Soekris Dam Dac

[nige2000]

well for the test i got 4 cells two pairs, each pair in series float charged to 8v dc
connect first pair of +8 to power on the dac and - to gnd on second pair + 8 goes to gnd on dac and - to power
magic we got plus and minus 8

this voltage is out of the comfort zone for these cells but ok so far

so if you try it dont be surprised if they leak explode etc

[rickmcinnis]

I think I will play it safe with two sets of three.

[randytsuch]

Nigel
So 4V a cell. Seems a little risky. I was just looking at the spec sheet for the batteries I bought, they have a 3.9V max voltage spec, so if your's are similar you're not too far above that.

If that works you don't need to try my idea.

Rick
For my idea, you need two voltages that add up to to 4V, and you put the two voltages in series. In my scheme, one voltage comes from a linear power supply, and the other voltage comes from a LIFEPO4 battery. I can't use two batteries because you can't run LIFEPO4's at 2V each, that is too low. .

So, if you were using a 1.2 Volt linear power supply, then you need to discharge the lifePO4 down to 2.8V. That gives you a 4V to power the DAC.
Make two of these, and connect the 2nd one up with + to DAC ground to get the -4V supply.

Randy

[nige2000]

Does this make enough difference to make you a little happier with the sound as compared to the PCM1794?

rechecked/tested a-b dam vs pcm1794
for detail dam dac is best
for dynamic, imagery, placement, foot tapping, head nodding etc the lifepo4 pcm1794 is best

[nige2000]

well im starting with the easy stuff
took the 3.3v LDO off the board and injected 3.3 from lifepo4



definitely more ease to the sound better cymbals, piano notes, trumpet etc with the 3.3v injected



also replaced the lowpass filter cap to a film cap on one channel too

difference in the caps is a closer call though it does seem slightly cleaner with film i will likely change the other one next time

pics stole from hifiduino

[randytsuch]

well im starting with the easy stuff
took the 3.3v LDO off the board and injected 3.3 from lifepo4

definitely more ease to the sound better cymbals, piano notes, trumpet etc with the 3.3v injected

also replaced the lowpass filter cap to a film cap on one channel too

difference in the caps is a closer call though it does seem slightly cleaner with film i will likely change the other one next time

pics stole from hifiduino

Cool stuff Nigel.

One thing that hifiduino mentioned is that there is only one 3.3V reg on the board (that you have now replaced). Wondering if there would be benefit to separate out so the main devices (FPGA, clock, isolator) have their own 3.3V.

This would harder to implement, probably would need to cut traces to do this. FPGA probably has 3.3V on multiple pins too.

Randy

[nige2000]

i think there is two v outs on that ldo
now i didnt check where the two 3.3v pads were going but i think each of the v out pads were going to separate devices because i had to bridge them
so that might be a place to seperate

i know with lifepo4, isolation and separation can be overrated if the dac components can be powered completely and directly by lifepo4 there will not be any noise worth isolating/seperating for
adding extra cells can test easy enough

you could try some 3.3v shunt regs theres a few available with tent, or ian (fifo) and other places im forgeting

[rickmcinnis]

I agree with nige2000 that with the capacity and the low esr of the A123s that the eed for regs for everything could well be moot.

We can hope this is so until reality shows us we cannot.

Any chance of a photograph of what you have done, O" great one? (said with complete sincerity).

What cap were you able to squeeze into that spot? What value?

Could this be a case where a transformer on the output could replace that filter entirely; just let the transformer perform the low pass - likely at a lower frequency - which would not make any difference. Would it?

Thought I was to receive my board today but it is actually tomorrow

[nige2000]

I agree with nige2000 that with the capacity and the low esr of the A123s that the eed for regs for everything could well be moot.

well on a pc its much different its a very noisy environment the more separate supplies the better

We can hope this is so until reality shows us we cannot.

if i succeed in removing all regs my hope is to only use 3 cells i really doubt i will need more than 3 sources of power to gain optimum preformance

Any chance of a photograph of what you have done, O" great one? (said with complete sincerity).

visual aids is not my strong point my test work is always a big mess because i like to test as i go

What cap were you able to squeeze into that spot? What value?

think its a 0805
the original cap is a 1500pf afaik i put in a 2200pf i will get some 1500pf to check/test

Could this be a case where a transformer on the output could replace that filter entirely; just let the transformer perform the low pass - likely at a lower frequency - which would not make any difference. Would it?

aghh sq difference between a transformer and a good film cap
on the 1794 i tested a film cap versus straight from pos and neg sq wise i wouldnt consider the cap to be much worse

Thought I was to receive my board today but it is actually tomorrow

thats pretty slow

[nige2000]

hmm
http://e-greenmotion.com/index.php?rout ... uct_id=188
1 Model NO. WL-9138190
2 Material Li-ion
3 Typical capacity 8-9AH
4 Nominal Volt 3.7V
5 N.W About 300g
6 Dimension for the cell body: 9x 138X190mm
7 Dimension for the cell with tabs 9x 138X210mm
8 Internal Resistance ≤0.8mΩ
9 Max. Charge Rate 120A
10 Recommend Continuous Discharge Rate 200A
11 Pulse Discharge Rate（3 seconds） 400A
12 Charge voltage 4.2V
13 Discharge cut off voltage 2.5V
14 Working Charge Temp 0~50°C
15 Working Discharge Temp –20°C~50°C
16 Storage Temperature –20°C~50°C