I used to, as the mck always showed up on I2S connections. But in fact on the PCB it is not connected to anything
Hello,
Most companies involved in diy audio suck just like in the real world.
I wish Doede would develop his own " streaming " solution. I mean nothing special but something that will work ( like all his products) and can easily be " combined "
I begin to get bored with some of the companies present here spitting out to many updates and leaving many customers with all kind of issues.
I don't think that the streaming part of digital is more difficult than the dac part.
Greetings Eduard
P.s some solutions offered here by other " companies" remind me of what i see here in North Vietnam
Most companies involved in diy audio suck just like in the real world.
I wish Doede would develop his own " streaming " solution. I mean nothing special but something that will work ( like all his products) and can easily be " combined "
I begin to get bored with some of the companies present here spitting out to many updates and leaving many customers with all kind of issues.
I don't think that the streaming part of digital is more difficult than the dac part.
Greetings Eduard
P.s some solutions offered here by other " companies" remind me of what i see here in North Vietnam
Attachments
For those interested in my latest post on playing around with DACs
check my blog: Link to Blog Post
But most important for posting here:
I would like to write a follow up with a “really good” R2R ladder DAC which claims to have gotten it right in terms of matching resistors.
Anyone out there who would be willing to lend me their R2R DAC for listening comparison and the exact same AP System Two tests?
That would be great! Just write me an E-Mail!
thanks,
Doede
EXPERIMENTS WITH A DISCRETE R2R LADDER DAC (XY2)
check my blog: Link to Blog Post
But most important for posting here:
I would like to write a follow up with a “really good” R2R ladder DAC which claims to have gotten it right in terms of matching resistors.
Anyone out there who would be willing to lend me their R2R DAC for listening comparison and the exact same AP System Two tests?
That would be great! Just write me an E-Mail!
thanks,
Doede
I promised some time go to post on my tests with chokes in the power supply. I just wrote a blog entry on this theme…
Link to blog post
Link to blog post
CLC is almost always better, but crc (and variants) is way less space hungry. So there's that 🙂
Hello,
Of course i am heavily biased because i always recommend people to use a more than decent choke.
I used some products from AE but never an amorphous core choke.
People always say with present day technology there is no need for chokes anymore. BUT there are a few technicians who will tell you a choke can do something where active devices fail.
You used the ll2733 with the coils in series to get the maximum number of Henry. Did you try connecting it in what Lundahl calls common mode. I would like to see if it is any better. The number of Henry stays the same but the filtering is different.
I have a really big Lundahl at home the Ll2771 which has a current rating of 400 mA, dcr 2,8 ohm and around 3500 mH with coils in parallel. I can send it to you first week of March if you like.
At Lundahl you can ask for a different airgap meaning you can get around 1H at 400 mA from the LL2733 as well. I know from using them in choke input that more Henry will do a LOT. Will the core material be the decision maker if the inductance is the same. I remember reading that amorphous core will saturate quicker but Doede ordered the choke for this job
I am a BIG fan of choke input and i will ask Doede to do a simple test to see if this will improve things more.. It is a bit more " tricky " because you must create a minimum current draw all the time and you will need a transformer with higher secundairy voltage. AND you need a quality choke, the Hammond might buzz.
Will you add a second low current choke like you did instead of the 300 ohm resistor on the board?
Greetings,Eduard
Of course i am heavily biased because i always recommend people to use a more than decent choke.
I used some products from AE but never an amorphous core choke.
People always say with present day technology there is no need for chokes anymore. BUT there are a few technicians who will tell you a choke can do something where active devices fail.
You used the ll2733 with the coils in series to get the maximum number of Henry. Did you try connecting it in what Lundahl calls common mode. I would like to see if it is any better. The number of Henry stays the same but the filtering is different.
I have a really big Lundahl at home the Ll2771 which has a current rating of 400 mA, dcr 2,8 ohm and around 3500 mH with coils in parallel. I can send it to you first week of March if you like.
At Lundahl you can ask for a different airgap meaning you can get around 1H at 400 mA from the LL2733 as well. I know from using them in choke input that more Henry will do a LOT. Will the core material be the decision maker if the inductance is the same. I remember reading that amorphous core will saturate quicker but Doede ordered the choke for this job
I am a BIG fan of choke input and i will ask Doede to do a simple test to see if this will improve things more.. It is a bit more " tricky " because you must create a minimum current draw all the time and you will need a transformer with higher secundairy voltage. AND you need a quality choke, the Hammond might buzz.
Will you add a second low current choke like you did instead of the 300 ohm resistor on the board?
Greetings,Eduard
Yes, don't remember if it's was just an OPT, or if it was for some kind of I/V.
Oh, I have found the post https://www.diyaudio.com/community/...1794-and-waveio-usb-input.224108/post-6847393
Oh, I have found the post https://www.diyaudio.com/community/...1794-and-waveio-usb-input.224108/post-6847393
Rearranging my dddac a bit.
I am currently needing single ended sources and amp sensitivity is 0,2V. Is anyone using lower value resistors with 1 deck dddac to lower its output? Is maybe output thd better? Been reading about this in other threads about the pcm1794s - I am thinking to parallel the resistors for 60-65ohm on positive side, would output be halved to 1,4V or so? and would it be ok to ground the negative side?
I am currently needing single ended sources and amp sensitivity is 0,2V. Is anyone using lower value resistors with 1 deck dddac to lower its output? Is maybe output thd better? Been reading about this in other threads about the pcm1794s - I am thinking to parallel the resistors for 60-65ohm on positive side, would output be halved to 1,4V or so? and would it be ok to ground the negative side?
I will come back to this - I did some work on this and need to look up my results. I did not post as it was all not exceptional.
Uhm..this is a special product from Audio Creative, please ask them if this is planned.
@geoturbo : I Posted about this already a few years ago. Yes, it is possible without any problem. in your case where you want 0,2V rms output you have to do the following:
for one Deck:
anyway, you can play around with these two parameters to get any output voltage you want as long as follow these rules of thumb:
hope this helps,
Doede
for one Deck:
- Replace BOTH Rloads 133 Ohm with 36 Ohm. (well actually all four)
- reduce the bias measurement voltage from 40mV to 25mV.
- Do NOT ground one side!
- THD will be roughly 0,1% instead of 0,15%
anyway, you can play around with these two parameters to get any output voltage you want as long as follow these rules of thumb:
- keep the bias below 50mV and higher than 25mV (this a range I did tests with, it will work outside, but I have no proof of correct working)
- keep the calculated product of (Rload x mV-bias) below 5500.
- Example 36 Ohm x 25mV = 36x25=900 so well below 5500. things are ok - you are not using the full output potential of the DAC, but that is what you need right now
hope this helps,
Doede
Last edited:
Hi Doede.
Main characteristic of that kind of topologies that have R-2R laddr at the end is:
1. MPU pack the format. Huge noise, ground bouces etc. of this input unit
2. Digital ICs driving R2R ladder. also has noise and different Rds-on-off values that impact on tolerance of R2R lader...
.
I listened more of that configurations bare boards. And was not satisfied. Main issue was noise that coming from HF into the audio band, giving unlistenable sound signature. for my opp.
Also I compared the same R2R concept with my design without MPU, only ditl ICs and R2R leadder. And was smaller amount of noise... So The MPU in front is significant generator of noise.
There is not much to do about it but it can be corrected anyway.
.
You listened and measured PCBs without any filter ant the end.
The analog multi order LC filter is a must and helps a lot in final audio result.
This is an voltage R2R ladder. THe output resistance is R.
Measure R, or read from component if it is possible, and design LC 5 to 7 order LC
with this value as driving resistance and some choosen R-terminating value.
With this program:
https://markimicrowave.com/technical-resources/tools/lc-filter-design-tool/
legendre is good also Bessel. Set the filter first something with standard values, easy to obtain, L as ferrite beads. And latter You can mount real L with much better Q factor and so.
Goal is to measure and listen with the filter output.
(If You want balanced filter, adopt values to half of that for SE line.)
But from my measurements and listening, balanced is giving more noise. It is not as analog ball for canceling... It is different kind of noise...
.
Add 10-47uF at decouple smd capacitors of digital ICa driving r2r ladder. On top of existing smd Cs. And for MPU if it is possible. Does not matter the "brand" 🙂
.
I used some buffers after passive filter. One strange thing was that simple Germanium buffer (after the passive filter) moved the sound quality a lot. JFET was also very good...
.
There is a need for buffer because even with lower values of r2r, it is still range in Kohms output and it is high output resistance to drive next component... You listened PCBs just with Co
.
I tried also with batteries, and there was not much improvement. But it was more convenient to manipulate boards...
.
Use "shunt first" option because of the lower values of L easier to find. The main factor for filter to design is R value of ladder as output resistance of unit. If it is higher then values of L is higher too 🙁
.
Example of 7 order Bessel filter 10K Rinput=Rladder, 100K output, Fo=50KHz, shunt first type.
This kind of DACs has alvays DC offset at the output. Typically of Voffset=1/2 Vdigital ICs supply.
Also typicall value of R2R ladder in theese dacs are 2R=10K, R=5K (mostly 10K in parallel)
These higher values of Rout is excluding direct use of transformer at the output, becaue it has to be from huge Lprim to satisfy just moderate LF range 🙁
Also typicall value of R2R ladder in theese dacs are 2R=10K, R=5K (mostly 10K in parallel)
These higher values of Rout is excluding direct use of transformer at the output, becaue it has to be from huge Lprim to satisfy just moderate LF range 🙁
Last edited:
this is something for about 5K output R, 25KHz, 7 order Bessel, shunt first.
For check with standard Values of FBs 10mH, 22mH, 33mH
I can recommend our member @abraxalito, with huge experience about passive filters, to consult? He has already made units. I think that He can even make custom very good L for these purposes. 🙂
For check with standard Values of FBs 10mH, 22mH, 33mH
I can recommend our member @abraxalito, with huge experience about passive filters, to consult? He has already made units. I think that He can even make custom very good L for these purposes. 🙂
Last edited:
Hi Zoran, thanks for the extensive reply! I did not use a 7th order filter, but I must say, that my analog chain is deliberately limited to 50kHz, so I do not expect too much effect. Building this filter in practice is quite a thing looking at the 10mH to 33mH inductors by the way....
But, I will measure again with the AP-cut-off filters in different positions (5-pole 22k 3-pole 30k&80k and flat to 500k Hz) and see if there is mensurable result in terms of extra noise and linearity. I still have a cut off filter from analog DSD experiments. I could use that to see if there still an audible difference.
also, your ideas to tweak the Ali DAC with components and buffers and see what works could be a nice rainy Sunday afternoon project 😎
BUT....
Highest priority is now finishing off the DDDAC1794MK3 project. First batch of DAC and PSU boards are currently at the SMD assembly company, and I will build two working models with both complete different set of electrolyte capacitors in the analog section of the DAC ("normal" and "organic") to see (listen) what the sonic impact is. Goals is to decide if it will be the one or the other or offering two options: Cap-kit-1 and cap-kit-2 for personal taste of the sonic result. Not sure, will see what comes.... Expect mid to end of march (need to go skiing first 😉 )
But, I will measure again with the AP-cut-off filters in different positions (5-pole 22k 3-pole 30k&80k and flat to 500k Hz) and see if there is mensurable result in terms of extra noise and linearity. I still have a cut off filter from analog DSD experiments. I could use that to see if there still an audible difference.
also, your ideas to tweak the Ali DAC with components and buffers and see what works could be a nice rainy Sunday afternoon project 😎
BUT....
Highest priority is now finishing off the DDDAC1794MK3 project. First batch of DAC and PSU boards are currently at the SMD assembly company, and I will build two working models with both complete different set of electrolyte capacitors in the analog section of the DAC ("normal" and "organic") to see (listen) what the sonic impact is. Goals is to decide if it will be the one or the other or offering two options: Cap-kit-1 and cap-kit-2 for personal taste of the sonic result. Not sure, will see what comes.... Expect mid to end of march (need to go skiing first 😉 )
Last edited:
Hello Doede,
Still in South East Asia but as far as i know there is not that much snow for skiing.
Here however temperatures will go up again in the next few days and i was told they will go up a little more in March before reaching insane levels in April. Have been here nearly 3 months and it is hard to imagine people working here at all.
Maybe i should pre order the boards because i have full confidence.
However i will try choke input. Will send you my big choke to try for choke input and pay for transport. Switching to choke input just requires a different transformer ( you have more than me) and if executed correctly with a load resistor ( and a zenerdiode, idea from Guide Tent to limit voltage to a safe voltage for the caps) it will be a cheap experiment. If you don't do it i will haha.
Greetings, Eduard
Still in South East Asia but as far as i know there is not that much snow for skiing.
Here however temperatures will go up again in the next few days and i was told they will go up a little more in March before reaching insane levels in April. Have been here nearly 3 months and it is hard to imagine people working here at all.
Maybe i should pre order the boards because i have full confidence.
However i will try choke input. Will send you my big choke to try for choke input and pay for transport. Switching to choke input just requires a different transformer ( you have more than me) and if executed correctly with a load resistor ( and a zenerdiode, idea from Guide Tent to limit voltage to a safe voltage for the caps) it will be a cheap experiment. If you don't do it i will haha.
Greetings, Eduard