[b0bb]

I'm waiting on the FIFO and clock boards from Ian here: https://github.com/iancanada/DocumentDownload and I have one of the new XMOS boards to use just for the USB input (it seems to have a better Linux DSD native support than Amanero) http://www.diyinhk.com/shop/audio-kits/107-xmos-dsd-dxd-768khz-high-quality-usb-to-i2sdsd-pcb.html but I was thinking of feeding the output from the above via the I2S Ethernet input of the 004.
If I understand correctly, I should do it via the Amanero header instead?

Looked at those a while back, Ian's board did not support DSD, the upcoming v2 does. I dislike having to change sources just to play PCM vs DSD.

The work goes beyond wiring it into the I2S header. You still have to deal with clock sync problems as there 2 clocks in the I2S stream, this is something the got Singxer wrapped around the axle with the clicks and bangs. Talk to @abartels if you want to find out how things can go south with Ian's board, he used it on the AKM dac.

The reclocker in the 9038 cannot be bypassed in the LKS, so most of work reclocking at the I2S side is wasted.
You end up with 2 reclockers in series, one in Ian's board and the other in the 9038, if the Singxer F-1 with on board reclocker is any indication, this is not a happy state of affairs.

It remains to be seen if the reclocker in Ian's FPGA can do a better job than the ESS one in the 9038.

I would toss the 575 controlling the reclocker on the 9038 and replace it with and ovenized XO (OCXO), I used 2 different OCXOs on the 003 and its on my todo list for the 004

One small saving grace for Amanero is that it now has an in-kernel Linux driver.

(If you really want to experiment to see what Ian's FIFO board is truly capable of, get one of the eval boards from DIYINHK and run it in sync mode to disable the 9038 reclocker.
The LKS 004 is hardstrapped in async mode and is ill suited for the purpose.
With the eval board you can experiment with the various USB interfaces out there, it is a lot cheaper if you make a mistake, 2 people have damaged their 003s here, one recovered, the other did not.

http://www.diyinhk.com/shop/audio-k...t-perfect-volume-control-and-spdif-input.html)

 

[vick_v]

Looked at those a while back, Ian's board did not support DSD, the upcoming v2 does. I dislike having to change sources just to play PCM vs DSD.

The work goes beyond wiring it into the I2S header. You still have to deal with clock sync problems as there 2 clocks in the I2S stream, this is something the got Singxer wrapped around the axle with the clicks and bangs. Talk to @abartels if you want to find out how things can go south with Ian's board, he used it on the AKM dac.

The reclocker in the 9038 cannot be bypassed in the LKS, so most of work reclocking at the I2S side is wasted.
You end up with 2 reclockers in series, one in Ian's board and the other in the 9038, if the Singxer F-1 with on board reclocker is any indication, this is not a happy state of affairs.

It remains to be seen if the reclocker in Ian's FPGA can do a better job than the ESS one in the 9038.

I would toss the 575 controlling the reclocker on the 9038 and replace it with and ovenized XO (OCXO), I used 2 different OCXOs on the 003 and its on my todo list for the 004

One small saving grace for Amanero is that it now has an in-kernel Linux driver.

(If you really want to experiment to see what Ian's FIFO board is truly capable of, get one of the eval boards from DIYINHK and run it in sync mode to disable the 9038 reclocker.
The LKS 004 is hardstrapped in async mode and is ill suited for the purpose.
With the eval board you can experiment with the various USB interfaces out there, it is a lot cheaper if you make a mistake, 2 people have damaged their 003s here, one recovered, the other did not.

http://www.diyinhk.com/shop/audio-k...t-perfect-volume-control-and-spdif-input.html)

Thanks for the pointers!
I am getting the new multi channel boards so they should support DSD and PCM. Also there was an ALSA patch that introduced a short mute when there is change in the stream which eliminated pops when changing from DSD to PCM (my understanding is that the DSD spec requires something like a 50ms mute at the beginning and end of the stream).
I did read abartels blogs about using Ian's FIFO board in his overbuilt DAC but I didn't see any info about him removing the board, I'll reach out to him for more details.
I think I'll try my initial idea and use the Ethernet input and if it's not worse than the Amanero I can at least use native DSD in linux since I'm not very confident he can fix that (it's been several months of almost done).
And maybe get a Pulsar at some point to replace the master XO of the 004.

 

[penguin69]

I use isolation transformer + tracking filter.
https://powervar.com/standard-power-conditioners/
http://www.digitaltele.com/pdf/emerson/islatrol_lric_ic_plus.pdf

Hi,

FWIW, I use the following for power conditioning/regeneration:

http://www.powerinspired.com/store/ag1500-ac-regenerator-1500w-psu-p-1742.html

It seems to keep the listening experience consistent, however I'm not knowledgeable enough to know whether using this type of product can bring with it trade-offs, e.g. loss of dynamics.

Anyway, on paper it seems to do a similar job to far more expensive 'audiophile' equivalents.

Looks like the price has gone up a little recently ... I managed to pick up 2 units in the past for 800GBP total.

There is also a lower wattage unit, that retails for 275GBP.

These units are fan-cooled, but the fan is not too noisy and can be cranked down during cooler months to the point where it is inaudible.

 

[vick_v]

Hi,

FWIW, I use the following for power conditioning/regeneration:

http://www.powerinspired.com/store/ag1500-ac-regenerator-1500w-psu-p-1742.html

It seems to keep the listening experience consistent, however I'm not knowledgeable enough to know whether using this type of product can bring with it trade-offs, e.g. loss of dynamics.

Anyway, on paper it seems to do a similar job to far more expensive 'audiophile' equivalents.

Looks like the price has gone up a little recently ... I managed to pick up 2 units in the past for 800GBP total.

There is also a lower wattage unit, that retails for 275GBP.

These units are fan-cooled, but the fan is not too noisy and can be cranked down during cooler months to the point where it is inaudible.

I got the 500W device too and while I can't tell if it has any effect I do like the fact of hopefully clean power going to all my devices.

 

[mb3]

The USB-I2S interfaces I use needs about 10W, so the relays will need to switch at least 3A@3.3V, the Takamisawa is not rated for this.

Gold has a low boiling point and will evaporate in the arc instead of oxidizing, I would use relays with Palladium or Indium alloys in the contacts minus the gold plating.The Takamisawa is using a Silver-Palladium alloy under the plating.

Gold plated relays like this one should only be used for switching low level circuits like DAC audio outputs.

TE makes a lot of relays and they have a good write up on relay contact materials, at the end of the article is some info on how to calculate the snubber network, aka the "wet" part.
I see this as a complication when using relays.
http://www.te.com/commerce/Document...v&DocNm=13C3236_AppNote&DocType=CS&DocLang=EN

I see designs using SSRs and I would not use these designs.

MOSFETS switches are another matter, for example the SSM6J0503N from Toshiba has an on resistance of 32.4 mOhm @ 3A, 4.5V
This is about 2/3 the contact resistance of the Takamisawa (50mOhm) and it carries a higher current without the contact wear concerns. Digikey sells them.

https://www.digikey.com/product-det...e/SSM6J503NU,LF(T/SSM6J503NULF(TCT-ND/3522397

The link below show the kinds of features I am interested in, namely microprocessor controlled transistor bank switching and an integrated charging circuit.
https://uptoneaudio.com/products/ultracap-lps-1

Just waiting for the rest of the industry to catch up.

LKS configured the 9038 to re-clock all incoming streams, the custom re-clock logic that you will be paying for on the SU-1 largely goes to waste.
The cost is within 50USD of the Pulsar OCXO, this replaces the XO in the DAC at the the very centre of the 9038's re-clock logic, I think that mod has better value for the money.

LKS paid little attention to the I2S interfaces, most of the logic is powered by the slow and somewhat noisy LM317, the LM317 is a huge step backward from the LT317 used on the 003.
Take a look at how PSAudio implemented I2S and compare to see how I2S is poorly done on the LKS.

I do not see any major changes in the I2S section between the 003 and 004 and expect the same kinds of problems with the SU-1, namely clicks and sometimes loud bangs when switching sample rates or mode changing between PCM and DSD.

Link below describes how the SU-1 does not play nicely with the LKS 003.
https://www.head-fi.org/f/threads/lks-audio-mh-da003.745032/page-25#post-12927931

The very same seller in your link blamed the LKS and walked away leaving the customer high and dry.
Suggest you check to see if the seller eventually made it right for the customer.

LKS utterly messed up the enhanced USB interface implementation on the 004.

• The USB thing is powered by a secondary winding off the digital transformer, creating a direct path for digital noise from the host + the USB board right into to the DAC. Toroids have very wide bandwidth giving the noise a very low impedance path coupled through the secondary windings, this bypassed the interwinding screen on the transformer rendering it useless.
• LKS did not insulate the USB input from the rest of the DAC metal work, this joined the ground on the host and the DAC completely bypassing the galvanic noise isolator.

This amateurish blunder robbed the 004 of a large chunk of its potential.

The fix is simple, power the USB board with its separate transformer with an interwinding screen between the primary and secondary windings and most importantly insulate the USB input from the rest of the metalwork.

The SQ difference after this change is not subtle.
Improvement on a level similar to the step up in SQ between the 003 and the stock 004



That is expected, the unbalanced output is driven by an fully buffered active filter, this allows it to drive a wide variety of cables without difficulty as the active stage compensates for the cable linearity issues.

The balanced stage has a passive filter after the I/V converter, this makes it highly dependent on the load represented by the impedances of the cables you are using and the input impedance of the power amps.

Balanced can be better if the rest of the audio chain can take advantage of differential drive, all of my amps, headphones and speakers use differential drive.

Not much can be done here except to use low(ish) capacitance balanced cable with linear response in the audio band designed for professional use like the Canare L4-6ES or the Gotham GAC/GAC+ series if you want to spend the money. I would steer clear of the Audiophile stuff for the XLR outputs on the LKS. A preamp in the middle might work depending on quality, I have used a Schiit Mjolnir as preamp with decent results.

LKS decision to use an unbuffered passive XLR output is an unfortunate triumph of audio "purity" over common sense.
The LKS004 is a decently designed unit for most parts but it has quite a few shortfalls when it comes to implementation.

Hi B0bb,
From your experience, would you expect that upgrading to isolated toroidal transformers will help to improve sound quality from this device? - which is already good imo. Thank you.

 

[b0bb]

Hi B0bb,
From your experience, would you expect that upgrading to isolated toroidal transformers will help to improve sound quality from this device? - which is already good imo. Thank you.

Yes but only if the LKS004 grounding and power routing problems are fixed first.

2 sets of transformers are required one for the digital devices such as the media player, computer, CD player etc and another for the analog side such as headphone amps, power and preamps and the DAC.
Only join the common ground and neutral at the power outlet on the wall (if you have access to an outlet on another circuit, that is better but not always practical and requires knowledge of the house wiring).

Keep in mind RFI filters are needed to isolate incoming interference as well as isolating the digital and analog parts of the audio chain from each other. (This is something LKS forgot when they wired in the enhanced USB board).

Avoid plugging the analog and digital devices on the same isolation transformer or AC re-generator, this is one reason why some people do not see a difference when using just 1 big unit for the whole setup.

If you have the enhanced USB board, the LKS 004 internally joins the digital and analog power lines, this must be corrected (separate transformer and insulated USB metal shield) first before any improvement can be seen.
On the standard USB board, the USB metal shield has to be insulated as it can join the digital and analog grounds thru touching the DAC metal work.

 

[Audio4Fun]

Hi, I’m new to the Head-Fi forum, but I had been reading this thread for a very long time, now I had purchased a MH-DA004 that is now burning for about 6 days, first test was not as good as I expected, but had to wait for new tests after 2 weeks burning period.

I want to change the stock fuse for a better quality one, but I found the stock fuse is rated at 3.15A (slow blow type), this seems to me to be a very big value for the MH-DA004 @220v/50Hz, is this correct?

Thanks

 

[penguin69]

Avoid plugging the analog and digital devices on the same isolation transformer or AC re-generator, this is one reason why some people do not see a difference when using just 1 big unit for the whole setup.

This is good to know. Mine are indeed separated in this way - but that is down to luck more than sound judgement!

 

[mb3]

Yes but only if the LKS004 grounding and power routing problems are fixed first.

2 sets of transformers are required one for the digital devices such as the media player, computer, CD player etc and another for the analog side such as headphone amps, power and preamps and the DAC.
Only join the common ground and neutral at the power outlet on the wall (if you have access to an outlet on another circuit, that is better but not always practical and requires knowledge of the house wiring).

Keep in mind RFI filters are needed to isolate incoming interference as well as isolating the digital and analog parts of the audio chain from each other. (This is something LKS forgot when they wired in the enhanced USB board).

Avoid plugging the analog and digital devices on the same isolation transformer or AC re-generator, this is one reason why some people do not see a difference when using just 1 big unit for the whole setup.

If you have the enhanced USB board, the LKS 004 internally joins the digital and analog power lines, this must be corrected (separate transformer and insulated USB metal shield) first before any improvement can be seen.
On the standard USB board, the USB metal shield has to be insulated as it can join the digital and analog grounds thru touching the DAC metal work.

Sounds good - thank you. Insulating the usb input works well. I have my dac and computer powered from an apc line-r 1200va with surge and emi filters in place.
http://www.apc.com/shop/au/en/products/Line-R-1200VA-Automatic-Voltage-Regulator/P-LE1200I
http://www.poweronaustralia.com.au/product/eaton-psfi/
I was thinking that upgrading the power supplies to something like this https://www.plitron.com/default-factory-specs/medical-grade-toroidal-transformers/
would be an improvement - along the lines of the kitsune level 3 upgrade...

 

[b0bb]

APC uses a switchmode type regulator and can be quite noisy, the Eaton RFI filter should be placed between the AVR and the dac.

Stock transformers are already quite good and has the required electrostatic shield between the primary and secondary windings, expect, at best an incremental improvement.
I think powering the enhanced USB interface with its own transformer will give better results and if you do not have the enhanced USB to I2S interface, get it as well as the other I2S inputs on the LKS004 is done poorly done in comparison.

 

[mb3]

APC uses a switchmode type regulator and can be quite noisy, the Eaton RFI filter should be placed between the AVR and the dac.

Stock transformers are already quite good and has the required electrostatic shield between the primary and secondary windings, expect, at best an incremental improvement.
I think powering the enhanced USB interface with its own transformer will give better results and if you do not have the enhanced USB to I2S interface, get it as well as the other I2S inputs on the LKS004 is done poorly done in comparison.

Hi b0bb, that's how I have it set up, avr >eaton>lks. Probably should use a different avr but this works ok. I have the upgraded version with the amanero USB card - which seems to be the best available option for USB input. Thank you for your tips.

 

[kirbyt]

I find all this very interesting read, a question directed to bobb is it plausible to wire the singxer directly into where the amenaro board pins in? Would that be of benefit as apposed to the HDMI i2s input? I'm a bit of a newb to all this but if I'm interpreting your comments correctly the amanero i2s input to board is treated with greater care vs the i2s sockets on the dac.

 

[b0bb]

I find all this very interesting read, a question directed to bobb is it plausible to wire the singxer directly into where the amenaro board pins in? Would that be of benefit as apposed to the HDMI i2s input? I'm a bit of a newb to all this but if I'm interpreting your comments correctly the amanero i2s input to board is treated with greater care vs the i2s sockets on the dac.

Possible in theory but whether it works in practice is a different matter.

There are no commonly agreed standards for I2S ,so basic items like the termination impedances is left up to the manufacturer. A 44.1k stream generates a 2.8MHz bitstream so the transmitter and receiver must be properly terminated.

Similarly, there no clear rules to govern behavior during a mode change, for example from PCM to DSD or a sample rate change. This one is responsible for much of the problems with click and bangs some people have experienced.

LKS matched the internal I2S input to the Amanero board and if you change this you have to re-do all the work they did with the Amanero card.

 

[kirbyt]

Well ok I went and took the questionable power side of the amenaro out and jumped the board 5v input via an lps-1. Seems to be an improvement for me as I really wasn't enjoying the amenaro input and preferred the singxer via i2s up till now.

 

[b0bb]

Well ok I went and took the questionable power side of the amenaro out and jumped the board 5v input via an lps-1. Seems to be an improvement for me as I really wasn't enjoying the amenaro input and preferred the singxer via i2s up till now.

Did you insulate the USB socket from the chassis?
The back panel comes off easily giving access to the USB socket.