RME ADI-2 DAC FS - AKM Versus ESS Measurements (DAC, Preamp & Headamp)

[dtaylo1066]

I just ordered an RME ADI-2 FS yesterday and happened to see in a comment on Amazon there were two different chips used. My first thought was, "Oh no, I bet mine has the bad one!" It was a relief to find this thread and see that whatever chip, RME is a company that is careful to make sure their products perform at a high level. I'll be curious to see which one I receive...

The one you will get is the worst chip, and you easily will notice the 1 DB difference in SINAD, but it will be made up for that shortfall by a lush, velvet sound with an expansive, all-engrossing soundstage that will envelop you, and highs will float like feathers in a light breeze, and mids will be sweet as Tupelo Honey. If you put white, hi-polymer erasers under the DAC's feet you will better solidify the bass, and make sure to trade out the power chord for the new Vodoo offering. https://voodoocable.net/

I also highly recommend bypassing the power supply entirely and running it off of a custom, rechargeable lithium ion DC power supply. The dead silent background is something not to be heard.

 

[VintageFlanker]

I'll be curious to see which one I receive...

It will be the ESS. The AKM is out of production for two years.

The one you will get is the worst chip, and you easily will notice the 1 DB difference in SINAD, but it will be made up for that shortfall by a lush, velvet sound with an expansive, all-engrossing soundstage that will envelop you, and highs will float like feathers in a light breeze, and mids will be sweet as Tupelo Honey. If you put white, hi-polymer erasers under the DAC's feet you will better solidify the bass, and make sure to trade out the power chord for the new Vodoo offering. https://voodoocable.net/

I also highly recommend bypassing the power supply entirely and running it off of a custom, rechargeable lithium ion DC power supply. The dead silent background is something not be heard.

Use emojis when you're sarcastic that way.

 

[rcstevensonaz]

The one you will get is the worst chip, and you easily will notice the 1 DB difference in SINAD, but it will be made up for that shortfall by a lush, velvet sound with an expansive, all-engrossing soundstage that will envelop you, and highs will float like feathers in a light breeze, and mids will be sweet as Tupelo Honey. If you put white, hi-polymer erasers under the DAC's feet you will better solidify the bass, and make sure to trade out the power chord for the new Vodoo offering. https://voodoocable.net/

I also highly recommend bypassing the power supply entirely and running it off of a custom, rechargeable lithium ion DC power supply. The dead silent background is something not to be heard.

A very well crafted tongue-in-cheek response! Of course, I have the older AKM so while I do gain that extra 1DB SINAD, I now fear how much of the velvet sound my AKM Velvet Sound chips are failing to deliver .

But to also add a +1 to what @VintageFlanker noted, adding a Humor Emoji is very important since (a) sarcasm in a forum message is not always obvious to even a native English speaker, and (b) such humor is even less obvious to non-native English speakers around the world who are also on this forum.

 

[auricom]

Is the V1 better or worse than V2 AKM?

 

[VintageFlanker]

Is the V1 better or worse than V2 AKM?

Worse, somehow. And it is also an AKM (4490). Competent enough for many use cases, tho. I just consider the V2 remote as a massive upgrade.

 

[sonitus mirus]

Worse, somehow. And it is also an AKM (4490). Competent enough for many use cases, tho. I just consider the V2 remote as a massive upgrade.

I really like the V2 remote. I had to buy mine separately for my 2018 DAC. Despite being within arm's reach of the unit and rarely use the remote, I'm still happy I got the new one.

 

[Jarlaxle]

Is the tested AKM unit Version 1 with AK4490? I couldn't see B letter in the serial number.

 

[Jimbob54]

What you are about to see is a direct comparison with measurements between the (discontinued) "Version 2", that came with AKM AK4493 DAC chip, and the revision that is actually available, now equipped with the ES9028Q2M

@Jarlaxle

 

[Jarlaxle]

@Jarlaxle

yes but

 

[Jimbob54]

yes but

I think we will have to let senor @VintageFlanker answer to that.

 

[unpluggged]

Is the tested AKM unit Version 1 with AK4490? I couldn't see B letter in the serial number.

It has the SD LD DAC filter, so it is the AK4493 version.

 

[cistercian]

It has the SD LD DAC filter, so it is the AK4493 version.

@VintageFlanker
Epic post and thank you. Happily listening to my AKM RME ADI-2 as I write this. I cried when the AKM fab burned down and bought an Icom R-30 based on the fab being trashed.
Because ICOM had a custom decoder chip built at that fab that disappeared after it burned down...and as a result discontinued the radio soon after.
All of my ICOM radios use AKM decoder chips. Because of this I am an AKM fanboy. Of course RME is staffed by excellent engineers and their product is so incredibly epic I trust them implicitly. But being an ICOM fan, I am biased. Great write up and best wishes!

PS the digital decode the chip in the ICOM R-30 has is epic. Its an amazing receiver with the best P25 decode of all my digital receivers...including SDR computer operated radios.

I am a radio geek first and an audiophile second by far. JBL 835 passive driven by Crown XLR 2506 driven by the RME should pretty much sum it up.
In my defense I have now a good headphone selection and use the Sennheiser HD-600 as my reference. YMMV but your write up is still outstanding!

 

[sonitus mirus]

What is an AKM RME ADI-2? Is it supposed to sound different from some other version?

 

[Rednaxela]

What is an AKM RME ADI-2? Is it supposed to sound different from some other version?

This is explained in detail here.

 

[Rja4000]

What is an AKM RME ADI-2? Is it supposed to sound different from some other version?

AKM is a diminutive for Asahi Kasei Microdevices, ESS stands for ESS Technology.
Both are chip manufacturers. In this case, it's about their DAC and ADC chips.

RME used to use mainly AKM chips.
As many other brands from the Pro audio industry were doing.
(My 15 years old Yamaha DM1000 V2 mixer uses AKM ADC and DAC chips, as an example)

In October 2020, a 3 days fire destroyed AKM's main factory, where DAC and ADC chips were produced.
It seems they were not able to produce chips in reasonable quantities until 2022, where they restarted with new products.

Most industrial actors using AKM chips were impacted. RME was one of them, although they said they had a serious strategic stock.

Due to the AKM shortage, though, RME decided to change the DAC chip in the ADI-2 DAC.
They also said they wouldn't change DAC or ADC in the ADI-2 Pro, since their stock was sufficient for those.
The limited edition RME ADI-2/4 Pro SE is using ESS chips, though.

I don't know for the UCX/UFX series.
I think it was confirmed that the UCX II is using AKM.
(Maybe @MC_RME may comment ?)

It seems RME managed to use the ESS ES9028Q2M in the ADI-2 DAC with doing just a few changes from the AKM base. There are differences in terms of available filters and some others, but they tried -and, it seems, succeeded- to keep functionality almost identical.

Note that it's not the first time RME changes the chip in the ADI-2 DAC. They did previously change from AK4490 to AK4493, with a small enhancement in measured performances.

As for an audible difference in sound, there are some legends around, but no evidence that there might be any difference.

The measurements clearly show transparency. At such a level of excellence, that leaves no room for actual audible differences.

Which, in other words, translates into:
There is no audible difference between the AKM and the ESS chip versions of the RME ADI-2 DAC.

 

[xaviescacs]

What is an AKM RME ADI-2? Is it supposed to sound different from some other version?

They also released a video on the topic.

 

[Rja4000]

They also released a video on the topic.

I didn't see that one. Great !

Matthias says the Intersample overload capacity in RME DAC may (partly) explain why, while ESS chip reputation is not too good, it won't make a difference in an RME device.

I measured here the capacity of supporting heavy intersample overloads for the the ESS-based RME ADI-2/4 pro SE
(test signal at 0dBFS has a level equivalent to +3dBFS for a normal test sine wave.).

The comparison between a good ESS chip based DAC (Topping D90LE) and the RME ADI-2/4 Pro SE (which uses the same ESS chip which uses a chip from the same family than the ESS-based ADI-2 DAC) shows clearly a strong difference: the Topping can't handle such overloads and starts saturation much earlier, to reach up to several % of THD at 0dBFS in this (admittedly extreme) example.

And some ESS-based DAC are probably much worse than the Topping here (which is at the top of the range of all Topping DAC models)

 

[unpluggged]

RME ADI-2/4 Pro SE (which uses the same ESS chip than the ESS-based ADI-2 DAC)

Just a small correction: ADI-2/4 Pro SE uses ES9038Q2M, as opposed to ADI-2 DAC's ES9028Q2M.

 

[Rja4000]

Just a small correction: ADI-2/4 Pro SE uses ES9038Q2M, as opposed to ADI-2 DAC's ES9028Q2M.

Sorry. My bad. I corrected above.

And the Topping at test is using the -at the time top-of-the-line- ES9038 Pro.

 

[Rja4000]

More data:

As expected, cheaper Topping performs worse.
RME UCX II is doing pretty good.

Test conditions:
DAC 48kHz
ADC 192kHz
Test signal: Sine wave 12kHz, phase 45°
BW limited to 50kHz (H2+H3+H4), (to avoid influence of High frequency noise)
Measured with Virtins Multi Instrument 3.9.8.1


Some reading about the subject:

Intersample Overs in CD Recordings

Intersample Overs are a Common Occurrence in CD Recordings We have frequently used Steely Dan's Gaslighting Abbie from Two Against Nature in our listening tests. This is a spectacular CD recording with lots of dynamics and a low noise floor. Nevertheless, in a little over 5 minutes, this track...

benchmarkmedia.com

In their list of examples, the track "Begin to Hope, Fidelity, Regina Spektor" should have overload peaks around 1.5dB on my charts.

At that level, you'd have 2,4% THD for the G5 and 0.4% for the D90LE.
And THD would be close to 0 (<0.01%) for the RMEs.