I haven't heard the passive mod.
You can consider this mod
- Replace AD847 with a high quality socket, put LME49990 in the socket.
- Replace the 2x 2200 uF / 35 V Sanyo with 2x 1000 uF / 50 V Elna Silmic II.
I used pure Silver / Tin solder for everything and replaced the little wires to the RCA terminal with Silver as well.
I'm listening to this mod now and the difference in sound quality is noteworthy.
I'd say everything sounds very nice and clean, analog-like, with a weird tonality, sort of like Egyptian or Arabic scales, with a cybernetic feeling as well.
I think female vocals sound a few years younger, after the Elna Silmic II.
Yup, I just lost all credibility I never had from the dScope team.
Quick update: I GOT IT WORKING!
Removed the 2 toppest dac chips (U8 and U9). Needed to remove U9 also to get U8 (the hot one) out. So the one of those chips were faulty.
Thanks Kastor L and abraxalito for helping out.
Removed the 2 toppest dac chips (U8 and U9). Needed to remove U9 also to get U8 (the hot one) out. So the one of those chips were faulty.
Thanks Kastor L and abraxalito for helping out.
Well done!
You just removed them and it started working?
I checked the heat in mine, from back, transformer, to front, the RCA, then DAC chips 1, 4, 5, 6 are the hottest, DAC chips 7 and 8 are the least hot, 2 and 3 are inbetween. 4, 5 and 6 are too hot for my finger if I push down hard, 1 as well, but the heat levels are quite similar overall.
Abraxalito recommends replacing them with TDA1387, I wonder if there are more alternatives?
The sound is pretty nice after my mod with the vanishingly low distortion of LME49990 and Elna Silmic II's, but the total sound is still quite warm and fuzzy, it's certainly lacking in crystal clarity, detail and transparency.
My other modern DAC's have more clarity and transparency, but the Lite DAC-AH has a natural flow, less analysis is needed when I listen, I always listen analytically so the Lite is a bit refreshing in that respect. I can just listen quite normally, less analysis, that is the definite plus here.
I wish I could remove the fuzziness and achieve higher detail, but that seems to require a different DAC chip.
Lite have many different DAC's using different DAC chips, they are all R-2R chips last time I checked.
You just removed them and it started working?
I checked the heat in mine, from back, transformer, to front, the RCA, then DAC chips 1, 4, 5, 6 are the hottest, DAC chips 7 and 8 are the least hot, 2 and 3 are inbetween. 4, 5 and 6 are too hot for my finger if I push down hard, 1 as well, but the heat levels are quite similar overall.
Abraxalito recommends replacing them with TDA1387, I wonder if there are more alternatives?
The sound is pretty nice after my mod with the vanishingly low distortion of LME49990 and Elna Silmic II's, but the total sound is still quite warm and fuzzy, it's certainly lacking in crystal clarity, detail and transparency.
My other modern DAC's have more clarity and transparency, but the Lite DAC-AH has a natural flow, less analysis is needed when I listen, I always listen analytically so the Lite is a bit refreshing in that respect. I can just listen quite normally, less analysis, that is the definite plus here.
I wish I could remove the fuzziness and achieve higher detail, but that seems to require a different DAC chip.
Lite have many different DAC's using different DAC chips, they are all R-2R chips last time I checked.
Last edited:
My recommendation is a bit more nuanced than that - replacing them but not implementing some passive filtering might not improve the sound. I've found CMOS DACs without downstream filtering can sound a bit harsh. Having already an opamp gain stage in place means a low impedance passive filter could be included, then the resident opamps can recover a line-level signal to output.
Transparency has never been a hallmark of TDA1543 in my experience, warm and fuzzy they always are 🙂
Is TDA1387 warm and fuzzy? How does it compare if we are splitting hairs?
I'm pretty satisfied with LME49990, I don't think I'll try the passive mod, but I may try replacing every single capacitor later to see if that makes any difference.
LT1115 is pretty nice in the line-out as well, it sounded sophisticated, a real tonally coalesced feeling, but then I noticed LME49990 was more revealling and I'm pretty happy with the slightly more atonal sound, somewhat aharmonic, I think that's why the final sound made me think of Arabic-like scales.
I'm a little surprised that the Elna Silmic II's made a slight difference, mostly with female vocals, younger as I wrote, maybe younger is
not the right term, perhaps softer, like frosted glass to normal glass as an extreme analogy. Just my views.
I'm pretty satisfied with LME49990, I don't think I'll try the passive mod, but I may try replacing every single capacitor later to see if that makes any difference.
LT1115 is pretty nice in the line-out as well, it sounded sophisticated, a real tonally coalesced feeling, but then I noticed LME49990 was more revealling and I'm pretty happy with the slightly more atonal sound, somewhat aharmonic, I think that's why the final sound made me think of Arabic-like scales.
I'm a little surprised that the Elna Silmic II's made a slight difference, mostly with female vocals, younger as I wrote, maybe younger is
not the right term, perhaps softer, like frosted glass to normal glass as an extreme analogy. Just my views.
Transparent to my ears but only when well filtered and with a very low noise and low impedance PSU. Transparent though is a relative thing, as I've not got direct access to the original so I might want to revise this view on hearing another DAC which sounds even clearer...
If it's more transparent than the TDA1543 then that's suitable enough for me!
Any other chips which will fit in there that I should be aware of?
I'm not aware of which Nos is the most transparent, the Lingmei XA2 on Taobao looks nice in the pictures and it's using the queen of all R2R.
Any other chips which will fit in there that I should be aware of?
I'm not aware of which Nos is the most transparent, the Lingmei XA2 on Taobao looks nice in the pictures and it's using the queen of all R2R.
No ill stick with TDA1543 🙂
This is what i have in mind:
- Passive mod with K75-10 2.2uf capacitors.
- 680uF/35V Panasonic FC for C28 & C37 (someone mentioned that this increased a little bit bass, have to test.)
- 220uF/25V Panasonic FC for C27,
- 2200/50V uF Panasonic FC for C25 & C26 (had in storage for another project)
- Maybe find some better 150 ohm resistors for R35&R36.
- Maybe do something to the pin7 of the dacs.. remove whole Vrail?
How to do the passive low pass filter that you mentioned?
Last edited:
There are several examples on my blog - it depends on which inductors you can get. Here's one that uses inductors available from Mouser - http://www.diyaudio.com/forums/blogs/abraxalito/1119-ozone-chebyshev-filter.html
I have not touched any resistors. I have a large amount of Takman resistors though from Japan and a store directly across the road from me which sells Audio Note resistors. 🙂))
The original pcb was designed to run dacs at 8volts (see silkprint on pcb at regulators, it says lm7808). Still it has assembled with 7806 regulators. R35 and R36 resistors (270ohm) are valued to run dac at 8volts. Correct value for resistors is 150ohm with 7806 regulators. I havent heard DAC-ah with 270 ohm resistors (already changed to 150 ohm) but i have heard that it affects a lot.
Why do I change to 3.3 V? When there is no sound? Sorry I don't follow you, please write in fine detail.
I have identified R36 and R36, I need to buy resistors for them soon.
Keep in mind I am changing this to 8x or 16x TDA1387.
I have identified R36 and R36, I need to buy resistors for them soon.
Keep in mind I am changing this to 8x or 16x TDA1387.
Interesting article here, please read
"Its a resistor string DAC meaning it hasn't got the glitching problems which arise in R2R type DACs. R2R DACs (like PCM1704) need to have very precise timing between the switches operating the DAC resistor elements or glitches are the result. Resistor string DACs are inherently low in glitch. If low glitching was the reason for the better sound then that hypothesis would also fit the observation that NOS sounds better than oversampling in any given DAC.
I decided to do a little listening experiment to test out the theory that its the absence of the digital filter which contributes to the NOS sound - I built my own filter which had no pre-ringing, implemented this as 2X OS and fed it to my TDA1387 DAC. The result was it didn't sound as good as NOS, there was more greyness to the sound. So my result tends to support the view that its lower DAC glitchiness which contributes to the sound of NOS, not absence of pre-ringing. Fortunately the TDA1387 isn't an R2R, its a better performer on gltich than this architecture."
Musings on audio design - DACs, part II | LinkedIn
"Its a resistor string DAC meaning it hasn't got the glitching problems which arise in R2R type DACs. R2R DACs (like PCM1704) need to have very precise timing between the switches operating the DAC resistor elements or glitches are the result. Resistor string DACs are inherently low in glitch. If low glitching was the reason for the better sound then that hypothesis would also fit the observation that NOS sounds better than oversampling in any given DAC.
I decided to do a little listening experiment to test out the theory that its the absence of the digital filter which contributes to the NOS sound - I built my own filter which had no pre-ringing, implemented this as 2X OS and fed it to my TDA1387 DAC. The result was it didn't sound as good as NOS, there was more greyness to the sound. So my result tends to support the view that its lower DAC glitchiness which contributes to the sound of NOS, not absence of pre-ringing. Fortunately the TDA1387 isn't an R2R, its a better performer on gltich than this architecture."
Musings on audio design - DACs, part II | LinkedIn
Last edited:
Sorry for not so clear instructions 🙂
Pls follow to this forum post, it explains a lot:
Circuit error in Lite DAC-AH: Easy fix
I dont know how this works with TDA1387. But i am very interested about hearing your results with TDA1387 dacs 🙂
- Status
- Not open for further replies.