RME LNI-2 DC honest talk

[FrantzM]

If you search for RME linear power supply, there are several vendors making money selling pointless upgrades to an already extremely well measuring and performing device. Perhaps RME thought "if customers feel they need such a thing, it would be better if we designed it, so that it properly compliments the internal regulators in our kit".

May I suggest an edit:

If you search for RME linear power supply, there are several vendors making money selling pointless upgrades to an already extremely well measuring and performing device. Perhaps RME thought "if customers feel they need such a thing, it would be better if we sell one, so that we can increase our profits a bit

Peace.

 

[KSTR]

TRACO THM 15-1212
Ripple and Noise: 75 mVp-p typ
Leakage current: less than 2.5 µA

Considering that galvanic isolation is more important than actual output characteristics, are you satisfied with this module? Have you measured its efficiency?

No. I trust TRACO's specs wrt efficiency. I've actually used the 15V output model simply because I had it already. I also added additional input and output caps (10uF MLCC) and some RCR-filtering (2x 0.47R + 470uF e-cap + 2x 0.47R to reduce any chance of CLC ringing, the L being the cable. ;-)

 

[KSTR]

May I suggest an edit:

we sell one, so that we can increase our profits a bit...

Peace.

... and re-invest it elsewhere.

Of course, that's the whole idea behind setting-up a company in the western world.

 

[Bob from Florida]

Friends, does this look like a good candidate as a competitor for LNI-2?
View attachment 396039

Take a look at this video of a comparable lithium battery being disassembled. You will discover that these type batteries maintain voltage via internal regulator circuits. These are highly unlikely to be an improvement over your existing power supply.

If you really want to use battery power, get an AGM Odyssey battery.

 

[johny_2000]

Take a look at this video of a comparable lithium battery being disassembled. You will discover that these type batteries maintain voltage via internal regulator circuits. These are highly unlikely to be an improvement over your existing power supply.

I'll be using this 18650x6 12V Li-Ion power bank. I checked the wiring and made sure that power from the batteries was going directly to the DC output jack using the rocker switch. The remaining components on the PCB serve as a battery level meter (LM339 + 5 LEDs) and a step-down DC/DC converter for the USB-A 5V output.

 

[Bob from Florida]

I'll be using this 18650x6 12V Li-Ion power bank. I checked the wiring and made sure that power from the batteries was going directly to the DC output jack using the rocker switch. The remaining components on the PCB serve as a battery level meter (LM339 + 5 LEDs) and a step-down DC/DC converter for the USB-A 5V output.

View attachment 396236 View attachment 396235

Still, very unlikely to make any difference. The internal power supply regulators in the RME are what matter. Unless the external supply is sending ground loop hum, the battery won't change anything. Batteries are very handy when running off "the grid"....

 

[johny_2000]

In a couple of days I will measure all three power sources - factory power supply, LNI-2 and battery power, under the same test conditions.

 

[johny_2000]

The few dozen measurements I've made over the last day look promising. There is definitely a correlation between the power source and the output results.

 

[restorer-john]

I'll be using this 18650x6 12V Li-Ion power bank. I checked the wiring and made sure that power from the batteries was going directly to the DC output jack using the rocker switch.

6 cells. 3.7V each. Series 22.2V. 3 series pair, twin parallel 11.1V.

What's on the other side of the board?

 

[KSTR]

The few dozen measurements I've made over the last day look promising. There is definitely a correlation between the power source and the output results.

Looking forward to your test results.

FWIW, I once tested the ADI-2 Pro with an undersized old-school 30 years old linear supply wall-wart (reservoir e-caps surely pretty much dried out), having more than 1Vpk-pk of 100Hz (and harmonics) ripple. Absolutely zero effect on the DAC output as measured with AP 2322 or a second ADI-2 Pro.

As noted before, we have to make sure we don't measure any disturbance coming solely from the interconnect cables and mistake it for DAC/ADC error, notably when looking at unbalanced outputs.
These days, I'm using Cosmos ADC with a Topping HS01 USB Isolator to fully float the measuring ADC, and before I used a second ADI-2 Pro with Intona USB isolator and the mentioned supply isolator. And of course use the shortest possible interconnects, well-shielded ones.

 

[restorer-john]

These days, I'm using Cosmos ADC with a Topping HS01 USB Isolator to fully float the measuring ADC

I use the Quantasylum QA403 which is also fully isolated.

And of course use the shortest possible interconnects, well-shielded ones.

Very true. And keep the BNC/XLR contacts clean.

 

[johny_2000]

6 cells. 3.7V each. Series 22.2V. 3 series pair, twin parallel 11.1V.

What's on the other side of the board?

This one has a 3S2P cell configuration.

Fully charged, it outputs 12.4V. During my testing (30-45 mins), it never dropped below 12.3V.

There are no components on the other side of the board. Just a few more power traces through vias.

I cut the trace to the positive input of the USB 5V step-down regulator so that it would also not make any noise.

 

[johny_2000]

As noted before, we have to make sure we don't measure any disturbance coming solely from the interconnect cables and mistake it for DAC/ADC error, notably when looking at unbalanced outputs.
These days, I'm using Cosmos ADC with a Topping HS01 USB Isolator to fully float the measuring ADC, and before I used a second ADI-2 Pro with Intona USB isolator and the mentioned supply isolator. And of course use the shortest possible interconnects, well-shielded ones.

Thank you!

I use a Cosmos ACDiso with a built-in USB power isolator, and tests by the author and Archimago have shown that using an additional USB isolator (e.g. Topping HS02) does not improve the ACDiso measurements. But it does not make them worse if one wants to use double isolation.

The E1DA Cosmos Stack: ADCiso and 2 APUs. An example of bad multichannel - Neil Young's 'Harvest' on DVD-A (2002), fixed in new Atmos mix.

A blog for audiophiles about more objective topics. Measurements of audio gear. Reasonable, realistic, no snakeoil assessment of sound, and equipment.

archimago.blogspot.com

My interconnects are short cables hand made using Canare L-2T2S cable, Neutrik XLR M/F and Amphenol 3.5mm connectors.

 

[johny_2000]

I'm finishing all the measurements. One more set and we can process the results.

 

[somebodyelse]

Take a look at this video of a comparable lithium battery being disassembled. You will discover that these type batteries maintain voltage via internal regulator circuits. These are highly unlikely to be an improvement over your existing power supply.

I think you misunderstand what the BMS (battery management system) does. It isn't a regulator. It's mostly a switch to disconnect the battery in various fault conditions, plus the relatively low current bit to keep cell voltages in balance.

 

[johny_2000]

So here we go!

Test conditions:
DAC: RME ADI-2/4 Pro SE
Output: XLR
Level: +19.5 dBU
USB isolator: Intona High-Speed USB Isolator 7054-DK (1kV)
ADC: E1DA Cosmos ADCiso
APU17 Notch filter: E1DA Cosmos APU@1 kHz@0 dB preamp
Power sources: Original 3-prong grounded Power Supply, Lithium 18650x6 12.4V battery pack, RME LNI-2 DC.
Power grid 60 Hz.

The summary:
Original PS:
Noise=-124.5 dB SNR=121.9 dB THD+N=-119.5 dB

Lithium 12V battery pack:
Noise=-124.7 dB SNR=122.1 dB THD+N=-119.7 dB

RME LNI-2 DC:
Noise=-125.4 dB SNR=122.8 dB THD+N=-120.1 dB

Conclusions:
As RME stated, the ADI-2 unit is not sensitive to power supply. It has a marginal +/- 1 dB improvement in any output specifications with most complex power supplies, such as the original grounded power supply, lithium battery pack, RME LNI-2 DC module. This is a big compliment to RME's engineering team, who managed to maintain the specification with all USB sources and 12V power supplies.

 

[Reddoc]

As RME stated, the ADI-2 unit is not sensitive to power supply.

I like the way that you were cynical about the equivalence of the power supplies, and then you proved their equivalence, rather than just believing what you were told. Empirical evidence that matched the claim.
Tick.
Next?

 

[johny_2000]

I like the way that you were cynical about the equivalence of the power supplies, and then you proved their equivalence, rather than just believing what you were told. Empirical evidence that matched the claim.
Tick.
Next?

Thank you! I am an engineer. I don’t trust words, but science and instruments.

 

[johny_2000]

Next?

There's not much you can do with this configuration - USB Isolator and Isolated DC µFilter Power Supply.
I think this is an engineering achievement of the highest caliber at this point in audio history, and I'm glad to be up to date.

Kudos to the German engineering teams!

 

[KSTR]

The summary:
Original PS:
Noise=-124.5 dB SNR=121.9 dB THD+N=-119.5 dB

Lithium 12V battery pack:
Noise=-124.7 dB SNR=122.1 dB THD+N=-119.7 dB

RME LNI-2 DC:
Noise=-125.4 dB SNR=122.8 dB THD+N=-120.1 dB

Conclusions:
As RME stated, the ADI-2 unit is not sensitive to power supply. It has a marginal +/- 1 dB improvement in any output specifications with most complex power supplies, such as the original grounded power supply, lithium battery pack, RME LNI-2 DC module.

Excellent, many thanks for the measurements.

I would tend to file all three results under essentially identical because the differences are quite likely within run-to-run variations. To double check, I often repeat the same measurement several times on different days, rigging up the tests again from scratch, and maybe it is slightly colder or warmer or the units haven't fully warmed up etc. +-0.5dB variance is nothing.

The only thing of interest I see in the plots are the slightly different mains spurs. For some reason the battery pack setup produced the largest peaks (at 4x, 240Hz) but is cleaner everywhere else. Of course all of this is irrelevant as it is 30dB below the RMS noise floor.

May I suggest another measurement that would expose the effect of different mains leakage currents (the battery pack shouldn't have any), if you want to take the extra effort&time, that is?

• Add a 2-prong 12V/2A SMPS to the set of tested supplies.
• Leave the Intona in place to keep USB clean.
• Measure the unbalanced output of the RME with a normal unbalanced interconnect.
• Don't use the APU notch.
• Connect the shell/GND of the unbalanced interconnect at the receiving end at the ADC Iso to mains earth, to emulate an earth-grounded amp or similar. Alternatively, when using the original three supplies, connect the receiving end shell to the 2-prong supply, emulating an amp with such a supply.
• Play a lower level sine like -30dBFS, sort of a test signal the represents a quiet passages in music
• Look at the resultant SNR and spectra

By this, we will see the huge impact of the supply's mains coupling in an unbalanced user scenario. Notably the 2-prong brick will degrade the system performance big time -- even the battery pack setup will suffer a bit -- and render all the SINAD race moot because system SINAD is dominated by cable effects (I*R drop on the shield), not devices ;-)