The capacitors mostly help the low end. The high end should be covered by the ceramic caps already in place. And, as your example demonstrates, effects of leads are completely out of the audio band (not only for humans but any primates).
As done in the picture (on the back of the PCB) I don't see a problem WRT to damaging the board, but I would prefer soldering caps as close as possible to the PCB via the provided holes. This works better with normal electrolytics of course. I used only Panasonic FR and FC caps and got a vast improvement in SQ.
Last edited:
I use this:
PL2303TA USB TTL to RS232 Converter Serial Cable module for win 8 XP VISTA 7 8.1 | eBay
Very cheap and it works perfectly!
Fedde
OK, thanks, to be sure I’m clear I put a picture from a setup as discribed. Seems easier to solder...
Not sure if those vias exist anymore on rev4 and 5... definitely not on 4...
Out of curiosity, do MLCC and electrolytic/polymer type caps have different “frequency response” or just the ESR curve? I originally thought capacitance and ESR are the only two factors that mattered.. thanks!
I'm no expert, but I think several parameters, including resonance, will vary.
Still felt that I ought to comment on what the cap mod did to the sound.
The low end was tight and punchy before the mod (rev4 dual mono); with the mod, it increased dramatically in quantity. It doesn't feel overwhelming but definitely much more free flowing and natural. It seems closer to what you'd hear in concerts (though maybe still lacking if you're into heavy metal/pop/...?), but others may have a better opinion on this.
It seems to have made other instruments more life-like as well. A huge improvement in naturalness I'd say. As I mentioned in earlier posts, dam1021 sounded much harsher and fatiguing last year with the adapted sigma22 PSU and greatly improved in naturalness after I installed Salas UltraBiB. This could be influenced by the pops I got from the disconnected I2S GND wire at first (in both PSU setup), but I'm relatively confident of the observation - harshness-induced headaches are hard to dream up...
Better soundstage/imaging/... I'm not super well-versed in these terminologies but I think all-round subjective improvements are hardly surprising given how a lower voltage ripple should just help everything.
There is less perceived details, but some could've been artifacts to start with. The new representation is probably closer to live performances in my opinion. Btw, I'm using HD650 on the buffered output - not very transparent equipment and I think I might be starting to hear limitations of the headphones rather than dam1021. Very happy with the mod and dam1021 in general. Definitely recommend adding caps.
Edit: I also improved the I2S cable and Amanero power supply when I added the caps, but I doubt there's much if any change in SQ due to the digital signal improvements - just expecting better stability.
The low end was tight and punchy before the mod (rev4 dual mono); with the mod, it increased dramatically in quantity. It doesn't feel overwhelming but definitely much more free flowing and natural. It seems closer to what you'd hear in concerts (though maybe still lacking if you're into heavy metal/pop/...?), but others may have a better opinion on this.
It seems to have made other instruments more life-like as well. A huge improvement in naturalness I'd say. As I mentioned in earlier posts, dam1021 sounded much harsher and fatiguing last year with the adapted sigma22 PSU and greatly improved in naturalness after I installed Salas UltraBiB. This could be influenced by the pops I got from the disconnected I2S GND wire at first (in both PSU setup), but I'm relatively confident of the observation - harshness-induced headaches are hard to dream up...
Better soundstage/imaging/... I'm not super well-versed in these terminologies but I think all-round subjective improvements are hardly surprising given how a lower voltage ripple should just help everything.
There is less perceived details, but some could've been artifacts to start with. The new representation is probably closer to live performances in my opinion. Btw, I'm using HD650 on the buffered output - not very transparent equipment and I think I might be starting to hear limitations of the headphones rather than dam1021. Very happy with the mod and dam1021 in general. Definitely recommend adding caps.
Edit: I also improved the I2S cable and Amanero power supply when I added the caps, but I doubt there's much if any change in SQ due to the digital signal improvements - just expecting better stability.
Last edited:
Everything you said, I only did one board so far for the left side, and already there is a significant imbalance in response across the board that I can tell. The bass is definitely more tighter!
Each cap is soldered directly to the bypass capacitor of each shift register (had no room for the 5th cap due to the vref series pass transistor. If you use the standard 6.3mm spacing it fits perfectly, and the bypass capacitors are slightly taller than the resistor arrays next to them, so the caps never touch anything other than the power pins and substrate of shift register.
Time to do the 2nd board.
Each cap is soldered directly to the bypass capacitor of each shift register (had no room for the 5th cap due to the vref series pass transistor. If you use the standard 6.3mm spacing it fits perfectly, and the bypass capacitors are slightly taller than the resistor arrays next to them, so the caps never touch anything other than the power pins and substrate of shift register.
Time to do the 2nd board.
Attachments
Nice! Great to hear that others are benefiting from adding extra capacitance. Am I right to assume that you added 4(now 5)*330uf per rail in addition to the 47uf Soren put in rev2 and a larger electrolytic per rail? Did you also have to do the transistor mod or is that something else?....
I have rev4 so went from 347uf to ~5000uf per rail - the change was nothing short of dramatic. I don’t have a lot of experience (and frankly interest in...) evaluating equipment. But the bass quantity (and at least mid-bass quality as far as I can tell on HD650) has greatly increased. Both continuously powered on for multiple days, the pre-mod bass was very tight and lean, feeling quite wanting, slightly artificial and attracting attention. The mod eliminated all those problems. The deep bass sounds ever so slightly loose with HD650 but it might have nothing to do with dam1021. Like I said earlier, everything is more natural and smoother. It might not make sense but I feel like I’m just hearing the headphones and recording now. If dynamics can be said to be somewhat lacking, it’s not the case anymore, possibly with the help of a much more engaging and natural base.
Soren, did you test higher vref capacitance when you decided on 1000uf for your dac line? You said rev4/5 don’t need the mod but it seems to me critical...
I'm sure he can hear great differences on dam1021 rev4/5 if he add caps. He's also probably the best person to try to provide some theoretical explanations for this dramatic effect... The more time you spend looking at the measurements compiled on Hifidrino the less they seem to explain the empirical results...
Thanks for providing a good laugh
I was saying exactly that to a friend the other day, that I think Soeren did not even listen to his creation prior to opening the sale, just a look at the measurements are enough.
Or maybe listening with some desktop speakers to check functionality.
But at least there is cooperation with the listening crowd here in some points.
BTW I managed to sucessfully solder the Si570 to my Dam1121, it is playing again. So far so good, but as often described the sound is still a little on the "technical" side
Did anyone add caps to Vref on the OEM version? Really want to like this dac, as I put a lot of effort in it.
Thanks for providing a good laugh
I was saying exactly that to a friend the other day, that I think Soeren did not even listen to his creation prior to opening the sale, just a look at the measurements are enough.
Or maybe listening with some desktop speakers to check functionality.
But at least there is cooperation with the listening crowd here in some points.
BTW I managed to sucessfully solder the Si570 to my Dam1121, it is playing again. So far so good, but as often described the sound is still a little on the "technical" side
Did anyone add caps to Vref on the OEM version? Really want to like this dac, as I put a lot of effort in it.
You don't mean that we could put any new oscillator on the board right?... (within power consumption specs of course) But I should stop myself there since there is probably little point in improving the clock further as Soren pointed out...
I have a strong feeling that the caps will solve the "technical" sounding problem. It seems there are 10 vref caps per rail on 1121? So maybe just parallel a large electrolytic to start with and see if it works?... Spikestabber used SMD caps successfully and perhaps that's an even better option for 1121.
Last edited:
I had to put a new clock in, because I stupidly killed the original one. 6 weeks lead time from Mouser on top as a penalty for being an idiot once more...
Yes, I also have the hope for those Vref caps. I am still tempted to traý batteries, as it is 3.8V with stock regs, and 3.6V with Lifepo4 batteries. So just a little lower.
But I am afraid to destroy it again in the process, so Vref caps seem to be easier, and maybe enough. But there are no through holes as in 1021 of course
Yes, I also have the hope for those Vref caps. I am still tempted to traý batteries, as it is 3.8V with stock regs, and 3.6V with Lifepo4 batteries. So just a little lower.
But I am afraid to destroy it again in the process, so Vref caps seem to be easier, and maybe enough. But there are no through holes as in 1021 of course
I had to put a new clock in, because I stupidly killed the original one. 6 weeks lead time from Mouser on top as a penalty for being an idiot once more...
Yes, I also have the hope for those Vref caps. I am still tempted to traý batteries, as it is 3.8V with stock regs, and 3.6V with Lifepo4 batteries. So just a little lower.
But I am afraid to destroy it again in the process, so Vref caps seem to be easier, and maybe enough. But there are no through holes as in 1021 of course
So what's actually stopping us from replacing SI514 on dam1021 (orSI570 on dam1121) with the best SI oscillator? FPGA code?...
Soren once said that the battery leads themselves could degrade the vref performance significantly. I would be careful to go for the battery mod due to the complexity and how much easier it is to break things on dam1121... Try the cap mods first to see if it satisfies you. There is already 1000uf I think per rail on 1121 but can't hurt to try. living sound reported improvements adding 8000uf to ~1000uf per rail on dam1021
Yes, probably firmware is stopping us. But honestly, I dont think it is a big difference between 514 and 570, there are more important tweaks as mentioned before.
Yes, that's why I am reluctant to go the battery route completely, too small everything. But I will try it for the clock though, letting the battery act as a huge cap being charged constantly by the inboard regulator. I did this with the USB board with very good results.
Yes, that's why I am reluctant to go the battery route completely, too small everything. But I will try it for the clock though, letting the battery act as a huge cap being charged constantly by the inboard regulator. I did this with the USB board with very good results.
The funny part is my original dam1021 that I did the 16 cap mod to sounded similar to what I hear now, I felt I lost something since going dual-mono and adding less capacitance to vref on the basis that it wasn't really needed, but it certainly is all back now! Running roughly 3200uF per rail now factoring in the ceramics. More caps you can tie closer to the shift registers are likely better than a few single larger ones. The smd caps I used specifically digikey part# 732-6399-1-ND
Hearing The Old Ways by Loreena McKennitt just has me grinning once again. Impressed.
Hearing The Old Ways by Loreena McKennitt just has me grinning once again. Impressed.
I had to put a new clock in, because I stupidly killed the original one. 6 weeks lead time from Mouser on top as a penalty for being an idiot once more...
Yes, I also have the hope for those Vref caps. I am still tempted to traý batteries, as it is 3.8V with stock regs, and 3.6V with Lifepo4 batteries. So just a little lower.
But I am afraid to destroy it again in the process, so Vref caps seem to be easier, and maybe enough. But there are no through holes as in 1021 of course
How did you kill it? So we know what not to do
//