Cool. Glad to hear this about the BIII 9028 being a drop in. Need to tie up loose ends and finalize another All-In-One project.
With this upcoming project I'm going to use this BAL SE kit - Balanced to Single Ended Kit | Joe audio - to convert the NTD1's balanced output to single ended output. I'm thinking of also maintaining the balanced output as well. Would having both this converter and balanced out be an issue? e.g. Is there signal loss?
With this upcoming project I'm going to use this BAL SE kit - Balanced to Single Ended Kit | Joe audio - to convert the NTD1's balanced output to single ended output. I'm thinking of also maintaining the balanced output as well. Would having both this converter and balanced out be an issue? e.g. Is there signal loss?
I'm a bit slow off the mark but I'm liking the look of the 9028 or even 9038 DAC boards from twisted pear. I'm currently running two 9018 in dual mono setting into an NTD1. Owen helped me to adjust the gain and current resistors so that I could run the NTD1 board at 32v.
Build Thread - A New Take on the Classic Pass Labs D1 with an ESS Dac
So, I have some decisions about what to do:
1) Literally swap out the 9018 boards for 9028 boards and deal with the mismatched DAC board problem that Russ White mentioned. I'm not keen on using the limiting function of the new DACs and having lived with the mismatching issue in the 9018 boards for so long (and not been bothered) I'm OK with that. I might need to tweak the heat sinks on my Paul Hynes 'tridents' but I think they'll be able to deliver enough current.
2) Remove the dual mono DAC boards and replace with a 9038. presumably, this would require a further reduction in the 4 resistors (R3& R4 to around 45R (Caddock MP930) and also R1&2 to around 100) and also a reduction in the power supply to the board to around 20-25V (these are all estimates only). This is quite a bit of work but I would then have the best board and would reduce the complexity of my DAC significantly. The digital signal benefited from the shortest possible path-length when I was doing initial tests but in the final case I needed to use slightly longer lengths of coax - not night and day differences but certainly noticeable.
3) Buy a 9038 board and the new mercury IV. This would certainly end up at the simplest configuration and I could probably re-use the power supplies from the NTD1 with some adjustment. I read that barrows is going to do this. I am also pretty sure that SCompRacer had a dual mono 9018 build and it looks like he is going for a single, stereo 9028.
4) Ask Russ White to progress his Dual mono, single PCB board DAC!
So...some decisions!
It looks like 4 is out for the time being. 1 is relatively easy but I'm drawn to the best board! I'll need to talk to some NTD1 owners who've heard the Mercury.
Any thoughts, input etc very welcome!
Crom
Build Thread - A New Take on the Classic Pass Labs D1 with an ESS Dac
So, I have some decisions about what to do:
1) Literally swap out the 9018 boards for 9028 boards and deal with the mismatched DAC board problem that Russ White mentioned. I'm not keen on using the limiting function of the new DACs and having lived with the mismatching issue in the 9018 boards for so long (and not been bothered) I'm OK with that. I might need to tweak the heat sinks on my Paul Hynes 'tridents' but I think they'll be able to deliver enough current.
2) Remove the dual mono DAC boards and replace with a 9038. presumably, this would require a further reduction in the 4 resistors (R3& R4 to around 45R (Caddock MP930) and also R1&2 to around 100) and also a reduction in the power supply to the board to around 20-25V (these are all estimates only). This is quite a bit of work but I would then have the best board and would reduce the complexity of my DAC significantly. The digital signal benefited from the shortest possible path-length when I was doing initial tests but in the final case I needed to use slightly longer lengths of coax - not night and day differences but certainly noticeable.
3) Buy a 9038 board and the new mercury IV. This would certainly end up at the simplest configuration and I could probably re-use the power supplies from the NTD1 with some adjustment. I read that barrows is going to do this. I am also pretty sure that SCompRacer had a dual mono 9018 build and it looks like he is going for a single, stereo 9028.
4) Ask Russ White to progress his Dual mono, single PCB board DAC!
So...some decisions!
It looks like 4 is out for the time being. 1 is relatively easy but I'm drawn to the best board! I'll need to talk to some NTD1 owners who've heard the Mercury.
Any thoughts, input etc very welcome!
Crom
Mercury vs. NTD-1 is going to be different for sure... It is hard to pin down these types of things, and I have no preamp to enable fast switching A/B comparisons here...
I am liking the total of 9038/Mercury better than 9018/NTD-1, but not necessarily in every sonic parameter. It is also still early, and I need to experiment with power supplies for the Mercury and small tweaks.
If pushed to say something right now, i would say that perhaps 9018/NTD-1 has a slight advantage in "natural" tone (maybe?), but that 9038/Mercury excels at detail retrieval, ultra low noise, and dynamics... But I am reaching a bit trying to come up with answers right now.
BTW, I oversample all PCM to DSD, and native DSD files play at their native rate, so that may be a factor as well. (I prefer ESS DACs with DSD).
I am liking the total of 9038/Mercury better than 9018/NTD-1, but not necessarily in every sonic parameter. It is also still early, and I need to experiment with power supplies for the Mercury and small tweaks.
If pushed to say something right now, i would say that perhaps 9018/NTD-1 has a slight advantage in "natural" tone (maybe?), but that 9038/Mercury excels at detail retrieval, ultra low noise, and dynamics... But I am reaching a bit trying to come up with answers right now.
BTW, I oversample all PCM to DSD, and native DSD files play at their native rate, so that may be a factor as well. (I prefer ESS DACs with DSD).
Thanks Barrows for the input, sad (for us 
) you did not bought the 9028 so you could have dropped it in the NTD1 and compare BIII themselves. As this is what I've asked to several people around; it seems the 9028 alone (Legato tested) is all you said (better details, dynamics...) than the 9018 and we just can't guess on how much each part plays in SQ. Tridents versions, older vs SR vs Haynes vs Dimdim etc... also into account is the NTD1 version but in that case you have the latest one, PSU included? Add to that what the 9038's spice added to the 9028... and your DSD conversion... DIY is "haute couture" =) No one size fits all here!
My guess, and it is only that for now, is that the very best will be top PSU with BIII9028 with NTD1 V4 & PSU will at least sound as good as full TPA BIII9028, if not better.
So Crom for me in your case it's options 1) or 3)
Anyway, keep these NTD1s alive and this thread also guys!
side note: few pics of your NTD1 maybe?
) you did not bought the 9028 so you could have dropped it in the NTD1 and compare BIII themselves. As this is what I've asked to several people around; it seems the 9028 alone (Legato tested) is all you said (better details, dynamics...) than the 9018 and we just can't guess on how much each part plays in SQ. Tridents versions, older vs SR vs Haynes vs Dimdim etc... also into account is the NTD1 version but in that case you have the latest one, PSU included? Add to that what the 9038's spice added to the 9028... and your DSD conversion... DIY is "haute couture" =) No one size fits all here!My guess, and it is only that for now, is that the very best will be top PSU with BIII9028 with NTD1 V4 & PSU will at least sound as good as full TPA BIII9028, if not better.
So Crom for me in your case it's options 1) or 3)
Anyway, keep these NTD1s alive and this thread also guys!
side note: few pics of your NTD1 maybe?
Last edited:
Yup, it is usually impossible to do true apples-apples comparisons! Especially with the 9038 as it requires a different output stage. But I look at the specs of the 9038 and know it is better than 9018 (and the ESS guys use a golden eared listening panel to inform their changes in addition to measurements), so now the task is: how to get the best out of the 9038. I know Russ worked on hard on this, and Mercury is the current result, so that is what I am working with.
My NTD-1 is V.4 with power supply, and Clarity CMR 3.0 µF coupling caps. The coupling caps are another variable, as I tried a few different approaches before settling on the CMRs. Each different cap choice changes the sound, transparency, and tonal qualities. At least with Mercury I need not worry about that!
Looks like there is no way to directly add photos here without first hosting them somewhere. That gets kind of convoluted. I did take a quick iPad snap of it for you, so if you want, PM me your e-mail address and I can send it out.
My NTD-1 is V.4 with power supply, and Clarity CMR 3.0 µF coupling caps. The coupling caps are another variable, as I tried a few different approaches before settling on the CMRs. Each different cap choice changes the sound, transparency, and tonal qualities. At least with Mercury I need not worry about that!
Looks like there is no way to directly add photos here without first hosting them somewhere. That gets kind of convoluted. I did take a quick iPad snap of it for you, so if you want, PM me your e-mail address and I can send it out.
Last edited:
Damn, another variable in setup!
Strange I can upload pics with the "Manage attachments" box under the text window, maybe system/browser dependant.
In short: 9028 with NTD1 is a nice working option, but a project with 9038 calls for Mercury or to go for unknown land where success is all but guaranteed...
Crom, as a "waiting while thinking" option, which I consider, why not try Trident SR on AVCC, as Russ said this is an audible upgrade?
Thanks for input guys.
Yes, Malefoda, I agree that 2 is probably a non-starter. I have taken a lot of time and effort optimising my NTD1 and it wasn’t really designed to be easily separated from the heatsink. I will end up breaking something! I’ve completely changed the power supply that I felt was the weak link in the design and my output caps have been upgraded a number of times. They are now Duelunds. I’ll post some pics at some point although it is nothing special to look at. It lives in a (well-aired) cupboard and I just switch it on - it does its job very well.
1 is definitely a possibility although I won’t be able to go back and forth between the 9018 and 9028 because I will need to remove the PH regs and re-use. Option 3 sounds interesting but I agree with Barrows that the power supply will need swapping to be a fair comparison with what I have. I would most probably then end up with 2 dacs to compare.
Good idea re upgrading the tridents Malefoda – power upgrades will definitely yield positive results. I’ve already replaced all the tridents with Paul Hynes three-pin regs and I haven’t found anything better than these! The only thing I think that I can do to further optimise the power to my dac boards is to individually power the AVCC reg (and possibly the others) from separate power supplies (rather than taking the input from the board). Given that I’m dual mono, that means at least 2 more regs and 2 more transformers (currently my DAC has 11 transformers I think!!) which I don’t have space for in the case at the minute.
To further muddy the water, I’ve had my interest piqued by the soekris R2R dac and I’m just reading up on that. I wonder whether it would be possible to run one of those boards balanced (without the buffers that are currently there – they do sell a balanced retail version) and run it into the NTD1. I wonder how R2R plays up against the 9028/38. Supporters do keep banging on about less clinical and more musical so it might be worth investigating!
Yes, Malefoda, I agree that 2 is probably a non-starter. I have taken a lot of time and effort optimising my NTD1 and it wasn’t really designed to be easily separated from the heatsink. I will end up breaking something! I’ve completely changed the power supply that I felt was the weak link in the design and my output caps have been upgraded a number of times. They are now Duelunds. I’ll post some pics at some point although it is nothing special to look at. It lives in a (well-aired) cupboard and I just switch it on - it does its job very well.
1 is definitely a possibility although I won’t be able to go back and forth between the 9018 and 9028 because I will need to remove the PH regs and re-use. Option 3 sounds interesting but I agree with Barrows that the power supply will need swapping to be a fair comparison with what I have. I would most probably then end up with 2 dacs to compare.
Good idea re upgrading the tridents Malefoda – power upgrades will definitely yield positive results. I’ve already replaced all the tridents with Paul Hynes three-pin regs and I haven’t found anything better than these! The only thing I think that I can do to further optimise the power to my dac boards is to individually power the AVCC reg (and possibly the others) from separate power supplies (rather than taking the input from the board). Given that I’m dual mono, that means at least 2 more regs and 2 more transformers (currently my DAC has 11 transformers I think!!) which I don’t have space for in the case at the minute.
To further muddy the water, I’ve had my interest piqued by the soekris R2R dac and I’m just reading up on that. I wonder whether it would be possible to run one of those boards balanced (without the buffers that are currently there – they do sell a balanced retail version) and run it into the NTD1. I wonder how R2R plays up against the 9028/38. Supporters do keep banging on about less clinical and more musical so it might be worth investigating!
A nice DAC and other output caps to try
We already are high-end here, how much for the same in shops? A lot of money but hey, we want higher-end =)
Another input on ES9018 vs ES9028Pro, even more with new Tridents/regs. Should sound terrific on NTD1!
http://www.diyaudio.com/forums/digi...lo-iii-upgraded-es9028pro-10.html#post5391869
Time to build guys (if $$$...)
We already are high-end here, how much for the same in shops? A lot of money but hey, we want higher-end =)
Another input on ES9018 vs ES9028Pro, even more with new Tridents/regs. Should sound terrific on NTD1!
http://www.diyaudio.com/forums/digi...lo-iii-upgraded-es9028pro-10.html#post5391869
Time to build guys (if $$$...)
Yes...
I work in the Industry and have some ideas on retail pricing. A commercial version of this build would have to go for at least $10K in a half decent chassis, perhaps more. Of course you could do some shortcuts such that you did not need so many separate PCBs without hurting performance.
I want to try BuffPRO 9028 with it, but no time or $ for that now. As I already have my BuffPRO 9038 I am working on, and a DSC-2 to build this summer...
A nice DAC and other output caps to try
We already are high-end here, how much for the same in shops? A lot of money but hey, we want higher-end =)
Another input on ES9018 vs ES9028Pro, even more with new Tridents/regs. Should sound terrific on NTD1!
http://www.diyaudio.com/forums/digi...lo-iii-upgraded-es9028pro-10.html#post5391869
Time to build guys (if $$$...)
I work in the Industry and have some ideas on retail pricing. A commercial version of this build would have to go for at least $10K in a half decent chassis, perhaps more. Of course you could do some shortcuts such that you did not need so many separate PCBs without hurting performance.
I want to try BuffPRO 9028 with it, but no time or $ for that now. As I already have my BuffPRO 9038 I am working on, and a DSC-2 to build this summer...
A quick update: Following a chat with Brian yesterday, I've ordered two 9028 boards from Twisted Pear. This happily coincided with a post from Russ releasing a dual mono firmware. The process to convert the boards to dual mono looks pretty much like it was for the BIIISE so that part of the journey should be relatively straight forward.
So Malefoda, option 1 it is ;-)
As far as I am aware I can keep the output stage and the power supplies completely identical to the 9018 boards and I should be able to play at least red book PCM material. I may end up having to heatsink the on-board regs if I want to play higher res/DSD material but we will see where we get to.
My current BIIISE's take their clock signal from Ian's McFifo board. Given Russ's comments about this not being worth the effort with the new boards, I will probably not bother with this step - I haven't decided yet. One other thing to investigate is whether I can feed the clock signal back from the McFifo to my Sonore USB card. I never got around to looking at this.
I will need to adapt the output cables running between the DAC boards and the NTD1 board as Brian tells me that the output phases are not reversed (as they were with the BIIISE boards).
I'll post some pics when I get a chance to start.
So Malefoda, option 1 it is ;-)
As far as I am aware I can keep the output stage and the power supplies completely identical to the 9018 boards and I should be able to play at least red book PCM material. I may end up having to heatsink the on-board regs if I want to play higher res/DSD material but we will see where we get to.
My current BIIISE's take their clock signal from Ian's McFifo board. Given Russ's comments about this not being worth the effort with the new boards, I will probably not bother with this step - I haven't decided yet. One other thing to investigate is whether I can feed the clock signal back from the McFifo to my Sonore USB card. I never got around to looking at this.
I will need to adapt the output cables running between the DAC boards and the NTD1 board as Brian tells me that the output phases are not reversed (as they were with the BIIISE boards).
I'll post some pics when I get a chance to start.
Last edited:
Yes but don't hold your breath, I hardly get any time to play with this stuff nowadays ;-)
re your Q. Not sure - I got the info from Russ's post here: Introducing the Buffalo III-SE-Pro 9028/9038
re your Q. Not sure - I got the info from Russ's post here: Introducing the Buffalo III-SE-Pro 9028/9038
Guys,
I'm out of steam when it comes to the circuit itself (V4 in my case), and my memory is a black hole.
There was that shet to calculate power and bias current, ok, but when I put what I do have, 200R/470R I got a lower 85mA when it seems best is around 100mA. It may change gain also. Anyone can explain me the parameters we balance here again?
Thanks
Matthieu
I'm out of steam when it comes to the circuit itself (V4 in my case), and my memory is a black hole.
Owen about the V3 said:
There was that shet to calculate power and bias current, ok, but when I put what I do have, 200R/470R I got a lower 85mA when it seems best is around 100mA. It may change gain also. Anyone can explain me the parameters we balance here again?
Thanks
Matthieu
Hi Malefoda, I've only been scanning threads and maybe I missed something but I'm not sure what you're asking. Happy to try to help if I can.
Right, I've been busy! I ended up going for option 1 which was a literal swapout of the 9018 boards for 2 x 9028pro boards. Russ and Brian produced a test dual mono firmware which Brian kindly included with my order. It didn't take me long to do the board exchange although I needed some help from Russ as I didn't realise that I needed to connect the DAC reset pins together along with the other tweaks to get dual mono to work.
I was worried when Russ stated that boards can have up to 14% differences in output levels. However, subjectively this doesn't appear to be a problem.Maybe I was lucky.
The DAC itself is now made up of:
Sonore USB board > Ian's McFIFO with Pulsar clocks > 2 x 9028Pro boards > NTD1 > duelund coupling caps
Everything is powered by Paul Hynes regs.
The hardware in this listening test was a constant except the DAC boards. Some listening was done at 44.1 but mostly PCM oversampled to 352/382 using the (excellent) upsampling facility in the USB card. Also, all the DAC board filters etc have been left at default for now.
I was hoping to finally be able to play some DSD given that now the FIFO is DSD-compatible and the DAC boards are effectively DSD plug-n-play but for some reason I cannot get any DSD to play. I suspect it's an issue with the USB card firmware but I'm investigating that separately.
Anyway, what do I think...
In a word, great! It is not a night-and-day difference but a glare or digital shimmer has been removed from the top end. This takes a little getting used to and a couple of tracks sounded a little bottom-heavy to start off with, but there is now much more nuance and detail across the frequency range. There is an increased precision and immediacy to the sound - I think caused by better transient response. This increases the naturalness of acoustic instrumentsianos, voices, guitars all sound more real - my wife commented that Ben Webster and Oscar Peterson could have been sitting across the room from us. My acid test of some ACDC and Pantera now has increased viscera. The soundstage has increased in width and depth and separation is better...again, not huge changes but noticeable and appreciated. Worth it? Definitely.
At some point I will probably buy a couple of 9038's and build a separate DAC but I want to wait until there is a decent selection of output stages. Perhaps Owen will come out of 'retirement' and upgrade his excellent NTD1. As long as I can get DSD to play then I will have scratched the itch for now and this upgrade will more than suffice for the time being
I was worried when Russ stated that boards can have up to 14% differences in output levels. However, subjectively this doesn't appear to be a problem.Maybe I was lucky.
The DAC itself is now made up of:
Sonore USB board > Ian's McFIFO with Pulsar clocks > 2 x 9028Pro boards > NTD1 > duelund coupling caps
Everything is powered by Paul Hynes regs.
The hardware in this listening test was a constant except the DAC boards. Some listening was done at 44.1 but mostly PCM oversampled to 352/382 using the (excellent) upsampling facility in the USB card. Also, all the DAC board filters etc have been left at default for now.
I was hoping to finally be able to play some DSD given that now the FIFO is DSD-compatible and the DAC boards are effectively DSD plug-n-play but for some reason I cannot get any DSD to play. I suspect it's an issue with the USB card firmware but I'm investigating that separately.
Anyway, what do I think...
In a word, great! It is not a night-and-day difference but a glare or digital shimmer has been removed from the top end. This takes a little getting used to and a couple of tracks sounded a little bottom-heavy to start off with, but there is now much more nuance and detail across the frequency range. There is an increased precision and immediacy to the sound - I think caused by better transient response. This increases the naturalness of acoustic instruments
ianos, voices, guitars all sound more real - my wife commented that Ben Webster and Oscar Peterson could have been sitting across the room from us. My acid test of some ACDC and Pantera now has increased viscera. The soundstage has increased in width and depth and separation is better...again, not huge changes but noticeable and appreciated. Worth it? Definitely.At some point I will probably buy a couple of 9038's and build a separate DAC but I want to wait until there is a decent selection of output stages. Perhaps Owen will come out of 'retirement' and upgrade his excellent NTD1. As long as I can get DSD to play then I will have scratched the itch for now and this upgrade will more than suffice for the time being
Hi Crom and fellows,
my request is for R17/R14 and siblings from Owen's schematic's pdf from here:
Build Thread - A New Take on the Classic Pass Labs D1 with an ESS Dac
I guess these are the "gain resistors" from the wiki here:
opc's NTD1 project
and these values are used in the sheet pointed in the wiki also, computing power dissipation.
And 200R/400R matches with values given by Owen here, 100mA current:
http://www.diyaudio.com/forums/vend...lable-bal-bal-se-se-lpuhp-70.html#post4284231
Maybe someone has links (stupid me, no memory and never bookmarked useful posts!) to how one choose values according to gain/current targeted. I want to understand, without learning all from scratch (transistors and such...), why in my circuit I have 200R/470R, so 85mA, which is less heat but away from the 100mA point (99mA with 200R/400R and 101 with 180R/390R), and maybe some gain issues I don't remember.
Hope that someone can help! That bugs me very much...
Thanks!
Matthieu
my request is for R17/R14 and siblings from Owen's schematic's pdf from here:
Build Thread - A New Take on the Classic Pass Labs D1 with an ESS Dac
I guess these are the "gain resistors" from the wiki here:
opc's NTD1 project
and these values are used in the sheet pointed in the wiki also, computing power dissipation.
And 200R/400R matches with values given by Owen here, 100mA current:
http://www.diyaudio.com/forums/vend...lable-bal-bal-se-se-lpuhp-70.html#post4284231
Maybe someone has links (stupid me, no memory and never bookmarked useful posts!) to how one choose values according to gain/current targeted. I want to understand, without learning all from scratch (transistors and such...), why in my circuit I have 200R/470R, so 85mA, which is less heat but away from the 100mA point (99mA with 200R/400R and 101 with 180R/390R), and maybe some gain issues I don't remember.
Hope that someone can help! That bugs me very much...
Thanks!
Matthieu
