Nice, saved for future reference.
Hello, your answer is in this post (only 4 pages back
)Read the following next posts as well for schematic examples.
@Zbunjen
have you solved that crackling noise yet?
that crackling is from i2s region, so don't go anywhere
just pay attention to this region, the chip itself (no broken) and the i2s isolator
the only important to this region is the power supply, must use a very2 clean psu for this isolator, especially transformer must be very good condition.
or, you can try to remove that isolator, and just use a 100r resistors between the i2s input and output, if your tda1541 chip is good condition then this setting will be okay.
have you solved that crackling noise yet?
that crackling is from i2s region, so don't go anywhere
just pay attention to this region, the chip itself (no broken) and the i2s isolator
the only important to this region is the power supply, must use a very2 clean psu for this isolator, especially transformer must be very good condition.
or, you can try to remove that isolator, and just use a 100r resistors between the i2s input and output, if your tda1541 chip is good condition then this setting will be okay.
Hello, thank you for your comment, no i haven't solved it. Due to time constraints and other projects i now just consider this logic part of the board broken and to be redone in another version. Xaled will send his signed magnitude converter and we'll see if it works, then i'll design new board (i need to do this anyway for some minor improvements).
But yes, the TDA1541's work, i am listening to them with outputs in parallel in regular simultaneous mode, nice sound. I've also tested I2S mode and that's fine too. So the error is somewhere in the i2s - signed magnitude logic yes.
Im not sure what you mean by isolator or transformer though. AC mains transformer is very far from circuit, with many thousand uF of filtration and jung/didden regulators after it, and on the board there is good local decoupling with C0G and tantalum caps close to the IC's.
Power Supply idea: low ESR LTO batteries + parallel Ultracapacitors
I am pleased with the performance of a relay-controlled float charge system managing parallel LTO batteries and Maxwell Ultracapacitors I worked out for my DIY TDA1541a board and thought it might be of interest to other enthusiasts of that chip.
Of course, a primary motive for this is a low noise level that would be more difficult using mains power.
Also, as I am sure you know, both the LTO batteries and the Ultracapacitors (AKA Supercapacitors) have low ESR = very good transient response, better enabled with this setup that places them immediately adjacent to the board with short wires and no relays or switching transistors in between, all that being handled on the other side.
Cc cv buck converters allow safe, controlled charge up of the Ultracapacitors, which I do before connecting the batteries to avoid an overly exuberant current rush party. (But that would make a good name for a band)
details in this thread.
And thank you to @ecdesigns and everyone else for this thread which has been a phenomenal resource informing my next build for that chip… which I expect will be a signed magnitude board.
cheers,
Chris
I am pleased with the performance of a relay-controlled float charge system managing parallel LTO batteries and Maxwell Ultracapacitors I worked out for my DIY TDA1541a board and thought it might be of interest to other enthusiasts of that chip.
Of course, a primary motive for this is a low noise level that would be more difficult using mains power.
Also, as I am sure you know, both the LTO batteries and the Ultracapacitors (AKA Supercapacitors) have low ESR = very good transient response, better enabled with this setup that places them immediately adjacent to the board with short wires and no relays or switching transistors in between, all that being handled on the other side.
Cc cv buck converters allow safe, controlled charge up of the Ultracapacitors, which I do before connecting the batteries to avoid an overly exuberant current rush party. (But that would make a good name for a band)
details in this thread.
And thank you to @ecdesigns and everyone else for this thread which has been a phenomenal resource informing my next build for that chip… which I expect will be a signed magnitude board.
cheers,
Chris
Last edited:
Hi Chris
I too have had very good results with supercaps on the +-5v supply to TDA1541a. I can imagine you have excellent results with your battery supercap approach.
I just posted before and after sound clips on Ryan's D3 thread. Post 1177 https://www.diyaudio.com/community/threads/diy-tda1541a-pcb-d3.328060/page-59#post-7009659
Thank you, @xaled.
I confess I was under the mistaken impression there was a signed magnitude board or two in operation, with maybe just a few kinks like chip output matching to work out… my bad for inconclusive thread skimming.
I will review the key posts in more detail … post 7092 was my intended reference for throwing together a Manhattan style board to start. If there is any wiggle room on the glue logic side, a part of me is inclined to dive into FPGA programming (new to me but I can learn) then maybe pivot to discrete to reduce noise, if need be, once all is worked out.
Excellent. I will have a listen with some good headphones to minimize the internet SQ loss… I suspect I will hear a significant difference, based on my experience with battery and Supercapacitor power supplies…my first listen on my AD1865 DAC with Maxwell Ultracaps on the 3 5V supplies was a revelation in, ummm, soundstage placement definition and depth might be the best description.
I would be great if you could validate the schematics from post 7092 in parallel to Zbunjen implementation attempts using his own board and my prototype PCB. I have some unpopulated PCBs and can send you one if interested.
From what I've understood FPGA implementation can not reach the quality of discrete logic due to intrinsic noise issues. However it would be better then nothing if you would go this way.
Welcome to the club JCMcNeil.
I too almost cant believe that ecdesigns put this much effort into presenting and developing it in the thread and that to this date nobody built the latest version of signed magnitude (and its been how long since #7092?) . Even the previous version there's only about one or two people who built it as far as i know. If nothing else i think we at least owe it to John
I will be testing xaled's board soon, and if it works then we can say there is an operational board. Im also developing a fully integrated solution.
I too almost cant believe that ecdesigns put this much effort into presenting and developing it in the thread and that to this date nobody built the latest version of signed magnitude (and its been how long since #7092?) . Even the previous version there's only about one or two people who built it as far as i know. If nothing else i think we at least owe it to John
I will be testing xaled's board soon, and if it works then we can say there is an operational board. Im also developing a fully integrated solution.
The last few pages show how i was struggling with the signed magnitude board, but it is not as helpful as it could be as we didn't reach any conclusion why it didnt work (which was my fault for not being better equipped for troubleshooting), so the journey continues.
Ryanj captured most of John's innovations on his Distinction3 or D3 board. He mentioned he has a few left but will not be making another run. Its a fine 4 layer pcb with the attenuation circuits, the 100Hz DEM etc. It's a tough thing to solder, but those that have done it like the results.
That's what im developing currently. Using 2x tda1541A in signed magnitude mode (17 bits, no zero crossing glitch current) or simultaneous mode. Or you can use just one if you want to (in simultaneous mode).
D3 is tried and tested but is only for single tda and is about where this thread was 5 years ago. Up to you..
Looking at the relevant Burr Brown datasheets would suggest that the problem might be that you are trying to bridge a 6ft gap with a 5ft plank.
I have an NEC CD730 which uses 2 x 1541A in balanced mode. I like the overall sound but not the complex maze of 8x sampling and DF it employs. Other CDPs I have are Creek CD60 with single crown S1 1541A sounds overall better, more analog, but doesn't have the low end control of the NEC so there are clearly other implementation factors going on here. Sine magnitude is interesting and a possible path to where I'd like to go, hopefully to get close to the resolution of the AKM used in my RME ADI pro 2 FSR. I believe it has more to do with clock and filtering than the DAC chip itself. I want the NOS analog sound with the low end of balanced 1541As.
I see, that's nice. No oversampling will be implemented in hardware (no dedicated IC for it is what i mean) for this board which is a no brainer as if you want you can do digital oversampling from the I2S source. A lot of people here, including myself are interested in signed magnitude. It is not for no reason that this thread is the hub of development for this dac, internet wide.
I suppose it is something you will have to see for yourself on your system and ears, no matter which board you go with.
If you simply want to move away from bipolar zero in smaller steps rather than one large jump then look at the BB PCM1702 datasheet under Advanced Sign Magnitude.Sorry...relevant Burr Brown datasheets? You must be playing 4D chess here or im missing something
At any rate, you probably did think of something and im courious, so if you would be so kind...
Received a nice package from xaled today, containing his implementation of the signed magnitude schematic. I will test it as soon as possible and report back what the situation is. He was so generous to even include some additional goodies that will help.
Looking at the datasheet it seems to me it is in agreement with the signed magnitude employed here. John describes it here also.
At any rate, it's not why its not working, as we should at least get sound.
Ah ok, now i understand what you were getting at
Looking at the datasheet it seems to me it is in agreement with the signed magnitude employed here. John describes it here also.
At any rate, it's not why its not working, as we should at least get sound.