Recent Posts

61

Transports / Re: Raspberry PI transport with I2S output into TDA1541A

« Last post by dannydigital on December 14, 2019, 02:39:59 PM »

You are not quite understanding how the whole thing works.

As I pointed out in last post, the 'old' JLsounds board (that you have) reclocks all I2S signals with a discrete flip flop chip. From memory it was LVC (but can't recall exactly).

So in fact your I2S lines to DAC have only 1 LVC logic chip between clock and DAC. That's why it sounds so good.  
Given this, the MUX switching chip could well make quite a difference. The only way to know is try.
 
The new board I2S has reclocking inside FPGA chip. I believe they take great care to keep jitter low but it won't be as good as a discrete FF logic chip.

Terry,

I understand exactly what you are saying wrt the differences implemented in the clocking o/p's of the two boards.
I'm not using the old board in this project.
Saving it for something else.

Ideally it would be best to have the USB board mounted right above the TDA1541 and extend the USB port to the rear panel instead but this is a retrofit project and I'm taking into consideration some practicality with the fitout.
Time to enjoy the results, using foobar2000remote on my phone. This is brilliant, it gives you full control of the system and allows full access in to the PC's directory structure. Can play whatever you have on your laptop, maybe network drives too, I will have to check this.

Fun fun fun!

62

Transports / Re: Raspberry PI transport with I2S output into TDA1541A

« Last post by dannydigital on December 14, 2019, 02:14:17 PM »

Well the older I2SoverUSB board won't do rates past 96kHz with the TDA1541. I asked Lyuben about this and his reply was;
 
"For a pity I don't have a PC with windows 7 anymore. As you know on our old board you can use it only as I2S to TDA1541. Your configuration is correct of TDA1541.
Can you install new Thesycon driver and adjust the buffer to be bigger?"

I set up a new WIN10 system and tried his recommendation as above plus tinkered with some other ideas but to no avail. When I try to input 176.4kHz or 192kHz signals it's noise. At 176.4 you can make out the music but it's heavily distorted.
I think it's a firmware issue in his module which could be fixed to support the higher rates but obviously Lyuben wouldn't be concerned about that. It doesn't really matter anyway as for the TDA1541 I think it's good enough.
Athough it would be nice for it to work to it's full potential since it can only operate in multiplexed mode 192kHz being the max.

63

Transports / Re: Raspberry PI transport with I2S output into TDA1541A

« Last post by zenelectro on December 14, 2019, 12:11:24 PM »

Terry,

The switching of SPDIF & USB feature was implemented purely for convenience as some would still prefer for their DAC to be compatible with legacy SPDIF devices, so that's my thought about that.

Yes, that's right. My point was by having spdif compatibility, you give away some amount of performance in USB mode. The difficult thing with
all this clocking / jitter biz is it's hard to get a perspective on what really matters. We are for the most part just listening which gets really tricky
with jitter. It would be great to have a phase noise test set but that's a lot of money    

In regards to the logic type used, I used HCT series which is what I had available, sure ACT would have been faster but I wasn't going to fuss over a single logic chip when compared with the digital signal traveling through LSI IC's like the Xmos > Xilinx which contains zillons of them.

You are not quite understanding how the whole thing works.

As I pointed out in last post, the 'old' JLsounds board (that you have) reclocks all I2S signals with a discrete flip flop chip. From memory it was LVC (but can't recall exactly).

So in fact your I2S lines to DAC have only 1 LVC logic chip between clock and DAC. That's why it sounds so good.  
Given this, the MUX switching chip could well make quite a difference. The only way to know is try.
 
The new board I2S has reclocking inside FPGA chip. I believe they take great care to keep jitter low but it won't be as good as a discrete FF logic chip.

With regards to the USB bridge, I am using the latest I2SoverUSB v.III board from Lyuben with the Xilinx re-clocking. (I ordered them just a month or two ago, they're a lot more versatile.)

http://jlsounds.com/products.html

I have one of the old boards too which I'm planing for another project.

I've only been able to get the old board working up to 96kHz with the TDA1541. Have you tinkered with these older ones? (using the  TDA1541)

Yes, I've had new and old boards working fine with TDA1541.

You should be able to get 192k / I2S from both and 384k / simultaneous from the new one.
Yep the new one is very flexible for just about any DAC out there.

My aim is to get a PCM1792 running native DSD256 (11MHz) with new board. 

All good fun.

TCD

64

Transports / Re: Raspberry PI transport with I2S output into TDA1541A

« Last post by dannydigital on December 14, 2019, 01:29:11 AM »

If you are switching I2S between spdif receiver chip and USB bridge, the USB bridge will have way lower jitter, especially when it is isolated and will sound better IME.

As such I would be optimizing the setup for USB bridge as the spdif will never be as good. My recommendation is to mount the MUX chip close to the USB bridge
so you can get good grounds. Every cm of wire will make a difference.

You have specced 74HC157 for MUX (switching) but for lowest jitter 74ACT157 (AC logic) or possibly even better SN74LVC157A (LVC logic) is better.
If you check the data sheet the prop delay on ACT is less than 1/2 that of the HC and the LVC is even lower. LVC is recommended for 3.6V operation but
absolute max is 6.5V so should be OK at 5V.

I have found that these USB bridges are very finicky with sound so everything will matter.
As far as silver cable blah blah.... all the other stuff matters much more.

Now the last (and worst part) of the story:

The JLsounds USB bridge you have there is an old model which have since been updated.
The older model has discrete (Flip Flop) re clocking of all I2S signals which is really good. The newer on does I2S re-clocking inside
what looks like FPGA chip and only master clock OP comes directly from actual clock. This is OK for new Delta Sigma DAC's because they
use Mclk for timing reference.

Terry,

The switching of SPDIF & USB feature was implemented purely for convenience as some would still prefer for their DAC to be compatible with legacy SPDIF devices, so that's my thought about that.

In regards to the logic type used, I used HCT series which is what I had available, sure ACT would have been faster but I wasn't going to fuss over a single logic chip when compared with the digital signal traveling through LSI IC's like the Xmos > Xilinx which contains zillons of them.

With regards to the USB bridge, I am using the latest I2SoverUSB v.III board from Lyuben with the Xilinx re-clocking. (I ordered them just a month or two ago, they're a lot more versatile.)

http://jlsounds.com/products.html

I have one of the old boards too which I'm planing for another project.

I've only been able to get the old board working up to 96kHz with the TDA1541. Have you tinkered with these older ones? (using the  TDA1541)

65

Transports / Re: Raspberry PI transport with I2S output into TDA1541A

« Last post by zenelectro on December 12, 2019, 11:06:46 AM »

Had SV, KLE and CC around today for a listen.

Rolled the output tubes on the KD,, played with power cables and IC's,, added some some super tweeters,,,,,

The raspberry pi wifi transport communicated all changes transparently.

Usb is redundant. 

Was lovely having you guys over this arvo.   Steps forward were made.

Yeah sounds good to me Oz.

At the end of the day it doesn't matter, R Pi or USB are both essentially asynchronous connections with a fixed
clock on the DAC end and that's what you want.

Then it just becomes a question of how it's implemented WRT stuff like grounding / isolation / reclocking etc etc etc.

I'll have to have a more detailed look at the R Pi / ethernet solutions by Allo, they look pretty well done.

The irony of all this is many Pro Audio facilities still rely on SPDIF / AES, so DAC's are still using this broken system.

TCD
 

66

Transports / Re: Raspberry PI transport with I2S output into TDA1541A

« Last post by ozmillsy on December 12, 2019, 12:41:15 AM »

Had SV, KLE and CC around today for a listen.

Rolled the output tubes on the KD,, played with power cables and IC's,, added some some super tweeters,,,,,

The raspberry pi wifi transport communicated all changes transparently.

Usb is redundant. 

Was lovely having you guys over this arvo.   Steps forward were made.   

67

Transports / Re: Raspberry PI transport with I2S output into TDA1541A

« Last post by zenelectro on December 11, 2019, 10:01:40 PM »

Installed & wired up

Single core copper wire power cables made up from high grade CAT6a data cable. (23AWG)

For the I2S bus, mil-spec OFC silver plated, Teflon insulated cabling is used. (22AWG)

The USBtoI2S module is implemented in completely isolated mode with two independent low noise regulated power supplies for the USB side & re-clocking side which are fed by a potted toroidal transformer from two (2 x 5VA) separate secondary windings. This provides better noise immunity from the computer environment that's fed on the USB side allowing for a cleaner re-clocked I2S bus feeding into the TDA1541A.

Danny,

I have no interest in TDA1541 any more (or zenclock) so if I may make some suggestions:

If you are switching I2S between spdif receiver chip and USB bridge, the USB bridge will have way lower jitter, especially when it is isolated and will sound better IME.

As such I would be optimizing the setup for USB bridge as the spdif will never be as good. My recommendation is to mount the MUX chip close to the USB bridge
so you can get good grounds. Every cm of wire will make a difference.

You have specced 74HC157 for MUX (switching) but for lowest jitter 74ACT157 (AC logic) or possibly even better SN74LVC157A (LVC logic) is better.
If you check the data sheet the prop delay on ACT is less than 1/2 that of the HC and the LVC is even lower. LVC is recommended for 3.6V operation but
absolute max is 6.5V so should be OK at 5V.

I have found that these USB bridges are very finicky with sound so everything will matter.
As far as silver cable blah blah.... all the other stuff matters much more.

Now the last (and worst part) of the story:

The JLsounds USB bridge you have there is an old model which have since been updated.
The older model has discrete (Flip Flop) re clocking of all I2S signals which is really good. The newer on does I2S re-clocking inside
what looks like FPGA chip and only master clock OP comes directly from actual clock. This is OK for new Delta Sigma DAC's because they
use Mclk for timing reference.

Older multibit DACs like 1541 and 1704 use I2S bitclock for timing reference so I feel the newer USB bridge is possibly not as good for
old multi bit DACs.

I have 2 PDX DAC's on my bench ATM, both have been upgraded with JLsounds USB bridges. One has the older one (like yours)
and the other has a MK3 newer version. I have to say they definitely sound different. They newer one is cleaner and more clinical
and the older one is more musical and ultimately more listenable.

I'm currently playing with power supplies and other components etc etc to see if I can get the newer one to sound more musical
and engaging. It's very frustrating. The old JLsounds bridge was *Killer* and he went and changed it.

Anyway, good luck.

Terry

68

Transports / Re: Raspberry PI transport with I2S output into TDA1541A

« Last post by dannydigital on December 09, 2019, 01:39:53 PM »

Installed & wired up

Single core copper wire power cables made up from high grade CAT6a data cable. (23AWG)

For the I2S bus, mil-spec OFC silver plated, Teflon insulated cabling is used. (22AWG)

The USBtoI2S module is implemented in completely isolated mode with two independent low noise regulated power supplies for the USB side & re-clocking side which are fed by a potted toroidal transformer from two (2 x 5VA) separate secondary windings. This provides better noise immunity from the computer environment that's fed on the USB side allowing for a cleaner re-clocked I2S bus feeding into the TDA1541A.

69

Transports / Re: Raspberry PI transport with I2S output into TDA1541A

« Last post by dannydigital on December 09, 2019, 01:29:18 PM »

KillerDACtoUSB  I2S Logic switching completed and working nicely.

Quad 2-input multiplexer 74HCT157 logic is used to switch I2S modes between COAXIAL/USB inputs. The spare 4th bit in the mux chip is used to indicate the switched state by an LED in the physical selector switch.

The logic chip also serves as a compatibility interface between the USB module, a 3.3V device and the TDA1541A which is a 5V device.
(Ensures reliability between active high & active low states in the transferred data.)

70

Transports / Re: Raspberry PI transport with I2S output into TDA1541A

« Last post by zenelectro on November 25, 2019, 06:51:52 PM »

I often wonder if anyone has ever attempted to build a semi-descrete DAC (no DAC chip) simply just using cmos logic, memory and analogue ccts on a large pcb. You could trim any part of the cct for whatever parameter you need at every stage.(Such as linearity)
This would be something different.
D

Lots of people have done it.

What sort of DAC do you want?

There are tons of discrete R2R DAC's, single ended and balanced.
There is one discrete, segmented R2R + unity weighted DAC (like AD1865 etc) being done that I know of.
There are discrete (1 bit) DSD DACs with 16 or 32 unity weighted OP bits arranged as an (averaging) FIR filter
There is another guy I know doing (1 bit) DSD differential current source DAC with various number of unity weighted elements

It's interesting, when you look at all these various approaches, they all have advantages and disadvantages.
To get any of them to work really well requires a very broad knowledge from PCB design to intricate details of logic families, clocking etc etc.

And then there are aa myriad of chips out there that can be used in so many different ways.

That's just the DAC - you have to get all the bits into the DAC without messing up.

So many options.

T