Not bored with it but as a result of not hearing a huge amount of difference between it and my reference DAC (the one with a capacitor tower under it) I upgraded the amps in my active speaker setup. Now differences are more apparent so I'm thinking of pulling out half the TDA1387s as I reckon 8 is probably overkill.... I've ordered more but the supplier told my ordering contact that he can't get the cases until this weekend.
No the on-board trafos don't have 110V taps. That adapter will certainly work as its 100W and the DAC's probably under 10W but I'd be rather concerned with the potential for switching noise. As its so small I'm guessing it has some kind of switching supply inside.
No the on-board trafos don't have 110V taps. That adapter will certainly work as its 100W and the DAC's probably under 10W but I'd be rather concerned with the potential for switching noise. As its so small I'm guessing it has some kind of switching supply inside.
I've gone for separate bypass on all pin7s individually. Use as large a cap as you can fit in.
<edit> My current state of mods now - I pulled out 4 of the TDA1387s because I wanted less stress on the power supply for my AD844 output buffers and less DAC generated LF noise on the DAC supply. The total DAC output current (now 4.8mA) has to be sourced by the AD844 - the feedback resistor has increased from 620ohm to 1k2. I've put more caps on its supplies too. Having only 4 DAC chips means more space for the pin7 caps - now I have 3300uF per chip. Since I lowered the AD844's supply it doesn't have enough output swing to absorb the DAC's output offset so I've added two two-transistor CCSs to bring the offset down to around 100mV.
The overall result is that its sounding closer to my Ozone Pagoda again. The main weakness seems to be in the bass, so more caps have to go on the DAC supply.
<edit> My current state of mods now - I pulled out 4 of the TDA1387s because I wanted less stress on the power supply for my AD844 output buffers and less DAC generated LF noise on the DAC supply. The total DAC output current (now 4.8mA) has to be sourced by the AD844 - the feedback resistor has increased from 620ohm to 1k2. I've put more caps on its supplies too. Having only 4 DAC chips means more space for the pin7 caps - now I have 3300uF per chip. Since I lowered the AD844's supply it doesn't have enough output swing to absorb the DAC's output offset so I've added two two-transistor CCSs to bring the offset down to around 100mV.
The overall result is that its sounding closer to my Ozone Pagoda again. The main weakness seems to be in the bass, so more caps have to go on the DAC supply.
Last edited:
Hi Richard,
I we want to go active IV conv what do you think should do?
the jfet IV ala fetishzator / lukas fikus IV by using bf245 or opamp use opa2604, or still the IV resistor do best for this dac chip?
I read somewhere that the IV R would do good by infinite impedance, and op amp will help the R IV to face the infinite impedance.
Please suggest
many thanks in advance
I we want to go active IV conv what do you think should do?
the jfet IV ala fetishzator / lukas fikus IV by using bf245 or opamp use opa2604, or still the IV resistor do best for this dac chip?
I read somewhere that the IV R would do good by infinite impedance, and op amp will help the R IV to face the infinite impedance.
Please suggest
many thanks in advance
Its a good question and a bit early to tell, I shall do more listening.
My current hypothesis is that its the DAC generating LF noise onto the rails (modulating its supply current) that requires so many shunt caps to get ultimate bass quality. Thus more DAC chips are going to create more noise - unless the noise they create can be decorrelated somehow. So I reckon that each chip needs a certain amount of rail C - more chips calls for more C.
There's also the issue of the output stage - more current needs to be supplied from the AD844 output to balance the DAC's current to its input, resulting in more sag on the rails. Hence again I reckon the output stage may well perform better (better dynamics) with lower input current (=fewer DACs). But this is all conjecture for now, more experiments are needed.
My current hypothesis is that its the DAC generating LF noise onto the rails (modulating its supply current) that requires so many shunt caps to get ultimate bass quality. Thus more DAC chips are going to create more noise - unless the noise they create can be decorrelated somehow. So I reckon that each chip needs a certain amount of rail C - more chips calls for more C.
There's also the issue of the output stage - more current needs to be supplied from the AD844 output to balance the DAC's current to its input, resulting in more sag on the rails. Hence again I reckon the output stage may well perform better (better dynamics) with lower input current (=fewer DACs). But this is all conjecture for now, more experiments are needed.
Been there, done that ...
Uh... this was me in 2006...
... parallel-processin' ... 8x1545A, then 4x, then 2x ... all in a diyparadise Monica mod ... added the discrete Rudolf I/V stage later which improved things a bit.
Never cared for parallel as much as single. W/o the Rudolf stage, I'd have salvaged the Monica... that I/V saved it from the recycler ... but, alas, it sits in a "museum" somewhere, unused for years.
Uh... this was me in 2006...
An externally hosted image should be here but it was not working when we last tested it.
... parallel-processin' ... 8x1545A, then 4x, then 2x ... all in a diyparadise Monica mod ... added the discrete Rudolf I/V stage later which improved things a bit.
Never cared for parallel as much as single. W/o the Rudolf stage, I'd have salvaged the Monica... that I/V saved it from the recycler ... but, alas, it sits in a "museum" somewhere, unused for years.
So it looks like my tda1387 x8 might actually be on its way. Given the overhead of shipping plus Taobao agent commissions, I ordered both the fully assembled unit and the DIY version (empty board + bag of components).
First order of business will be to actually power the thing up, and it looks like it ships with 220v support; I'm on 110 here in the USA. Would one of you guys be willing to see if there's a model number on the stock 220v transformers? Hopefully I can track down a datasheet and find something equivalent in a 110 variety. Maybe I can order those in advance of receiving the unit to avoid yet another delay.
I'm trying to decide if I should start by hacking the pre-built unit, or incorporate the hacks as I build up the other one... maybe best to start with something that works, so if it stops working, I know it's my fault?
If nothing else, I needed an excuse to bump this thread.
First order of business will be to actually power the thing up, and it looks like it ships with 220v support; I'm on 110 here in the USA. Would one of you guys be willing to see if there's a model number on the stock 220v transformers? Hopefully I can track down a datasheet and find something equivalent in a 110 variety. Maybe I can order those in advance of receiving the unit to avoid yet another delay.
I'm trying to decide if I should start by hacking the pre-built unit, or incorporate the hacks as I build up the other one... maybe best to start with something that works, so if it stops working, I know it's my fault?
If nothing else, I needed an excuse to bump this thread.
Happy to see you're still moving forward with this DAC
I've got two new PCBs (only) with this - as the seller said the delay was due to inability to get cases I asked for stuffed boards minus the cases. I just pulled out a couple of the res caps to see the details of the trafos.
They come from this manufacturer - 5VA power transformer PE3518D-I - Dechang Electronics that shows the physical dimensions. The digital trafo is 7.5+7.5VAC (two separate windings) and the analog one is 15+15VAC. Both are 5VA.
How about one of these rather than pulling out your trafos - http://www.ebay.com/itm/Goldsource-STU-200-W-Watt-Step-Up-Down-110V-220V-Voltage-Converter-Transformer/390915984585?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D1%26asc%3D20131003132420%26meid%3D8b16c1d411354ea6ae214394b46beb63%26pid%3D100005%26rk%3D2%26rkt%3D6%26mehot%3Dpp%26sd%3D160733095687&rt=nc
I've got two new PCBs (only) with this - as the seller said the delay was due to inability to get cases I asked for stuffed boards minus the cases. I just pulled out a couple of the res caps to see the details of the trafos.
They come from this manufacturer - 5VA power transformer PE3518D-I - Dechang Electronics that shows the physical dimensions. The digital trafo is 7.5+7.5VAC (two separate windings) and the analog one is 15+15VAC. Both are 5VA.
How about one of these rather than pulling out your trafos - http://www.ebay.com/itm/Goldsource-STU-200-W-Watt-Step-Up-Down-110V-220V-Voltage-Converter-Transformer/390915984585?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D1%26asc%3D20131003132420%26meid%3D8b16c1d411354ea6ae214394b46beb63%26pid%3D100005%26rk%3D2%26rkt%3D6%26mehot%3Dpp%26sd%3D160733095687&rt=nc
Last edited:
hello. the static noise still there. only in right channel, though. and the static increase decrease along with music. much like some method of oscillation. where can oscillation cause this noise? 1387? ad744 i/v? i think 1387 because noise always right channel, even when i/v channel swapped.
Alright, finally got it! Now I'm waiting on a 220-to-110 transformer so I can actually test it out. In the meantime, what do you guys think is the best ordering of mods? I read Abraxalito's list on his blog---is there any meaning to the ordering? Basically looking to start with the biggest bang-for-buck mods.
Do the stock opamps definitely need to go? Or should I give them some time before I chuck 'em?
Do the stock opamps definitely need to go? Or should I give them some time before I chuck 'em?
No particular meaning in the ordering - perhaps I should re-write it in the order of bang for the buck. I'd say top three mods are:
1) either add the 22pF cap (between pins5 and 8) to make the 5534s stable as I/V or swap them out for a unity gain stable amp. Remove the feedback cap.
2) improve decoupling on this opamp - if you keep the current 15V rails you'll need 35V rated caps - find the lowest ESR you can get (Rubycon ZLH; Nichicon PW,HW; NCC KY, KZE)
3) swap out the TDA1387 pin7 decoupling caps (currently ceramics) for 1000uF electrolytics.
1) either add the 22pF cap (between pins5 and 8) to make the 5534s stable as I/V or swap them out for a unity gain stable amp. Remove the feedback cap.
2) improve decoupling on this opamp - if you keep the current 15V rails you'll need 35V rated caps - find the lowest ESR you can get (Rubycon ZLH; Nichicon PW,HW; NCC KY, KZE)
3) swap out the TDA1387 pin7 decoupling caps (currently ceramics) for 1000uF electrolytics.