Hi guys,
There are two things to change besides op amps - first the tcxo (rectangular silver lump) which should be 24.576Mhz ? It should be running off 3.3V so the best replacement is Crystek cchd-957, next Tent labs xo, and then Vanguard 0.3ppm (not 1ppm - that is 5v only). The clock will make a BIG difference.
Next, the cap at the bottom of the adjust pin of the 317/337. Big caps (100uF) kill noise but dampen the dynamics a little. Small caps (0.047uF) are noisier but the sound is much more dynamic. Playing with this cap will change the output impedance so you'll notice changes in the midrange - some quite unpleasant - but when you get the right value, it's heaven. I use Wima black box on the adjust pins in my CS4398 dac.
Have fun !
BTW here's a link to someone who's done this - I can confirm it's right.
http://tech.juaneda.com/en/articles/lm317.html
There are two things to change besides op amps - first the tcxo (rectangular silver lump) which should be 24.576Mhz ? It should be running off 3.3V so the best replacement is Crystek cchd-957, next Tent labs xo, and then Vanguard 0.3ppm (not 1ppm - that is 5v only). The clock will make a BIG difference.
Next, the cap at the bottom of the adjust pin of the 317/337. Big caps (100uF) kill noise but dampen the dynamics a little. Small caps (0.047uF) are noisier but the sound is much more dynamic. Playing with this cap will change the output impedance so you'll notice changes in the midrange - some quite unpleasant - but when you get the right value, it's heaven. I use Wima black box on the adjust pins in my CS4398 dac.
Have fun !
BTW here's a link to someone who's done this - I can confirm it's right.
http://tech.juaneda.com/en/articles/lm317.html
Last edited:
Hi,
Please refer to the attached link for detail of the various mini kit on Salas shunt reg,
http://www.diyaudio.com/forums/group-buys/188974-gb-thread-salas-ssr1-1-bib-shunt-reg-4.html
I got both BJT and BIB kit sets and the DCB1.
I have disconnected the IC, coax and usb, so there is no data comming in, I measure the DC offset before the divider and at the RCA port and both didnot register any reading at all, don't know if my meter is good enough or not.
I think the 2 Tonerex caps 47uF 50V + the 104 and 222 are PS decoupling caps. Please correct me if wrong.
Attach are images of the opamp section for better viewing.
I think the 2 Tonerex caps 47uF 50V + the 104 and 222 are PS decoupling caps. Please correct me if wrong.
Attach are images of the opamp section for better viewing.
Last edited:
Hi,
I also bought some black box and planed to replace that cap on 317/337. Will a LT317/337 have better noise rejection? I can change from LM to LT then lower the 100uf to 4.7uf BG N or 47 BG N. As for the tcxo, If I change to Vanguard 0.3ppm then I will need to pop the 3.3 reg and replace it with 5V reg?
Thanks for all the information.
If you look at the schematics, there's not much a Salas supply or any other may do if you don't eliminate what's already there.
That's why I had suggested starting "from the outside" and work inwards in mods.
That is:
1) Improve the output. Try other ICs to see if they improve on what's already there. Some dual chips were already suggested. I ordered some AD827s and will also try some LM4562 when i get them.
2) Improve the input. A transformer might do that. But you will need to add another at the source. Finding out more about that mod.
3) Improve the raw supply. An external CRC would clean the raw DC being fed to the regulators.
Besides that I already suggested on the correct implementation for the 317/337 regulators, increasing the adj leg bypass cap and decreasing the output cap.
Those are the first changes I will perform on my DAC when it gets here.
That's why I had suggested starting "from the outside" and work inwards in mods.
That is:
1) Improve the output. Try other ICs to see if they improve on what's already there. Some dual chips were already suggested. I ordered some AD827s and will also try some LM4562 when i get them.
2) Improve the input. A transformer might do that. But you will need to add another at the source. Finding out more about that mod.
3) Improve the raw supply. An external CRC would clean the raw DC being fed to the regulators.
Besides that I already suggested on the correct implementation for the 317/337 regulators, increasing the adj leg bypass cap and decreasing the output cap.
Those are the first changes I will perform on my DAC when it gets here.
Last edited:
Hi,
I don't have any experience with LT317 but it seems from the datasheet that this has been made better so it has better accuracy and line/load regulation, but is otherwise an LM317. So it's very likely the same rules apply re. adjust cap and this will be how you get the best sound out of the reg. Remember that caps burn in so don't be too quick to swap stuff around. Get used to leaving it on for a few days before critical listening. Slow, but hey, you wanna listen to music anyway right ;-)
You can go down to 0.0047uF IIRC before the impedance changes are out of the audio band. It is 1uF or less that you'll really notice the change. Try changing it in decades - 100 to 10, 10 to 1, 1 to 0.1, 0.1 to 0.01, etc. When you've tried a few changes, you'll know what you like and can fine tune.
The Vanguard 0.3ppm will work on both 3.3V and 5V so no need to change regs. If you know Vocative Audio - Chris and his partner - are great for Vanguard and they do mod work if you want too.
V.A. Hi Fi www.VocativeAudio.com The advanced Hi-Fi shop in China-Hong Kong
For the offset, is your dmm set to dc mV ? Measure at the RCA output. If it's zero, then either you have a super DAC or there is a dc blocking cap. A good way to know is to measure the offset when the DAC is fully warmed up, and then from cold when you first turn it on. They should be a little different.
There's no way to tell without seeing the underside with markings for what's what. However, you can measure the dc voltage across each electro cap. Anything in the low mV would indicate it is a DC blocking cap. Most will be the +/-12.5V supply.
I don't have any experience with LT317 but it seems from the datasheet that this has been made better so it has better accuracy and line/load regulation, but is otherwise an LM317. So it's very likely the same rules apply re. adjust cap and this will be how you get the best sound out of the reg. Remember that caps burn in so don't be too quick to swap stuff around. Get used to leaving it on for a few days before critical listening. Slow, but hey, you wanna listen to music anyway right ;-)
You can go down to 0.0047uF IIRC before the impedance changes are out of the audio band. It is 1uF or less that you'll really notice the change. Try changing it in decades - 100 to 10, 10 to 1, 1 to 0.1, 0.1 to 0.01, etc. When you've tried a few changes, you'll know what you like and can fine tune.
The Vanguard 0.3ppm will work on both 3.3V and 5V so no need to change regs. If you know Vocative Audio - Chris and his partner - are great for Vanguard and they do mod work if you want too.
V.A. Hi Fi www.VocativeAudio.com The advanced Hi-Fi shop in China-Hong Kong
For the offset, is your dmm set to dc mV ? Measure at the RCA output. If it's zero, then either you have a super DAC or there is a dc blocking cap. A good way to know is to measure the offset when the DAC is fully warmed up, and then from cold when you first turn it on. They should be a little different.
There's no way to tell without seeing the underside with markings for what's what. However, you can measure the dc voltage across each electro cap. Anything in the low mV would indicate it is a DC blocking cap. Most will be the +/-12.5V supply.
Try reading this. You might decide there is no correct, only what you like, or perhaps what sounds good in a given environment.
Simple Voltage Regulators Part 2: Output Impedance
Last edited:
What's also possible is to use leds (infared led have the lowest noise) at the adjust pin instead of a resistor, or a 11V zener (for 12.25V output). Here's an example :
I've used 2 green leds at the adjust pin instead of a resistor for 5v output and that works well. You can do a google search to find out more if you want to. Here's a site :
http://www.acoustica.org.uk/t/3pin_reg_notes1.html
I've used 2 green leds at the adjust pin instead of a resistor for 5v output and that works well. You can do a google search to find out more if you want to. Here's a site :
http://www.acoustica.org.uk/t/3pin_reg_notes1.html
This is a DIY forum, and we are certainly capable to, probably obliged to, test different combinations to get best results. Audibly, of course.
There might be certain differences with the subjective test results made by Eric Juaneda, as he used a single supply. Dual supplies, by nulling several parameters, might provide different results from single ones.
Juaneda findings that using no capacitors in adj leg and output are systematically good make me doubt of the whole test, though.
Using zener diodes (properly filtered) on the adj leg should improve things too, as well as LEDs. Using a transistor there it's also good.
There might be certain differences with the subjective test results made by Eric Juaneda, as he used a single supply. Dual supplies, by nulling several parameters, might provide different results from single ones.
Juaneda findings that using no capacitors in adj leg and output are systematically good make me doubt of the whole test, though.
Using zener diodes (properly filtered) on the adj leg should improve things too, as well as LEDs. Using a transistor there it's also good.
Hi Tom,
If you've read this entire thread, you know where I stand on 'too' many mods - if any.
I really have no intention of making a lot of changes to the board - especially if the payback in terms of sonic qualities is either; not that noticeable audibly, or negligible at best. That's not to say, I won't try rolling a few different op amps to improve the sound in some way. I just don't want to be continually chasing the holy grail of DAC's and op amps in search of audio nirvana. I just feel in the end, there are other pieces of gear in my chain that would likely mask any improvement an op amp, or other circuit mod might make. But just to be clear though, I certainly don't begrudge others for doing so, because in the process I usually learn something. That's what DIY'ing is all about!Having said that... some different op amps showed up in the mail today . One is a drop-in I believe (AD827), and the other one (LM4562) needs a Browndog adapter which hasn't arrived yet. Maybe tomorrow.
These were recommended by others earlier in thread. Time to go back and re-read some posts.... My philosophy is very much like yours on mods.
Besides the "more external" areas I mentioned, which more or less controls what goes in and out, there's not that much that might be gained swapping passive parts.
Now that most resistors are standard 1% MF, only filtering, output or bypass capacitors might need some looking into. I'm not sure what type the small caps now on the output trio are, as polysytyrenes and polypropylenes do make a difference usually. But that will very much depend on the rest of the chain.
The DAC main crystal (which is not on the schematic!) change might bring a large improvement, as many serious people have done it and the feedback is very exciting, but I believe those other things should be swapped first.
Going after the Nirvana is one of the things DIY is about, as long as you do not get obsessive over it.
Besides the "more external" areas I mentioned, which more or less controls what goes in and out, there's not that much that might be gained swapping passive parts.
Now that most resistors are standard 1% MF, only filtering, output or bypass capacitors might need some looking into. I'm not sure what type the small caps now on the output trio are, as polysytyrenes and polypropylenes do make a difference usually. But that will very much depend on the rest of the chain.
The DAC main crystal (which is not on the schematic!) change might bring a large improvement, as many serious people have done it and the feedback is very exciting, but I believe those other things should be swapped first.
Going after the Nirvana is one of the things DIY is about, as long as you do not get obsessive over it.
qusp,
Are you recommending these as drop-in replacements to the OP275, or just stating that these are some other DIPS dual opamps? I'm putting together a summary of recommended op amps as replacements for this DAC. Just want to be sure.
Since there's been a lot of good activity, comments and suggestions in this thread so far, I thought I would re-cap what op amps have been recommended so far. There seems to be an interest by others to try this DAC too, so hopefully this list will help future owners.
My list includes the following. If I missed any, my apologizes.
LME49720 - Drop-in
OPA827 - x 2 on Browndog adapter
LME49860 - Drop-in
LM6172 - w/cap mods
AD827 - Drop-in
OPA1642 - on adapter
LM4562 - Can Casing on Browndog adapter
OPA2134PA
OPA2132PA
AD8620 - on adapter (Browndog I presume)
I swapped in the AD827's this afternoon and have been doing some casual listening at my computer. Don't know if there's a required break-in period, but a couple things I've noticed. Seems there is an ever so slight emphasis on the bass. Not in a 'boomy' way, but rather a more subtle solid punch to it (good). Seems I can hear a little more sparkle to the piano too. Very,very subjective.
My list includes the following. If I missed any, my apologizes.
LME49720 - Drop-in
OPA827 - x 2 on Browndog adapter
LME49860 - Drop-in
LM6172 - w/cap mods
AD827 - Drop-in
OPA1642 - on adapter
LM4562 - Can Casing on Browndog adapter
OPA2134PA
OPA2132PA
AD8620 - on adapter (Browndog I presume)
I swapped in the AD827's this afternoon and have been doing some casual listening at my computer. Don't know if there's a required break-in period, but a couple things I've noticed. Seems there is an ever so slight emphasis on the bass. Not in a 'boomy' way, but rather a more subtle solid punch to it (good). Seems I can hear a little more sparkle to the piano too. Very,very subjective.
I agree. However, it would be nice to know prior to making a mod if you can expect a 'real' tangible and noticeable difference in the sound. To me, those are the mods worth investigating further and possibly implementing. The decision will vary with individual.
Last edited:
You are asking a bit too much. After a certain point, differences may add a little, and quite likely it will be barely noticeable.
But I don't think you have reached that point yet. Different ICs will bring changes to the sound. Not necessarily improvement, just changes, which you will judge if you like or not.
But you're right that things have to burn-in, the DAC itself has too. Leave it playing all doing long for several days, 10 or 15 days at least.
Every now and then listen to it if you want to, and leave it playing.
You have to discover what sound you're looking for. I myself look for real-sounding instruments, voice too.
But I don't think you have reached that point yet. Different ICs will bring changes to the sound. Not necessarily improvement, just changes, which you will judge if you like or not.
But you're right that things have to burn-in, the DAC itself has too. Leave it playing all doing long for several days, 10 or 15 days at least.
Every now and then listen to it if you want to, and leave it playing.
You have to discover what sound you're looking for. I myself look for real-sounding instruments, voice too.
Hi Tom,
If you've read this entire thread, you know where I stand on 'too' many mods - if any.
Having said that... some different op amps showed up in the mail today . One is a drop-in I believe (AD827), and the other one (LM4562) needs a Browndog adapter which hasn't arrived yet. Maybe tomorrow.
Hi,
A low jitter xo will make a very noticeable difference. The best is the Crystek CCHD-957. It's not expensive and will make noticeable improvements in definition, space, decay, distortion, etc. How much difference depends on how good or otherwise the existing clock is.
AD827 is really for video and is quite noisy. Compare that to AD8599 which is designed for audio pre-amps.
The ad1955 really wants an op amp that is fet or j-fet input, not bjt, in the I/V pair because it won't strain the ad1955 output, hence AD8620 or OPA2134 as possible substitutes. But try and see how it sounds eh ? If it blends with your system well, then job done. I'll be interested to hear your comments.
For the buffer, you can try all different op amps. I know people who love the LT1028... and LM4562 is great there too. Op275 is much warmer with a distinct lower midrange "boost".
Too many op amps...not enough time
Last edited:
Hi,
If you look at the output impedance plots on one of the links I posted, you'll see the cap on the adjust pin increases the output impedance at low to mid frequencies, which supports Eric's (and mine) observations that the dynamics improve substantially. The cap on the output lowers the impedance at mid and higher frequencies. So balancing them to work together and avoiding low esr caps on the output are the way to go...
In a perfect world, I want a flat output impedance across the audio band for a balanced sound, and as low as possible for maximum dynamics.
I don't particularly like LM317 because it is hard to achieve a flat output impedance and it is never that low. This is why shunt regs and very good series regs like AMB's are better. I even prefer the sound of the humble 7805 which is mostly flat until the higher treble range. Some people have said LT1084 sound better than LM317 but I've yet to try.
Just my 2 cents.
Thanks for that confirmation on some (opamps) that have already been mentioned. Sounds like an xo replacement may be where I notice an audible difference. I've already been looking at the TentLabs offerings for some insight. Sounds like if we're trying to reduce noise then the AD827 is not the one to use.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.