L.K.S. Audio MH-DA004 Dual ES9038PRO DAC - Van Damme's double impact?
Apr 25, 2020 at 6:49 PM Post #2,551 of 4,419
@b0bb Thank you a million times for your help with this! I successfully flashed both firmware, and now have DSD512 over USB.

Unfortunately, I also blew the Mudorf Supreme fuse with a startling "pop" and spark unplugging the power cord just before reattaching the back plate. That was a $100 or so mishap.

Fortunately, I had one last fuse for the LKS, a Synergistic Red fuse.

The sound quality took a significant leap, both with redbook playback and DSD vs any previous configuration I used. I think it may be time to sell my Sonore ultraDigital.

b0bb and others with this firmware update: did you experience improved sound quality after the firmware update as well?
 
Apr 25, 2020 at 7:42 PM Post #2,552 of 4,419
LKS STOCK IMPRESSIONS

So this DAC has been on my list for a couple years. I have done a lot of R&D and testing with different types of circuits and design approaches. I will say that if I had to create and design a DAC, this LKS DAC does almost everything right. Got the DAC on February 24th, but waited until now to post my experiences (it took this long to go through all that).

The current release LKS looks a little different then the pictures I have seen, probably something like Gen 3 at this point. Those yucky silver mica caps are gone and replaced with normal film caps (good thing!). That would have been one of the first things I did. This DAC is quite heavy. Excellent build quality.

Initial sound was extremely clean, but with a tight sound and a tight edge (first power up). Just a couple of hours made a big difference in sound. The stock sound is so clean! With excellent high frequency resolution and accuracy. There have been complaints in the forums for the ESS DAC chip having a “glare”. Listening to this, I can sort of see what they are talking about, but it really depends on system. The stock DAC is so high resolution that any silver cables will definitely be too bright! I swapped the stock fuse with a Furutech and that got rid of some of glare characteristic. ..wiating to see how this breaks in.

Actually, I’m wondering if some of the glare you guys are talking about are coming from those silver mica caps on the I/V stage. I do know that they are usually bright in the highs and silver mica are known to be microphonic, which supports the idea that putting in EFI shielding against the transformers will help the glare.

========

My system uses all Audioquest NRG-4 power cables chopped and re-terminated with Furutech rhodium plugs. All interconnect are 20awg solid core OCC copper braided and using Furutech rhodium XLR connections. Preamp is highly modified rebuilt Rane stereo line mixer with outboard linear power supply and uses a Sonic Imagery 994 op amp for input stage and two Burson V6 Vivid op amps for balanced output stage. In the line mixer I have actually completely bypassed two op amp circuits in the middle. Connected to Yamaha HS8 studio monitors that are internally rebuilt with upgraded capacitors and dynamat on interior cabinet. All have Furutech fuse upgrades.

========

Well, I’ve fully burned it in with a new 3.15A Furutech fuse (last one available).

The sound is extremely clean. However, it’s somewhat forward in the upper mids and has a bit of a glare. It’s also not as natural sounding as my highly modified/rebuilt Emotiva DAC. It also doesn’t have as much high frequency extension and air. The bass is also not as hard hitting as my Emotiva DAC. At this point, I can see why the LKS dac has not won some DAC shootouts.

The glare could be from a few different things:
  • Glare from Nichicon MUSE capacitors
  • 575 clock for DAC needs to be upgraded
  • EM/noise or other elements
Modification project to follow.
stock-lks.JPG
 
Apr 25, 2020 at 7:43 PM Post #2,553 of 4,419
LKS MODIFICATIONS

I have done a significantly more amount of work than most people would do. However, it all adds up in the end. With all the upgrades, this DAC no longer resembles stock. The resolution and attack are amazing. High frequency detail is significantly improved and much more realistic. Bass is significantly stronger (think “punch you in the chest” type of bass). This DAC is now absolutely beautiful to listen to.

This is the work I have done:

A/C input
--------------------
- Furutech gold IEC => Upgraded to Furutech Rhodium IEC
- Fuse => Upgraded to Furutech Rhodium 3.15A (not available anymore)
- Transformer Secondary wires => twisted like everyone else
- Rectifier Diodes => All 24 diodes upgraded to 8A Vishay Schottky Diodes (VS-8TQ100-M3)


Digital Power Supply
--------------------
- 4x6800uf Gold Tune caps => Initially I upgraded these to 4x10,000uf Nichicon KW. I can’t remember where I read this, but I’ve read that Gold Tune are not as tight sounding and somewhat laid back. Anyways, I wanted to boost the capacitance, so I used KW caps. Later, I decided to go and upgrade to all Mundorf. I used a Mundorf 22,000uf 25V for the left most largest cap. I used Mundorf 10,000uf 25V for the 3 remaining. I did some measurements and these caps are only seeing 9V DC after the rectifier diodes, so the 25V Mundorf caps are totally safe and provide much larger capacitance. They are also 125 degree caps. I soldered Vishay 0.1uf MKP film caps to the leads underneath the board because they wouldn’t fit in the spaces between these large caps.
- 390uf post-regulator caps => Upgraded to 470uf Nichicon FG.
- 0.1uf MKP post-regulator => Upgraded to Vishay 0.1uf MKP (BFC241641004) except for one spot where I had to use a smaller WIMA 0.1uf MKP.
- LM317 regulators => I thought about upgrading these to the Sparkos SS1117 discrete regulators, but the Sparkos are internally limited at 1A. The LM317 allow up to 1.5A of current, and we know the DAC chips love the current. I don’t think I’m getting any less of performance at this point. Upgrading to Sparkos would require tweaking of the SMD resistors in front of the regulators.


Digital Circuits
--------------------
- Replaced the input coupling caps (big 0.1uf WIMA) on the S/PDIF inputs with straight wire (Ric Shultz trick). I used Neotech 20awg solid-core OCC copper that has been stripped.
- 270uf Polymer caps => All these were upgraded to Nichicon 330uf polymer. It’s not much, but every little bit helps! The stock 270uf caps are actually very good, so you can skip this part if you want. Make sure you re-solder those film caps on the bottom leads.
- 100uf Polymer caps => I left these alone because the stock Panasonic SEPC had such a great ESR (8 mohm). I couldn’t beat that.
- 0.1uf film caps => replaced these with Vishay 0.1uf MKP
- Crystek 575 => I did several clock tests before I decided on the Accusilicon AS318-B-100 (clock comments are in following post). I soldered a small 0.01uf film cap across the GRND/POWER pins on the bottom side of the board. For testing purposes, Mil-Max has a nice 14-pin socket with only 4 pins loaded so that you don’t have to melt out those other pins (digikey ED90428-ND). For final clock, I remove this socket and soldered the clock directly in.



Analog Power Supply
--------------------
- 2x4700uf Mundorf => Upgraded these to 2x10,000uf 40V Mundorf
- 2x1000uf post regulator => Upgraded to 2x1000uf Nichicon FG
- 220pf regulator MKP => Upgraded to 220pf Kemet PFR film-on-foil caps
- 0.1uf film caps => Upgraded to 0.1uf Vishay MKP


Analog I/V section
--------------------
- 22uf MUSE caps => Upgraded to 22uf + 47uf Nichicon FG. The 47uf caps used the hole for 0.1uf film caps. I soldered WIMA 0.1uf MKP to the leads underneath the board. For many years, I thought Nichicon MUSE were the best caps to use. However, about a year ago I did some critical comparisons for the heck of it and found that FG (Fine Gold) actually sounded better. The MUSE caps are nice and very high resolution, but they have a little bit of glare in the upper mids/highs. The FG caps actually sound much more natural and realistic and they have excellent high frequency response! I increase the capacitance for each discrete section. I found that having a bit more capacitance here helps with bass/midbass punch and strength. You definitely do not want to starve these discrete Class A circuits!
- 220pf MKP caps => Upgraded to 220pf Kemet PFR film-on-foil caps. These are the 3 caps in the middle if the discrete circuit (on earlier models, these are the silver-mica caps) I have found that these film-on-foil are superior to any MKP. However, they are only available in small values (up to 2200pf picofarad).
- 2200pf film caps => Upgraded to 2200pf Kemet PFR. These are the two inline rows of polyester caps above the circuit.
- 61.8 ohm resistors => Upgraded to Vishay CPF3 3-watt 61.9 ohm resistors. Like many have recommended, I could have gotten the TX2575-3 z-Foil resistors, but I didn’t want to deal with the lead time and $70 per resistor. In past projects, I have found that a higher wattage resistor actually is more transparent and provide better response. I learned this when comparing a .5w resistor to a 1w resistor in a op amp negative feedback circuit. The CPF3 are only $3.25 each (mouser 71-CPF361R900BEE14). I think they give excellent results! They are long and have to be soldered into the extended holes. These resistors have excellent specs at 0.1% tolerance and Temperature Coefficient of 25PPM/C (we all know this section likes to heat up). There are resistors that have lower PPM specs, but they are generally a much higher resistance. This was the best resistor I found between Digikey/Mouser for this application.

Mod-lks1.JPG
 
Apr 25, 2020 at 7:45 PM Post #2,555 of 4,419
LKS MODIFICATION NOTES

Some things to keep in mind if you jump into this level of modification.

The two 4700uf Mundorf caps in the analog power supply section do not have a drain circuit and keep their charge for a very long time. These caps see 20V DC. Even after 24 hours of being unplugged, the caps are only down to 15V. You need to have a way to discharge these caps before working on this board. I use a Fluke digital multi-meter that has a built-in discharge function. Just measure the leads on the caps for capacitance and the Fluke will automatically show a “DiSC” discharge mode on the screen. Just sit there and let the caps discharge (may take 1-2 minutes). It should discharge down to something like 1V. Otherwise, you may be best just leaving the LKS unplugged for about a week or so to naturally discharge these caps.

If you want to replace the I/V section caps/resistors, you will have to remove the heat-sinked transistors from the bottom of the board. This transistor heat sink requires a lot of room, which means you cannot solder/cut the leads normally. The height of this transistor heatsink must match the height of the Sabre DAC chip heatsink blocks on the bottom side of the board. If the transistor heatsink is too high, then the DAC heatsink blocks will not touch the case and you will actually force the circuit board to flex when you screw it down into the case. When you install the new 220pf film caps, insert them into the hole and then use electric tape to tape/hold them down. On the bottom side, use a precision wire cutter to snip the lead so that they are flush with the surface of the board. Then just apply a small dab of solder just to hold them in place and make contact. This should be done with the leads of the resistors as well. When re-installing the transistor/heatsink, make sure the transistor leads are bent at a perfect 90degree angle. This should get you where you want.

I used Arctic Silver thermal compound (very thin layer) on the bottom of the heat sinks for both transistors and DAC chip blocks.

When removing the stock Crystek 575 clock. Most of the users here have recommended using a hot air re-work station. You can definitely do this. However, you really need to be patient for those solder points to heat up. I’ve sat there for 3-7 minutes at a time and it’s painful to do. You also have to make sure you don’t use too much air force or you will start blowing off other small SMD components! I used a Hakko FM-2022 Parallel Tweezer with some 8mm shovel tips (Hakko T16-1006). It literally took 2 seconds to heat the solder and I just lifted the clock right off the board! If you have two soldering irons, you can get a couple flat shovel tips which will do the same thing. It’s much cheaper than getting a tweezer iron with proper power supply.
 
Apr 25, 2020 at 7:47 PM Post #2,556 of 4,419
LKS CLOCK IMPRESSIONS

Crystek 575 – I only used this with the LKS as “purely stock”. Of course, this presented the expected result. Extremely clean, but very forward in the upper mids (can come across as a glare). Stock DAC also did not have the high frequency response and realism. Lacked the bass punch/strength that I wanted.

The following impressions were all done after I completely rebuilt the DAC, so I don’t know exactly which modification did what. I used the Rhea board for the ABLNO and 950X with a 0.01uf 0808 SMD cap on the bottom side (digikey 399-14547-1-ND)

Abracon ABLNO => I only had this one in for 24 hours because I was doing my final modifications and needed to make a decision. That being said, I wanted to try this clock to see how it results on a fully modified LKS because it had very good PPM numbers. In the beginning, the sound was very sharp in the upper mids and high frequencies. After 30-60 minutes, I could start hearing a ringing/brightness in the highs. It came across as annoying sounding (listening fatigue). After 24 hours, the sharpness somewhat calmed down, but it still had that high frequency brightness/ringing. None of the other clocks had this problem, even in the first 24 hours. I don’t think this clock is the right one for LKS / ESS.

Crystek 950X => This clock actually has a very slightly soft attack on the sound throughout the entire frequency range. Don’t get me wrong, all the frequencies are there and high frequencies are represented, but it’s just a little bit soft (a little bit laid back). Kind of tilts toward a McIntosh type sound (maybe not quite as laid back). I can definitely see where people complain about it being too laid back with the stock Gold Tune caps and components. Nice clock, but not what I wanted.

Accusilicon AS318-B-100 => This clock is actually pretty damn amazing. It has none of the problems of previous clocks and much more resolution than 950X. Everything comes across in a very exciting presentation. The music and sounds are just so much more realistic with significantly more impact. An amazing result. I think internally, this may be a VCXO clock. It’s completely sealed like an OCXO and the documentation talks about “Pin 1 OE being pulled up to VDD by default”. Not sure, however. This is my recommendation if you don’t want the laid back style of the 950X. I purchased this from Audiophonics. Cost was $90USD when I added in the TNT express shipping. If you do order from them, make sure you message them as ask them to use a box for shipping. Mine came in a padded envelope and it actually bent down the pins. I was able to salvage it, however.
 
Apr 26, 2020 at 9:37 AM Post #2,557 of 4,419
Hello guys. First of all, I apologize for my supreme ignorance but, as I said earlier, this is my first dac and my first step in digital audio. I already have the LKS at home but I am still waiting for my "Audioquest" usb cable. As "b0bb" recommends, while I wait for the cable to arrive, I will use the stock cable. I have my pc with windows 7 and, at the moment, I will use Foobar. I know I have to go to "Amanero firmwares" and since I am afraid of doing something wrong due to my ignorance plus the non-existent LKS instruction manual, this is where my doubts begin:
-... which firmware exactly should I download for my W7 + Foobar ...?
-... must I have the usb cable connected to pc and lks ...?
Please, if someone (... I am already ashamed to quote "b0bb" to ask for their help ...:flushed:) has the patience and wants to indicate the exact steps to proceed from 0, I will be very grateful. Greetings to all.
 
Apr 26, 2020 at 12:34 PM Post #2,558 of 4,419
-... which firmware exactly should I download for my W7 + Foobar ...?
The "official" support : http://mu-sound.com



The latest Amanero usb driver (not firmware)
https://www.amanero.com/drivers.htm

After driver installation, in foobar, select DS: ASIO 1.03 driver
play some music files, make sure that your Windows 7 foobar can communicate with LKS .

For upsampling your PCM files (much better than foobar's stock DSP)
https://www.foobar2000.org/components/tag/DSP

install SRC Resampler 1.0.11
in foobar.
 
Apr 26, 2020 at 3:43 PM Post #2,559 of 4,419
the one I bought is for uk plug 25mm 13A. you can contact him if you want to buy https://www.facebook.com/audiolinksince1997/
20190817_220058.jpg

I've found these from several different HK-based sellers across the internet, and they are all asking about $375 (US) for one fuse. That is more than twice what I would be willing to pay to try it. Has anyone here found it at a lower price, or had an ebay offer accepted at a lower price?
 
Apr 27, 2020 at 10:28 AM Post #2,561 of 4,419
I've found these from several different HK-based sellers across the internet, and they are all asking about $375 (US) for one fuse. That is more than twice what I would be willing to pay to try it. Has anyone here found it at a lower price, or had an ebay offer accepted at a lower price?
I have replaced the Zero fuse with SR Orange at the IEC of dac. The Orange is significantly better. Anyone using LKS004 should replace the stock fuse, it sounds so harsh and bright.
 
Apr 27, 2020 at 9:48 PM Post #2,564 of 4,419
I have replaced the Zero fuse with SR Orange at the IEC of dac. The Orange is significantly better. Anyone using LKS004 should replace the stock fuse, it sounds so harsh and bright.
Well, now we're getting into less outrageous territory. :) The SR Orange fuses are receiving complaints all over the internet for blowing on first or second power on after installation at values even 50% higher than rated value. What value SR Orange did you go with for the LKS?

Out of curiosity, have you compared the various SR fuses?

Thanks!
 

Users who are viewing this thread

Back
Top