bit of a progress report on my test bench
running of 3 lifepo4 battery pack supplies
been float charged by linear supplies
3.3v direct bypass to onboard 3.3v ldo reg
+/-4v direct bypass to vref
new 1.2v ldo reg powered by the 3.3v rail to bypass the onboard switch mode reg
have previously tried
ac power which was horrendous
+/- 12 volt batteries on main power input (better)
3.3v lifepo4 bypass to ldo reg for clock fpga etc (removed ldo) (improved clarity + resolution)
removed output opamps
+/- 7v dc lifepo4 to main power input
dropped to +/-5v main input, removed and bridged 5v regs
2000uf per rail on vref (great improvement in dynamics but a bit boomy at times)
5 f supercaps on vref (twice as good as 2000uf per rail)
direct 4v to vref (hallelujah!!!)
add 1.2v ldo for fpga to bypass sm reg (i did this so i could test vref at 3.3v as the sm reg wont operate under 4v, but this surprised me the sm reg must have been causing a bit of degradation of sq as low level detail was significantly improved)
very simple mods with outstanding results
unfortunately i dont know anyone with the likes of a totaldac to compare with
but i will say (if im allowed 🙂 ) that this dam implementation in its current state is well beyond the capabilities of any sigma delta implementation
running of 3 lifepo4 battery pack supplies
been float charged by linear supplies
3.3v direct bypass to onboard 3.3v ldo reg
+/-4v direct bypass to vref
new 1.2v ldo reg powered by the 3.3v rail to bypass the onboard switch mode reg
have previously tried
ac power which was horrendous
+/- 12 volt batteries on main power input (better)
3.3v lifepo4 bypass to ldo reg for clock fpga etc (removed ldo) (improved clarity + resolution)
removed output opamps
+/- 7v dc lifepo4 to main power input
dropped to +/-5v main input, removed and bridged 5v regs
2000uf per rail on vref (great improvement in dynamics but a bit boomy at times)
5 f supercaps on vref (twice as good as 2000uf per rail)
direct 4v to vref (hallelujah!!!)
add 1.2v ldo for fpga to bypass sm reg (i did this so i could test vref at 3.3v as the sm reg wont operate under 4v, but this surprised me the sm reg must have been causing a bit of degradation of sq as low level detail was significantly improved)
very simple mods with outstanding results
unfortunately i dont know anyone with the likes of a totaldac to compare with
but i will say (if im allowed 🙂 ) that this dam implementation in its current state is well beyond the capabilities of any sigma delta implementation
not the original dac build this is the 3.3v tester similar idea
this is a second build 3.3v version its not my benchmark build but similar idea
its not finished i have to take of some chips thats causing interference its probably the 4v regs
this is a second build 3.3v version its not my benchmark build but similar idea
its not finished i have to take of some chips thats causing interference its probably the 4v regs
Thanks - I t would have been interesting to see the latest creation. Some close ups would be nice.
//
//
I know 2 owners of Totaldac D1 dual... To be fair, comparing against d1 dual should be done with a dual DAM in balanced configuration.
I guess a lot of unexperimented dam owners (like me 😀) are looking for a detailed how to perform bypass mods like yours (specially for the Vref mod).
I was thinking to someting like opc The Wire LT3042 parallel regulators for the direct Vref mod. But looking at the datasheed, I was wondering how important the load transient response is in this application (doesn't look nice in the datasheet, I hope that's better with more caps).
I'd just like to add that I'm hovering over buying one of the latest boards and have been following the threads.
I use a 12v SLA battery for my current ES9023 DAC and it was miles batter than any solid state PSU, even with a choke in it. So the possibility of using batteries would tip me over into buying the board.
Nige - when your "hideous creature" turns into an elegant swan, I'd love to see your (reasonably) final layout and wiring so I could grovellingly copy it slavishly.
I use a 12v SLA battery for my current ES9023 DAC and it was miles batter than any solid state PSU, even with a choke in it. So the possibility of using batteries would tip me over into buying the board.
Nige - when your "hideous creature" turns into an elegant swan, I'd love to see your (reasonably) final layout and wiring so I could grovellingly copy it slavishly.
not sure that would prove anything not already known
i could still do with my 3.3v build as its still stock if the mods are removed
i dont really see the point as ive sigma deltas id rather listen too than a stock revision 1 dam (the revision 2/factory vref mod might change that view as that was a great improvement )
thanks for sharing exp
Almost all mentioned mods were realized and wide tested by me as well
after dozens of listening sessions and comparisons with top-level dacs want to approve your input for "right" amount of vref caps. For sure more uf is better dynamics but everything that more than 5-7uf and less than 25 000 - 40 000 uf are negative affect on tonal balance.
3.999/-3.998v
i knew someone would ask that 🙂
the build was to test this as a possibility i didn't expect it to sound "finished"
all i was hoping for was it to be plausible
if i skew it say 4.0v and -3.6v its difficult to notice any difference sq wise
did i read somewhere that should cause distortion ?
maybe so but its more about been able to compare with dacs in the higher league its superseded anything available to me to test
well using lifepo4 cells as a super duper capacitor is kinda like cheating the test, i guess it will be difficult to better
I reported testing that early on, if I remember correctly it resulted in only 2nd harmonic distortion at 14 db lower than difference in voltage.
So 4.0 / 3.6 Volt would result in 2nd harmonic distortion at -34 db.
I can only say that everything is optimized and tuned for +-4V vref, changing that is not a good idea....