.
Since the current discussion is around PS decoupling and also noting that there is a bit of disdain here for audiophile electronic designers, I thought the job on
my bench was ironic. I'm looking at various channel strips of an English mix console, I guess you could call it a semi-pro, 32 ch -> 8 sub grp -> 2 master
section in for upgrade.
Inspecting the subgroup card, all 9 opamps are JRC TL072's. Looking for PS decoupling caps there is total of *1 per rail, 22 uF - that's it.
These cards include a 12-LED level meter driven by 3 x LM339 comparators powered off the +-17V audio supplies.
Those 2 caps are in the LED switching section and there are no PS caps anywhere to be seen near the OPA's
So... lot's of work to do.
TCD
my bench was ironic. I'm looking at various channel strips of an English mix console, I guess you could call it a semi-pro, 32 ch -> 8 sub grp -> 2 master
section in for upgrade.
Inspecting the subgroup card, all 9 opamps are JRC TL072's. Looking for PS decoupling caps there is total of *1 per rail, 22 uF - that's it.
These cards include a 12-LED level meter driven by 3 x LM339 comparators powered off the +-17V audio supplies.
Those 2 caps are in the LED switching section and there are no PS caps anywhere to be seen near the OPA's
So... lot's of work to do.
TCD
TL072 has 3MHz GBW
Ironically it will probably perform better with less caps to inject power supply noise into the ground traces...
You may think so but that is not the case, it just has to be done carefully.
All under control.
TCD
When a business is factored in, that's a given. As the saying goes, business is business.If nothing else, I admire your persistence and enthusiasm.
Been thinking some more about BD34301EKV and the eval board design.
Interesting that there are so many different pins, or sets pins, labeled with AVCC in the name. Why? Since the dac is segmented architecture maybe the three most significant bits are thermometer coded, and the lower order bits are binary coded (something like that anyway).
https://www.analog.com/media/cn/training-seminars/tutorials/MT-016.pdf
https://www.lumerink.com/courses/ece517/s17/Lecture Notes/DAC Architectures.pdf
In such a case, all the AVCC signals would need to be at the same nominal DC voltage, but at least they could be bypassed separately so that switching transient currents of the few highest order bits do not excessively glitch, distort, or otherwise disturb the lower order bits.
If something more or less like the above were true, then the recommended AVCC decoupling and or possibly the odd PCB layout might make a certain type of sense.
Okay with me so far, but here is a concern: I listened to aluminum polymer and tantalum polymer electrolytics for AK4499 AVCC. They were located at the dac chip AVCC pins. To me they made sort of a low level 'gritty' or 'popping' type of noise. It was different that Panasonic FM (which have their own characteristic noise sound). Other electrolytics such as Silmac don't have that noise effect to me, but they have dynamically flat sound sort of like an audio compressor (lots of good app notes on compressor design at That Corp, if anyone is interested). As said in other posts and threads, to me good film caps are the most benign sounding, but they are bulky and expensive (which brings its own set of problems).
What I have been leading up to is this: Now wondering if the sound of this dac will be limited by the sound of the AVCC bypass caps in a way that might be very difficult to avoid.
Time will tell, I guess.
Interesting that there are so many different pins, or sets pins, labeled with AVCC in the name. Why? Since the dac is segmented architecture maybe the three most significant bits are thermometer coded, and the lower order bits are binary coded (something like that anyway).
https://www.analog.com/media/cn/training-seminars/tutorials/MT-016.pdf
https://www.lumerink.com/courses/ece517/s17/Lecture Notes/DAC Architectures.pdf
In such a case, all the AVCC signals would need to be at the same nominal DC voltage, but at least they could be bypassed separately so that switching transient currents of the few highest order bits do not excessively glitch, distort, or otherwise disturb the lower order bits.
If something more or less like the above were true, then the recommended AVCC decoupling and or possibly the odd PCB layout might make a certain type of sense.
Okay with me so far, but here is a concern: I listened to aluminum polymer and tantalum polymer electrolytics for AK4499 AVCC. They were located at the dac chip AVCC pins. To me they made sort of a low level 'gritty' or 'popping' type of noise. It was different that Panasonic FM (which have their own characteristic noise sound). Other electrolytics such as Silmac don't have that noise effect to me, but they have dynamically flat sound sort of like an audio compressor (lots of good app notes on compressor design at That Corp, if anyone is interested). As said in other posts and threads, to me good film caps are the most benign sounding, but they are bulky and expensive (which brings its own set of problems).
What I have been leading up to is this: Now wondering if the sound of this dac will be limited by the sound of the AVCC bypass caps in a way that might be very difficult to avoid.
Time will tell, I guess.
Last edited:
Last weekend I was able to return back to my board. But unfortunately, I made a mistake connecting the power supplies.
The result was a loss of ~100 dollar chip, and what is the most important - the time, since I had only one
I have to think well before spending another $100, but in any case BD34301 is not available now. Digikey and others said - in August.
P.S. During the previous session, I did not have time for documented measurements, and also the listening time was rather short to make any conclusions.
Most likely for the same reason it has a ton of ground pins: for low inductance and low coupling between pins.
You could probe AVCC to check for noise, please post results. You'll probably get a bit different noise on each pin.
Eval board is starting to sound pretty good playing DSD256. Mosfet output buffer DC blocking cap is slowly opening up. Possibly the organic polymer AVCC caps are clearing up too. Seems to sound best to me so far with AD797s on two-to-one Brown Dog adapters. Don't know if the extra inductance from the adapter and dip socket is helping or hurting in some audible way; maybe some of both. Could be a little filtering of RF from the dac outputs. Eval board is back in the steel server case, running around the clock for maybe another week before further assessment.
Another update on the Rohm eval board. Just added optional Arduino control of the dac chip. Now able to configure the dac chip for DSD256 phase mode operation. Letting it sit and run for awhile. Phase mode allows dsd playback with MCLK = 2x BLCK. The Rohm recommended dsd mode is where MCLK = BCLK. However, I will need to implement a high quality MCLK divider to really know if it makes much difference. At the present time I would say that reverb tail and room ambience reproduction may be a little more audible than with AK4499. Should have a better idea about that after doing some more work with the clock & DSP board that provides I2S and MCLK to the Rohm eval board. Currently using the original prototype board which was later improved somewhat in the next rev. The only populated 2nd rev board is still attached to the AK4499 dac. Thinking about whether to do a third rev just for Rohm, or just populate and slightly mod a 2nd rev board for now. The latter probably wouldn't be quite as good, but maybe good enough for now for some more testing.
Last edited:
TAS has a review of the Luxman CD player up now : Luxman D-10X CD/SACD Player and DAC - The Absolute Sound