnt right.

The sound of the dac/adc is not a belief at RME.

 

[Skeptischism]

i'm not talking about the sound... where did I mention the sound?

I forgot, they are using the ES9028Q2M, not one of the PRO series. the q2m dacs are somewhat easier to deal with. They do still require an IV stage. the ESS ADCs hit some higher marks than the AKM as well

Why? Doesn't it just mean that RME got it right?

well, if they were using the 9028PRO, or 9038PRO, they are significantly higher performance than the AKM, so if the performance is the same, theyve screwed up somewhere. I havent used the 9028q2m. only the 9038q2m and the ES9018/28/38PRO.

I just checked the datasheet. the ES9028q2m comes in at -120dB THD+N vs -113dB THD+N. of the AK4493 and DNR is also meaningfully better.

So no, they shouldnt measure the same.

the sound? very likely impossible to pick between them.

 

[Skeptischism]

The sound of the dac/adc is not a belief at RME.

and dude, please dont be such a cliche ...

I would have thought it was pretty obvious I was making comment on the objective data and performance ...

 

[Skeptischism]

also, the ES9028q2m does not officially support 768kHz (like the AKM), max PCM rate is 384kHz. running it without the OSF you may be able to get it to run 768kHz. again, just talking objective specs here, not saying its important to run PCM at 768kHz.

 

[Trell]

Well, it will have made a pretty large difference to the designer and the entire layout of the DAC and ADC sections. the ADI-2 used the AK4493. it most certainly makes a difference. the AK4493 is a voltage out DAC chip, so does not require an IV stage, just buffering and filtering. the AK is a much easier chip to get right than the ESS and if there is no difference in the performance, something isnt right.

From firmware released 2021.08.02: https://forum.rme-audio.de/viewtopic.php?id=33593

"
Topic: Firmware Update ADI-2 DAC 66 / 41 - Loopback

Since about one month the ADI-2 DAC ships with an ES9028Q2M as DAC chip. This is our first unit with ESS. In the transition process we had to learn a lot, about the chip itself, how it behaves, how to best use it, what influence the clocking has, what tech specs are missing in the 'secret' data sheet (tons of..), and how to make the DAC operate and behave as similar as possible to the former AKM chip.

The result was the shipping version FPGA 60 and DSP 39. This is a big step up from the former 35 / 35, the numbers indeed equal many steps of development. And we aren’t finished, some fine-tuning is always possible and will be done.

Such fine-tuning is the main reason for the new version FPGA 66 DSP 41. We managed to get the THD+N performance identical between asynchronous and quasi-synchronous mode, so the new option Varipitch has been removed. That mode (DAC fully synchronous to the incoming SteadyClock) is now always active. Its big advantage (and the reason to add it in the first place) was that in this mode the chip does not perform a random, short mute state after sample rate change with audio already playing. So we can now offer the performance of the ESS chip with the better behavior of the AKM chip.

But this update is also useful for older DACs, as it officially adds the new option Loopback to USB. It is described in text and block diagram of the updated manual 2.8:
http://www.rme-audio.de/downloads/adi2dacr_e.pdf
http://www.rme-audio.de/downloads/adi2dacr_d.pdf

"

 

[Skeptischism]

As I was saying

 

[Trell]

But RME implementation is close enough that the specs are the same but you don't believe that is true, if I understood you correctly. Given their reputation and history, I believe them, but this will be shown when a device with the new chip is tested.

 

[Music1969]

so if the performance is the same, theyve screwed up somewhere.

So if the RCA/XLR/Phones analogue output performance (not DAC chip output - we don't listen to the direct output of the chip) remains 'state of the art', then RME have 'screwed up ' ?

 

[Skeptischism]

I dont know how else I can say this. perhaps you guys atr just audio users and not terribly involved in hardware from a design POV?

If it's the same, they are not getting good performance out of the 'New' chip (this is not a new part, its a last generation mobile DAC, designed for phones)). Even in spite of this, the performance should be minimum 6dB BETTER with the new version. as long as you didnt totally screw up the layout, you should be able to match the AKM even just running the ESS in voltage output mode.

 

[Skeptischism]

So if the RCA/XLR/Phones analogue output performance (not DAC chip output - we don't listen to the direct output of the chip) remains 'state of the art', then RME have 'screwed up ' ?

yes

 

[Music1969]

the performance should be minimum 6dB BETTER with the new version. as long as you didnt totally screw up the layout, you should be able to match the AKM even just running the ESS in voltage output mode.

This would require quite the re-design of analogue stage?

Why should they do this, given the already state of the art performance?

 

[Skeptischism]

why? because if RME dont get datasheet performance from a new part they are using, they have screwed up. I really dont know why this is so difficult for you guys to understand ...

 

[Skeptischism]

The old ADI-2 is a great unit. the overall specs are still great, all aspects considered, but I would not call -112 sinad 'state of the art'. It may have been a few years ago, but it isnt now and any designer than only got -112dB from an ESS dac would have failed in their implementation. IVX's E1DA beats that with a DAC that costs less than $200 (but it lacks features vs the ADI-2) actually I think its less than $100. the power of this unit is in all its features, a high performing ADC and headphone amp, as well as their trademark rock solid DSP and firmware, but as far as headline specs are concerned, I have a 7 year old DIY ES9018 PRO that handily outperforms it. I thought i would go back and check to see if they got better than datasheet performance out of the AKM ... and ... they didnt ... very slightly worse actually.

 

[Skeptischism]

Basically that would be the equivalent of only getting -104dB SINAD out of the AKM. Trust me, if that happened they would be trying again and if the designer was new, they might be looking for another job.

But, I doubt very much it measures the same. I have a lot of respect for RME. the first serious audio interface I bought, back in the 90's was an RME Fireface400; it's where my audio journey began. That interface and a set of Alesis active monitors changed the way I thought about sound and the drivers were absolutely solid, in a time when that wasnt the norm. The matrix mixer was the first of its kind and others emulate it to this day.

I would expect the new ADI to come in at at least -116dB, or -118 SINAD. not massively better, but better.

the ADI-2 was able to stay at the top of the recommendations here, long after the AK dac was getting a bit long in the tooth.

 

[3125b]

I would expect the new ADI to come in at at least -116dB, or -118 SINAD. not massively better, but better.

The AK4493 version reached close to 117dB SINAD at full output in Amirs review, wich makes it the best 4493 implementation I can think of right now.
But who buys these kind of things for the last dB of SINAD or the DAC chip anyway. Some companies sell mainly by the DAC chip used (while actual performance might or might not be good), but RME never put any emphasis on that. They sell the product by their own measured specs and (mainly) features and useability.

 

[devink]

According to the RME manual, this is the updated THD+N for the ADI-2 DAC: THD+N @ 0 dBFS: -116 dB, 0.00016 %

Which is better than the what the ADI-2 Pro FS R BE manual quotes: THD+N @ 0 dBFS: -115 dB, 0.00018 %

This seems like an odd situation to me where now the cheaper DAC actually has better specs than the "Pro."