, yes, they are not the best for sure . But they are 15x cheaper than crystek. Good for people with low budget, like me Qusp , I have one doubt about the need of the voltage offset in current mode output, in the datasheet of the es9018 they use the voltage offset . But some users and the twisted pear I/V convertors have the inputs referenced to ground, do you know more about this issue ? Thanks for any information about this matter.
Last edited:
@redrabbit, i'm currently using
XG-1000CA/CB | Real time Clock Module | Products | Epson Quartz Crystal Devices
frequency @ 75 MHz in async with an amanero, does sound good listening @ DSD 5.6 MHz, even if it just an upsample
Q: What did you put in front of those TPS eval boards ? Any experience you'd like to share ?
I'm feeding my DAC with seriel 78xx regs + LM1117 for both DVCC & AVCC, I guess I've got some room for improvement ....
XG-1000CA/CB | Real time Clock Module | Products | Epson Quartz Crystal Devices
frequency @ 75 MHz in async with an amanero, does sound good listening @ DSD 5.6 MHz, even if it just an upsample
Q: What did you put in front of those TPS eval boards ? Any experience you'd like to share ?
I'm feeding my DAC with seriel 78xx regs + LM1117 for both DVCC & AVCC, I guess I've got some room for improvement ....
using the XG-1000CA/CB | Real time Clock Module | Products | Epson Quartz Crystal Devices
sounds quite nice, even on 78xx + lm1117 regs ....
sounds quite nice, even on 78xx + lm1117 regs ....
An externally hosted image should be here but it was not working when we last tested it.
Here are some pictures of my V3 board using an older set of TP Tridents and AVCC module. I cut off the IV section since I am going to test this with a tube output stage using passive IV. I did run it with the stock IV for a while. It was pretty darn good for the money invested.
This has the AVCC separated as shown on HiFiDuino. Its not pretty, but it works for my needs. I am using this for testing before I risk doing any work on my BIII. Sorry for the lousy pictures, I am supposed to be working.
This has the AVCC separated as shown on HiFiDuino. Its not pretty, but it works for my needs. I am using this for testing before I risk doing any work on my BIII. Sorry for the lousy pictures, I am supposed to be working.
Attachments
wow, that looks like hardcore DIY, after a 30 yrs break from electronics I wouldn't trust myself in doing that already, if ever ...
@ 1st I had wrongly mounted my x-tal just b4 I wanted start ...
luckely my bro had his eyes open ... after a quick destroy and replace (I had the luck having another x-tal laying around) and had to burn a couple of parts but it looked not only better
it just played instantly into my regular Denon amps ... wow ... that's what they mean with music ... I can't still get my ears of it ... this is cool, very cool ... I can hear details in the music I didn't hear b4 and my name is not Luke ... probably never listened good enough that's what my old man used to say as well but I think he meant something different ...
This is all hand soldered, even the 0603 C1, so excuse my not perfect soldering look ... the tip of the soldering iron was almost bigger then C1 ...
Any recommendation about a good AVCC (except the TP stuff)
Is Salas a good candidate ?
Any other ???
@ 1st I had wrongly mounted my x-tal just b4 I wanted start ...
luckely my bro had his eyes open ... after a quick destroy and replace (I had the luck having another x-tal laying around) and had to burn a couple of parts but it looked not only better
it just played instantly into my regular Denon amps ... wow ... that's what they mean with music ... I can't still get my ears of it ... this is cool, very cool ... I can hear details in the music I didn't hear b4 and my name is not Luke ... probably never listened good enough
that's what my old man used to say as well but I think he meant something different ... This is all hand soldered, even the 0603 C1, so excuse my not perfect soldering look ... the tip of the soldering iron was almost bigger then C1 ...
Any recommendation about a good AVCC (except the TP stuff)
Is Salas a good candidate ?
Any other ???
The TPS7A47 has very low noise 4uv , NCP700B has also low noise (10uv) and easy to use, you can use a pré-regulator with a output voltage of 5 volts and then a NCP700B 3.3v .
Despite the fact that shunt regulators have lower efficiency . But a good shunt regulator like Salas is hard to beat, in terms of noise and psrr.
also see the :
Services
lots of information about regulators . Like this one
http://waltjung.org/PDFs/Improved_PN_Regs.pdf
Despite the fact that shunt regulators have lower efficiency . But a good shunt regulator like Salas is hard to beat, in terms of noise and psrr.
also see the :
Services
lots of information about regulators . Like this one
http://waltjung.org/PDFs/Improved_PN_Regs.pdf
Ok, I have already find the reason for using the offset voltage. Thanks anyway
yeah sorry about that, didnt check back till now; its really pretty straight forward as you found out. it nulls the offset from having a +VCC only supply, it also means less unwanted current flows in that AVCC node, meaning lower distortion
Last edited:
es9018 has a strange output behavior. It looks like a voltage source followed by a resistor. If is this the case , I fail to understand how working in Current mode lower the distortion. I need to investigate further , so I bought one of this dac to do some tests, and also test my new i/v converter for the es9018/12.
one bad thing of using the vcc/2 in the I/v converter of the datasheet is that without the output offset current the opamps will work in class b , if one use the reference to ground the offset current will force the opamps in class A. In my type of I/v converter the absence of the offset current is an advantage but when using opamps is better to have a offset current.
i've never used the offset nulling with opamps and es9018, only with a common gate follower and jfet based. the fact is, it DOES lower distortion and noise, but its not critical if it matches AVCC/2 exactly. the most important factor for distortion performance is low impedance, low noise AVCC supply/layout and unimpeded low impedance connection (signal and return) to the IV convertor, which should also be high gm to complete that picture. the performance difference of 'current mode' vs 'voltage mode' is at least as large as the datasheet says, any research/measurements i've seen to do with this have been more profound if anything
Last edited:
My Diyinhk es9018 + amanero + hifiduino combo is playing for a week. Digital power supply is LT1763 + LM317 (enhanced) as preregulator. It started with another LT1763 for analog power supply, but changed it for a pooge derivate with LT1963 as a prereg. The difference between LT1763 and pooge is quite big - former being hard and glassy in comparation. I started with current mode opamp output stage (ada4898+LM49990) but to my ears passive voltage out with Cinemag cmli15 output xfomers (1:1) sounded better so that's what I'm listening now. Oscillator is 100MHz SAW. 74hcu04 is used for spdif to ttl conversion. Good quality components are used but nothing special except 0402 smd caps directly on ps pins of es9018.
It sounds nice.
Thanks to diyinhk, glt & amanero!
It sounds nice.
Thanks to diyinhk, glt & amanero!
Attachments
My Diyinhk es9018 + amanero + hifiduino combo is playing for a week. Digital power supply is LT1763 + LM317 (enhanced) as preregulator. It started with another LT1763 for analog power supply, but changed it for a pooge derivate with LT1963 as a prereg. The difference between LT1763 and pooge is quite big - former being hard and glassy in comparation. I started with current mode opamp output stage (ada4898+LM49990) but to my ears passive voltage out with Cinemag cmli15 output xfomers (1:1) sounded better so that's what I'm listening now. Oscillator is 100MHz SAW. 74hcu04 is used for spdif to ttl conversion. Good quality components are used but nothing special except 0402 smd caps directly on ps pins of es9018.
It sounds nice.
Thanks to diyinhk, glt & amanero!
Good job!, especially soldering those 0402 caps. Are you using this pooge:
http://www.andrewweekes.talktalk.net/pooge_551.htm ?
Perhaps diyinhk should have 0402 caps per-installed in all of the power pins in his next version of the board.
Last edited:
Hi there,
how did you hook up that display, grasping up the right data to display from the diyinhk board ?
TIA,
Peter.
Everything is explained on glt's site.
I didn't use level converter for I2C connection, just 10k resistors on sda and scl lines. Tried ADUM isolator, but couldn't get it to work - as soon as I connected ground to dac side of isolator arduino would go berzerk.
Yes, that's the schematic - I used 1n4148 for reference, jfet for simple current source (I know bjt would be better) and adjusted the values for 3v3 output.
0402 caps are a bit larger than pins spacing so it's really tacky job to solder them without shorting anything
; for proper job PCB revision would be needed.
Troublesome night
Crying out for some help here after some desperate nights ...
I've been wondering if someone had the same experienced b4.
Build my DIYINHK board, added an Amanero and plugged it in into a W7 Laptop downloaded and configured foobar and audiogate. Got Audiogate to play DSD, 96, 48 and some others of which the names may nt be mentioned here.
After playing a couple of days I thought let's ask the DAC what sample rate it is playing. This can be read in registers 31-28 according to the datasheet and a lot of info I got from glt's hifiduino site although I do use another MCU. I even succeeded in that and displayed the I2S samplerate on a 3.2'' LCD from 4D-Systems.
Now there is where the fun started, it always replied 44100 rounded and calculated with single floats, even playing a 48 or 96 kHz sampled piece of music. Source verified, it is 48 or 96. The DAC insisted in replying 44100. Now I found in the W7 sound devices, the one can setup the sampling rate/bits per sample to use when playing in shared mode. This can only be set to one rate/bits per sample from a list of possibilities. I got as high as 48000 and 24 bits per sample anything higher went to the situation that the DAC is not being able to get a lock on the I2S signal coming out of the Amanero (driver 1.0.56).
Any ideas, tip or tricks
Also, the highest sr/bps selectable is 192/24 and as far as I know, the Amanero is capable of more (384/24 ?). Any tips are welcome here as well.
This pmo (excuse my english) real time, since I hoped listening to some quality music here, I was listening to f..g 44.1 that was not really the expectation I was going for. So I turned to foobar and set the thing up described on the hifiduino side as well, same story, plays all well if I do not exceed 48 kHz sr higher rates get same behavior -> no lock on signal.
It is even getting worse, trying to play 2.8 DSDs. played alright but only with a very disturbing undertoon like from a bad psu. That never happened using Audiogate. DAC clock is 75 MHz.
Any help from you guys will be much appreciated, and if you once get nearby my place you are in for some beers and if the season allows some nice bbq taste ....
Crying out for some help here after some desperate nights ...
I've been wondering if someone had the same experienced b4.
Build my DIYINHK board, added an Amanero and plugged it in into a W7 Laptop downloaded and configured foobar and audiogate. Got Audiogate to play DSD, 96, 48 and some others of which the names may nt be mentioned here
. After playing a couple of days I thought let's ask the DAC what sample rate it is playing. This can be read in registers 31-28 according to the datasheet and a lot of info I got from glt's hifiduino site although I do use another MCU. I even succeeded in that and displayed the I2S samplerate on a 3.2'' LCD from 4D-Systems.
Now there is where the fun started, it always replied 44100 rounded and calculated with single floats, even playing a 48 or 96 kHz sampled piece of music. Source verified, it is 48 or 96. The DAC insisted in replying 44100. Now I found in the W7 sound devices, the one can setup the sampling rate/bits per sample to use when playing in shared mode. This can only be set to one rate/bits per sample from a list of possibilities. I got as high as 48000 and 24 bits per sample anything higher went to the situation that the DAC is not being able to get a lock on the I2S signal coming out of the Amanero (driver 1.0.56).
Any ideas, tip or tricks
Also, the highest sr/bps selectable is 192/24 and as far as I know, the Amanero is capable of more (384/24 ?). Any tips are welcome here as well.
This pmo (excuse my english) real time, since I hoped listening to some quality music here, I was listening to f..g 44.1 that was not really the expectation I was going for. So I turned to foobar and set the thing up described on the hifiduino side as well, same story, plays all well if I do not exceed 48 kHz sr higher rates get same behavior -> no lock on signal.
It is even getting worse, trying to play 2.8 DSDs. played alright but only with a very disturbing undertoon like from a bad psu. That never happened using Audiogate. DAC clock is 75 MHz.
Any help from you guys will be much appreciated, and if you once get nearby my place you are in for some beers and if the season allows some nice bbq taste ....
Crying out for some help here after some desperate nights ...
I've been wondering if someone had the same experienced b4.
Build my DIYINHK board, added an Amanero and plugged it in into a W7 Laptop downloaded and configured foobar and audiogate. Got Audiogate to play DSD, 96, 48 and some others of which the names may nt be mentioned here
After playing a couple of days I thought let's ask the DAC what sample rate it is playing. This can be read in registers 31-28 according to the datasheet and a lot of info I got from glt's hifiduino site although I do use another MCU. I even succeeded in that and displayed the I2S samplerate on a 3.2'' LCD from 4D-Systems.
Now there is where the fun started, it always replied 44100 rounded and calculated with single floats, even playing a 48 or 96 kHz sampled piece of music. Source verified, it is 48 or 96. The DAC insisted in replying 44100. Now I found in the W7 sound devices, the one can setup the sampling rate/bits per sample to use when playing in shared mode. This can only be set to one rate/bits per sample from a list of possibilities. I got as high as 48000 and 24 bits per sample anything higher went to the situation that the DAC is not being able to get a lock on the I2S signal coming out of the Amanero (driver 1.0.56).
Any ideas, tip or tricks
Also, the highest sr/bps selectable is 192/24 and as far as I know, the Amanero is capable of more (384/24 ?). Any tips are welcome here as well.
This pmo (excuse my english) real time, since I hoped listening to some quality music here, I was listening to f..g 44.1 that was not really the expectation I was going for. So I turned to foobar and set the thing up described on the hifiduino side as well, same story, plays all well if I do not exceed 48 kHz sr higher rates get same behavior -> no lock on signal.
It is even getting worse, trying to play 2.8 DSDs. played alright but only with a very disturbing undertoon like from a bad psu. That never happened using Audiogate. DAC clock is 75 MHz.
Any help from you guys will be much appreciated, and if you once get nearby my place you are in for some beers and if the season allows some nice bbq taste ....
I'd recommend you download and install the correct Amanero's ASIO driver (the one that matches your Windows 7 version -> 32 or 64bit) which you can find here:
Index of /drivers/asio
You may not have configured Foobar and Audiogate to use ASIO or WASAPI which are needed to get bit perfect.
For Foobar setup you can try this guide:
Foobar 2000 for Dummies (Part 1) | DIY-Audio-Heaven
Foobar 2000 for Dummies (Part 2) | DIY-Audio-Heaven
Check you have selected ASIO or WASAPI as "Driver Type" in Audiogate and Combo384 in the device drop down selection list, Sample rate as "Automatic" though I'm not sure it is capable of sending native DSD to the sound card.
- Status
- Not open for further replies.