[b0bb]

Hello guys. Please, I have had the lks for a couple of weeks and I play my music through Foobar upsampling to DSD512 and everything is great, but how can I know if my Amanero board is updated or not ...? Is there a simple way (... within the reach of a newbie ...) to find out ...? Thanks guys for all your knowledge ...

Your LKS shows it is locked on to the DSD512 clock, you have a DSD512 capable version of the firmware on the Amanero.

If you must have the latest version, you will need to get help at your location, partial disassembly of the DAC is required to get this done.

 

[Gilberto 62]

Thanks b0bb. Well then I stop worrying about the firmware and keep enjoying this cool lks ....

 

[hpsxrb]

Does anyone know how much power draw the LKS 004?
Thank you.

 

[JBP3]

Question for you guys, can I drive very high efficiency speakers directly from this DAC? In other words does it have approximately 2 Watts of audio output more or less? Has anybody actually tried this? For example I understand this is done with the Chord Dave. Thanks in advance for any insight!

 

[Forgisound]

Question for you guys, can I drive very high efficiency speakers directly from this DAC? In other words does it have approximately 2 Watts of audio output more or less? Has anybody actually tried this? For example I understand this is done with the Chord Dave. Thanks in advance for any insight!

Not. That's not a good idea. DACs typically have very little power. Often they can't even drive amplifier.

 

[MartinWT]

My LKS drives a Burson buffer before the power amp, which greatly improves the heft and slam in delivery. There is no chance that the LKS could drive speakers directly.

 

[JBP3]

Thanks to all for the reply. And also thank you for the heads up on the person buffer, I looked into it and it looks intriguing.

 

[JBP3]

Sorry, that would be Burson buffer.

 

[hpsxrb]

b0bb please,
Following a statement from you on this subject, can you please tell me if there are already proven benefits in replacing the rectifier diodes provided for in the LKS update (MBRF10H100-E3/45) with the Schottky Silicon Carbide Diodes?
I looked for something like 100V / 10A but with a good price and only found these on Mouser:
511-STPSC8065D (8A / 650V)
726-IDH06G65C6XKSA1 (16A / 650V)
Thank you.

 

[b0bb]

b0bb please,
Following a statement from you on this subject, can you please tell me if there are already proven benefits in replacing the rectifier diodes provided for in the LKS update (MBRF10H100-E3/45) with the Schottky Silicon Carbide Diodes?
I looked for something like 100V / 10A but with a good price and only found these on Mouser:
511-STPSC8065D (8A / 650V)
726-IDH06G65C6XKSA1 (16A / 650V)
Thank you.

Mouser Germany sells the MBRF10H100 for 1.48€
https://www.mouser.de/ProductDetail.../MBRF10H100-E3-45?qs=cfR2GKjf2Rmfh8eE2yEj/w==

They are fully insulated, this reduces the chance short-circuit accidents.

I hve not used the alternatives but the datasheet numbers look good, on paper at least.

 

[hpsxrb]

Mouser Germany sells the MBRF10H100 for 1.48€
https://www.mouser.de/ProductDetail/Vishay-Semiconductors/MBRF10H100-E3-45?qs=cfR2GKjf2Rmfh8eE2yEj/w==

They are fully insulated, this reduces the chance short-circuit accidents.

I hve not used the alternatives but the datasheet numbers look good, on paper at least.

b0bb but these are not Silicon Carbide Schottky (these are the ones mentioned earlier in the LKS update)!

 

[b0bb]

b0bb but these are not Silicon Carbide Schottky (these are the ones mentioned earlier in the LKS update)!

Never tried the SiC diodes, not yet available when I made the mod.

Not sure about why you are excited over the SiC diodes.
SiC leaks more current, 25X worse (105μA vs 4.5μA). This is a path for noise, and the LKS004 is very sensitive.

To put things in perspective, the fullscale (FS) current output of the LKS004 is 60mA or 60000μA
105uA is 55dB below FS (-55dBFS)
4.5uA is 83.5dB below FS (-83.5dBFS)

There is a 28.5dB difference in the noise profile, the best case noise performance of the ES9038 is -135 dBFS
That is a significant chunk of the noise headroom gone if you use the noisy SiC diodes.


SiC drops more voltage, 2X worse (1.45V vs 0.77V) higher, the voltage regulator loses working headroom.
On the diode setup on LKS004, you lose 2.9V on SiC and that is 20% of the supply voltage vs 10% (1.44V)

Regular 10H100 Schottky

STPSC8065 SiC Schottky

 

[tdrood]

Never tried the SiC diodes, not yet available when I made the mod.

Not sure about why you are excited over the SiC diodes.
SiC leaks more current, 25X worse (105μA vs 4.5μA). This is a path for noise, and the LKS004 is very sensitive.

To put things in perspective, the fullscale (FS) current output of the LKS004 is 60mA or 60000μA
105uA is 55dB below FS (-55dBFS)
4.5uA is 83.5dB below FS (-83.5dBFS)

There is a 28.5dB difference in the noise profile, the best case noise performance of the ES9038 is -135 dBFS
That is a significant chunk of the noise headroom gone if you use the noisy SiC diodes.


SiC loses more voltage, 2X worse (1.45V vs 0.77V) higher, the voltage regulator loses working headroom.
On the diode setup on LKS004, you lose 2.9V on SiC and that is 20% of the supply voltage vs 10% (1.44V)

Regular 10H100 Schottky


STPSC8065 SiC Schottky

b0bb, I have a question for you related to the data in this post... the 10H100 table lists only the maxima for each parameter while the 8065 table includes a min, max and typ (ok, so there are no values in the min cells, so scratch the min). I guess you have to default to the worst case (max) when doing this evaluation, but I am curious if the typical values have any merit/meaning. In my world, data must be statistically significant to a 99% probability/95% confidence, (to which we add an additional 50% margin) but I design structural repairs for commercial aircraft, so lives are literally at stake and typical values are, therefore, useless to me. But, that may not be the case here(?), and the typical values may be more realistic, real world values - hence my question. To be clear, I guess the question is, generally speaking, do the ‘typ’ values have any merit?

But I digress... I see that the reverse current in the H100 is up to 6,000 microA (6.0mA) @ 125C - a very sharp rise over the 25C value. In fact, it is an order of magnitude greater than the 8065@150C (0.75mA). Based on the slope of that curve, one would rightly exect the H100 values to be significantly higher still at 150C, even if that rise is linear. I don’t know what the actual service temps of these devices are in our LKS004s, but it does seem conceivable that the reverse current may actually be better in the 8065 devices(?), unless the curve on the H100 is very non linear and/or the operating temps are low. That said, the data definitely favors the H100 for fwd voltage drop, regardless of temp. I am keen to hear your comments regarding the reverse current data.

I really appreciate your thoughtful - and knowledgeable - comments on this thread. As I alluded above, I am an aerospace structures engineer; I am an electronics hobbies wannabe at best, and you have taught me a LOT, thank you!!

Peace,

T

 

[b0bb]

b0bb, I have a question for you related to the data in this post... the 10H100 table lists only the maxima for each parameter while the 8065 table includes a min, max and typ (ok, so there are no values in the min cells, so scratch the min). I guess you have to default to the worst case (max) when doing this evaluation, but I am curious if the typical values have any merit/meaning. In my world, data must be statistically significant to a 99% probability/95% confidence, (to which we add an additional 50% margin) but I design structural repairs for commercial aircraft, so lives are literally at stake and typical values are, therefore, useless to me. But, that may not be the case here(?), and the typical values may be more realistic, real world values - hence my question. To be clear, I guess the question is, generally speaking, do the ‘typ’ values have any merit?

But I digress... I see that the reverse current in the H100 is up to 6,000 microA (6.0mA) @ 125C - a very sharp rise over the 25C value. In fact, it is an order of magnitude greater than the 8065@150C (0.75mA). Based on the slope of that curve, one would rightly exect the H100 values to be significantly higher still at 150C, even if that rise is linear. I don’t know what the actual service temps of these devices are in our LKS004s, but it does seem conceivable that the reverse current may actually be better in the 8065 devices(?), unless the curve on the H100 is very non linear and/or the operating temps are low. That said, the data definitely favors the H100 for fwd voltage drop, regardless of temp. I am keen to hear your comments regarding the reverse current data.

I really appreciate your thoughtful - and knowledgeable - comments on this thread. As I alluded above, I am an aerospace structures engineer; I am an electronics hobbies wannabe at best, and you have taught me a LOT, thank you!!

Peace,

T

The relationship is highly exponential as shown in the picture below.
The 10H100 runs very cool approx 35C

This came from an appnote from ST
https://www.st.com/resource/en/appl...osses-in-a-power-diode-stmicroelectronics.pdf

Tj = 35

For the 10H100
Tjref2-Tjref1 = 100
c = 0.0719
Ir = 9.24uA

For STPSC8065 SiC Schottky
Tjref2-Tjref1 = 125
c = 0.0157
Ir = 122.9uA

SiC Schottky is 13.3x worse @35C

In practice the SiC will run hotter given 2x the voltage drop it will run hotter and with a consequent increase in leakage current.
At 45C, leakage increases to 144uA, this makes it 15.6x worse than the conventional silicon schottky diode.

This is -52dBFS, assuming the voltage regulator reduces noise by 30dB, the net SNR would be 82dB.
This is about as good as 1980's era DACs. Good 16bit DACs are capable of 90dB SNR.

A 24bit DAC has a theoretical SNR of 144dB.

 

[b0bb]

To be clear, I guess the question is, generally speaking, do the ‘typ’ values have any merit?

Max values are guaranteed by the manufacturer, typ values are not.