Hi, posting here on the possibility someone can help.
I have a dual mono Buffalo III (ES9018s) stack which was working fine. I tried to upgrade to Trident SRs with SE firmware chip which sent everything crazy, and I've been troubleshooting since.
In summary, channel behavior is now very strange. I checked my system and its fine with other sources so its the twisted pear stack.
Using test tones flac files, I get this behavior:
Test Tone Type....................................Left Speaker....................................Right speaker
-Right channel only...................................silent.............................................test tone
-Left channel only....................................test tone.......................................test tone
-Left + Right channel..............................test tone.........................................silent
So what should play only on the left channel gets played on both channels, and what should play on both channels only plays through the left channel.
Almost as if the left dac (master dac with firmware) is playing stereo and instructing the right dac to play Mono. I think. I'm trying to work out the logic.
Anyone got any ideas about what is happening and any fixes?
regards
Cedric
I have a dual mono Buffalo III (ES9018s) stack which was working fine. I tried to upgrade to Trident SRs with SE firmware chip which sent everything crazy, and I've been troubleshooting since.
In summary, channel behavior is now very strange. I checked my system and its fine with other sources so its the twisted pear stack.
Using test tones flac files, I get this behavior:
Test Tone Type....................................Left Speaker....................................Right speaker
-Right channel only...................................silent.............................................test tone
-Left channel only....................................test tone.......................................test tone
-Left + Right channel..............................test tone.........................................silent
So what should play only on the left channel gets played on both channels, and what should play on both channels only plays through the left channel.
Almost as if the left dac (master dac with firmware) is playing stereo and instructing the right dac to play Mono. I think. I'm trying to work out the logic.
Anyone got any ideas about what is happening and any fixes?
regards
Cedric
I may be out of touch because it has been a long time since I had 9018s. I’m not sure what you mean by SE firmware or how it might be unique. But…
The notion of an “upgrade” is a bit dubious - it’s more of a binary situation of “it works” or “it won’t work”. So if you changed the microcontroller chip, you should first test by reversing those changes. Hopefully normal channel routing is restored and you can then work better directed toward the original objective.
The notion of an “upgrade” is a bit dubious - it’s more of a binary situation of “it works” or “it won’t work”. So if you changed the microcontroller chip, you should first test by reversing those changes. Hopefully normal channel routing is restored and you can then work better directed toward the original objective.
Hi Francolargo, thank you very much for responding much appreciated.
.
"SE firmware" ... I don't know what I meant either, he he. Must have been a typo, I was referring to the firmware update 3 micro-controller as per TP website:
When used with Buffalo-III/IIIse DACs, Firmware Update 3 is required to compensate for the slightly longer settling time for the Trident-SR.
.
Actually writing the above helped clarify my thoughts, as it being a problem with mono/stereo configuration. So I reset the dual/mono dip switch (switch 3 on switch Block 2) to stereo and back to mono again with the DAC live and now its fixed thankfully. Something so simple!
.
Indeed my setup is about 10 years old now. I spent days last week trying to revert back to the initial configuration with no luck, checking connections and voltages, re-soldering, troubleshooting more self inflicted stupid errors and generally driving myself mad. If it aint broke dont fix it he he, I HATE troubleshooting digital, give me amps or speakers any day.
Perhaps the DIP switches were a little corroded or some register got stuck, who knows.
.
By upgrade I mean sound quality upgrade. I didn't hear any immediate difference changing just the AVCCs from the old shunts to the SRs to be honest, but I'll do some focused A-B listening tests when I have some time and introduce the V3 micro-controller and other trident SRs step by step.
.
Though now also considering the upgrade to 9028 or 9038s and Mercury as I've started down this path.
Any advice on the configuration that would provide the best sound quality?
Is it worth doing dual mono with these dacs? 9038 significantly better that 9028?
.
"SE firmware" ... I don't know what I meant either, he he. Must have been a typo, I was referring to the firmware update 3 micro-controller as per TP website:
When used with Buffalo-III/IIIse DACs, Firmware Update 3 is required to compensate for the slightly longer settling time for the Trident-SR.
.
Actually writing the above helped clarify my thoughts, as it being a problem with mono/stereo configuration. So I reset the dual/mono dip switch (switch 3 on switch Block 2) to stereo and back to mono again with the DAC live and now its fixed thankfully. Something so simple!
.
Indeed my setup is about 10 years old now. I spent days last week trying to revert back to the initial configuration with no luck, checking connections and voltages, re-soldering, troubleshooting more self inflicted stupid errors and generally driving myself mad. If it aint broke dont fix it he he, I HATE troubleshooting digital, give me amps or speakers any day.
Perhaps the DIP switches were a little corroded or some register got stuck, who knows.
.
By upgrade I mean sound quality upgrade. I didn't hear any immediate difference changing just the AVCCs from the old shunts to the SRs to be honest, but I'll do some focused A-B listening tests when I have some time and introduce the V3 micro-controller and other trident SRs step by step.
.
Though now also considering the upgrade to 9028 or 9038s and Mercury as I've started down this path.
Any advice on the configuration that would provide the best sound quality?
Is it worth doing dual mono with these dacs? 9038 significantly better that 9028?
So glad to hear that the channel alignment issue was almost trivial. Had it been related to the workings of a Linux source player, I have some working knowledge to fix it, which is why I spoke up.
One of the reasons that Brian upgraded basically all of Twisted Pear’s power supplies, as well as the tridents, to be comfortable sourcing more current is that the 9028 and particularly the 9038 demand more power. So your new tridents would be perfect should you decide to swap in a newer Buffalo board. For the 9038 you would need to be sure the Buffalo’s power source was adequate, and you would need to add the Mercury I/V to manage the 9038’s higher output. A few years ago I did a project allowing me to compare all three versions of the ESS DAC’s. The Mercury I/V under the older 9018 was a solid improvement over previous TPA I/V designs. And of course the 9028 and 9038 were increasingly detailed. I suggest going for the Mercury - you won’t be disappointed.
One of the reasons that Brian upgraded basically all of Twisted Pear’s power supplies, as well as the tridents, to be comfortable sourcing more current is that the 9028 and particularly the 9038 demand more power. So your new tridents would be perfect should you decide to swap in a newer Buffalo board. For the 9038 you would need to be sure the Buffalo’s power source was adequate, and you would need to add the Mercury I/V to manage the 9038’s higher output. A few years ago I did a project allowing me to compare all three versions of the ESS DAC’s. The Mercury I/V under the older 9018 was a solid improvement over previous TPA I/V designs. And of course the 9028 and 9038 were increasingly detailed. I suggest going for the Mercury - you won’t be disappointed.
Sorry, your additional questions… (All IMHO, of course…)
The internals of 9028 and 9038 used the same strategy as ‘going mono’ with the 9018 - the 9038 being four times more dense than the 9028. So, my opinion would be to use only one, not run them in a mono configuration.
As for 9028 vs 9038, easy in theory - 9038. Does Brian even have ‘28’s left? In practice, I think the difference between the two depends on other factors - from room acoustics all the way back to what the DAC plugs into. With more highly resolving systems, the extra performance of the 9038 becomes more obvious.
The internals of 9028 and 9038 used the same strategy as ‘going mono’ with the 9018 - the 9038 being four times more dense than the 9028. So, my opinion would be to use only one, not run them in a mono configuration.
As for 9028 vs 9038, easy in theory - 9038. Does Brian even have ‘28’s left? In practice, I think the difference between the two depends on other factors - from room acoustics all the way back to what the DAC plugs into. With more highly resolving systems, the extra performance of the 9038 becomes more obvious.
Last edited:
Well sat down and did a critical A-B comparison between the original AVCC shunt tridents vs the Trident-SRs (AD7151s) on my dual mono ES9018s.
So the Trident-SRs sound absolutely fantastic- Detailed, much more 3D, incredibly dynamic and huge instrumental separation. Unfortunately also unlistenable after circa 10 mins, my ears were on edge and had to switch off. Which confirmed my background listening over the last week i.e. I wasn't inclined to do any.
Plopped the old shunt Tridents back in, everything receded but my goodness what a relief, can listen to music again.
So whats going on? Well I'm feeding my Buffalos with ALW (Jung design) Super Regulators rather than Placids. Jung Superregs don't generally like lots of capacitance nor very low impedance output caps; makes 'em oscillate. And on investigation, the Trident-SR have big fat 100uf Murata ECAS caps on their inputs which also have crushingly low ESR (4.5-10mOhms).
Voila there's oscillation, I don't even need to reach for a scope to know. The Trident-SRs do appear to have a small inductor on their inputs but clearly not meaty enough for my particular setup. A salutory reminder that it's all about keeping an eye on basics:
1. Critical listening- something can superficially sound better, but long term listening & emotional response must be your ultimate guide. This was quite an extreme example where 'long term' was <10 minutes and it was a fault he he, but illustrates the point.
2. There is no such thing as a general purpose 'better component', just good potential candidates. One listens to a system not a collection of components. Systems mean interactions, so it's the complementary integration of components within the whole that matters. The "Put an 800BHP engine into a VW Beetle and launch into a tree," metaphor.
Happily don't think this will be a big issue to rectify. Either an inductor on the Jung design SR, playing with its output cap or swapping out the input caps on the Tridents for something smaller and/or higher ESR should fix it.
Plunging for ESS9038 + Mercury so the problem needs to be addressed.
So the Trident-SRs sound absolutely fantastic- Detailed, much more 3D, incredibly dynamic and huge instrumental separation. Unfortunately also unlistenable after circa 10 mins, my ears were on edge and had to switch off. Which confirmed my background listening over the last week i.e. I wasn't inclined to do any.
Plopped the old shunt Tridents back in, everything receded but my goodness what a relief, can listen to music again.
So whats going on? Well I'm feeding my Buffalos with ALW (Jung design) Super Regulators rather than Placids. Jung Superregs don't generally like lots of capacitance nor very low impedance output caps; makes 'em oscillate. And on investigation, the Trident-SR have big fat 100uf Murata ECAS caps on their inputs which also have crushingly low ESR (4.5-10mOhms).
Voila there's oscillation, I don't even need to reach for a scope to know. The Trident-SRs do appear to have a small inductor on their inputs but clearly not meaty enough for my particular setup. A salutory reminder that it's all about keeping an eye on basics:
1. Critical listening- something can superficially sound better, but long term listening & emotional response must be your ultimate guide. This was quite an extreme example where 'long term' was <10 minutes and it was a fault he he, but illustrates the point.
2. There is no such thing as a general purpose 'better component', just good potential candidates. One listens to a system not a collection of components. Systems mean interactions, so it's the complementary integration of components within the whole that matters. The "Put an 800BHP engine into a VW Beetle and launch into a tree," metaphor.
Happily don't think this will be a big issue to rectify. Either an inductor on the Jung design SR, playing with its output cap or swapping out the input caps on the Tridents for something smaller and/or higher ESR should fix it.
Plunging for ESS9038 + Mercury so the problem needs to be addressed.
Interesting reading. 
Certainly, it would be preferable to focus on the main supply to the new Buffalo in this case. The ESS DACs like their current, and by some reports, regulators with higher ESR are audibly less engaging. Let me see if I can find anything...
ESS makes regulators fairly well optimized for their DACs and this data sheet will give you insights:
https://www.esstech.com/wp-content/uploads/2024/03/ES9312_DataSheet_v0.2.1.pdf
Also, some Buffalo users have experimented with directly supplying individual tridents (or subbing other regulators). Here's one user who did so:
http://twistedpearaudio.com/forum/p...lass-library-for-the-ES9028-38-chip#post25711
If you dig into this, remember that VCC_l, VCC_R, and DVDD should be at 1.3v, as per the ES9038 data sheet.
Edit: I once tried individually powering the Buffalo clock with 3.3v from an LT3045 and didn't like the result. The idea was to switch power on/off for different 9038 operational modes. But when I changed the LT3045 to put out ~4.5v and ran that to the trident, it was slightly better than the trident fed from the main Buffalo power circuit.
Cheers,
Frank
Certainly, it would be preferable to focus on the main supply to the new Buffalo in this case. The ESS DACs like their current, and by some reports, regulators with higher ESR are audibly less engaging. Let me see if I can find anything...
ESS makes regulators fairly well optimized for their DACs and this data sheet will give you insights:
https://www.esstech.com/wp-content/uploads/2024/03/ES9312_DataSheet_v0.2.1.pdf
Also, some Buffalo users have experimented with directly supplying individual tridents (or subbing other regulators). Here's one user who did so:
http://twistedpearaudio.com/forum/p...lass-library-for-the-ES9028-38-chip#post25711
If you dig into this, remember that VCC_l, VCC_R, and DVDD should be at 1.3v, as per the ES9038 data sheet.
Edit: I once tried individually powering the Buffalo clock with 3.3v from an LT3045 and didn't like the result. The idea was to switch power on/off for different 9038 operational modes. But when I changed the LT3045 to put out ~4.5v and ran that to the trident, it was slightly better than the trident fed from the main Buffalo power circuit.
Cheers,
Frank
Last edited:
One other little link from the same user as the first one above:
http://www.twistedpearaudio.com/forum/posts/m25955-Cronus-and-Hermes-sound-upgrade#post25955
(S)He reports dramatic improvement from isolating the XO 3.3v from the analog supply. I would say 'noticeable improvement' when I added an LT3045 to supply the clock's Trident, but I was only using that clock for fairly messy SPDIF input. The rest of the time the DAC is slaved to the source clock.
http://www.twistedpearaudio.com/forum/posts/m25955-Cronus-and-Hermes-sound-upgrade#post25955
(S)He reports dramatic improvement from isolating the XO 3.3v from the analog supply. I would say 'noticeable improvement' when I added an LT3045 to supply the clock's Trident, but I was only using that clock for fairly messy SPDIF input. The rest of the time the DAC is slaved to the source clock.
Hi Franco, thanks for the links and tips!
I remember I used to use separate regulated PSU supplies for my clock and DAC on my CD player years ago, one of the few mods that really paid dividends on it. Looks like should try that again.
The specs on the LT3045 do indeed look better than the AD7151 chips, mainly rejection though. Probably less an issue if using fully separate psus, maybe one day.
Sigh, I do see a full rebuild coming now though. What a PITA, but I'm getting a taste for blood again he he.
cheers C
I remember I used to use separate regulated PSU supplies for my clock and DAC on my CD player years ago, one of the few mods that really paid dividends on it. Looks like should try that again.
The specs on the LT3045 do indeed look better than the AD7151 chips, mainly rejection though. Probably less an issue if using fully separate psus, maybe one day.
Sigh, I do see a full rebuild coming now though. What a PITA, but I'm getting a taste for blood again he he.
cheers C
In fact, while I'm at it. You mentioned The rest of the time the DAC is slaved to the source clock.
.
You're using your Buffalos in synchronos mode is that right then? Ive been working my way through the original threads on the release.
There were some reports of a good improvement going this route but custom firmware needed to loaded onto the microcontroller.
Is that still the case or do current firmware chips enable synchronos?
.
You're using your Buffalos in synchronos mode is that right then? Ive been working my way through the original threads on the release.
There were some reports of a good improvement going this route but custom firmware needed to loaded onto the microcontroller.
Is that still the case or do current firmware chips enable synchronos?
"Somewhere out there" are firmware codes for synchronous ESS Pro DAC operation. In my builds I have a SOC computer interfaced and don't use the onboard firmware chips so A) I have the ability to test and tune various operating modes without need of coding by others, B) I have experience with using the I2C port on the Buffalo and, C) I've experienced some of the upsides and downsides of that flexibility. On the downside, the ESS Pro DAC memory registers are not the most stable, and the I2C controller and interface must be really 'clean' with respect to shielding and isolation. That's a whole different topic in itself, but once that is achieved the controls operate with little worry. On the upside, I can extract every available shred of -140dB goodness out the hardware. Now, in my system I prefer the synchronous mode but I was also surprised at how good the system sounded in default asynch mode after separately powering its onboard clock. I would advise doing that, if even just adding a clean-up regulator (or whatever) to a branch of the clean 5v supply to the buffalo. That's all I did with the LT3045. If you're comfortable with linux for I2C communications, I can share some Python scripts that have worked for me. My to-do list includes putting all the DAC controls on a web page served within the BBB or RPI built into the player/DAC. ...eventually... Last thought for now: in my humble experience, the new DAC will wow you straight away. But the best is yet to come after a few months of 'burn-in' if you will. If you're like me, you'll become a critic of music production and mastering within the content you hear. That's how revealing this hardware can be. I even sometimes listen to artists about whom I'm not particularly ecstatic - just because their recording and mastering engineers made such a great recording!
Thanks again for your advice, Franco!
OK i'll skip synchronous for now and fully isolate the Buffalo supplies with separate psus instead. I find digital and coding infuriating, so the less I have to do the happier I'll be.
I look forward to being Wowed by the presentation in due course then, he he. Listening to music you wouldn't normally is a good sign to me, something very right is happening there.
Better start designing the chassis layout then. Happily I have a couple of prototype PSU boards and parts from last year that I put on the back-burner due to an minor but annoying PCB bug. That will be something like 5 transformers and 11 isolated rails in total for this dac/music server, ridiculous!
OK i'll skip synchronous for now and fully isolate the Buffalo supplies with separate psus instead. I find digital and coding infuriating, so the less I have to do the happier I'll be.
I look forward to being Wowed by the presentation in due course then, he he. Listening to music you wouldn't normally is a good sign to me, something very right is happening there.
Better start designing the chassis layout then. Happily I have a couple of prototype PSU boards and parts from last year that I put on the back-burner due to an minor but annoying PCB bug. That will be something like 5 transformers and 11 isolated rails in total for this dac/music server, ridiculous!