What happens if I see that new resistor as the input impedance of the amp instead of output impedance of the DAC?
How is it? The cap is not on the actual output of the DAC.. as in "last component"? Because if it where, the 626 ohm ladder would be the caps driving impedance and not what it sees?
I became confused now... help!
//
Attachments
Last edited:
Ahh the right thread 
I didn't realise there was an updated version of this (facepalm).
I just need to sweet talk the Mrs into allowing me to get a 1941 1% board.. Given that a decent Xmos board is a few quid, having twice the ladders..
I looked at designing & building an R2R from scratch before finding the earlier board... my only annoyance is that it has a Spartan6 which is very minimal resources compared to an Artix7A with additional DDR for example. However nothing stopping some PC side audio processing before it gets sent over i2s with FIR1 & FIR2 set to bypass.
I'm just writing a Python Jupyter Notebook for modelling filters as we speak.
I didn't realise there was an updated version of this (facepalm).I just need to sweet talk the Mrs into allowing me to get a 1941 1% board.. Given that a decent Xmos board is a few quid, having twice the ladders..
I looked at designing & building an R2R from scratch before finding the earlier board... my only annoyance is that it has a Spartan6 which is very minimal resources compared to an Artix7A with additional DDR for example. However nothing stopping some PC side audio processing before it gets sent over i2s with FIR1 & FIR2 set to bypass.
I'm just writing a Python Jupyter Notebook for modelling filters as we speak.
Thanks, Peter.
I have looked at ALL of your posts many times. The first one is ambiguous at best! The next one makes it clear but then (without gong to look) I swear there was a photo of the board with only four of the BG .1uF caps somewhere.
I think TNT was writing about the 100R resistor I thought I would need to attenuate the DAC.
My SOEKRIS is driving a pair of xilica XP2040s not power amps.
I am going to use about 12 db attenuation at the amps and then split the difference between the DAC and the xilica for further attenuation. The amplifiers that handle the bass have pots so that part is easy. I am using the raw balanced output into the xilica's balanced input. I have found there is no need for the ground wire.
I would like to use Mr. Yazaki's SOSHIN silver micas - but eight would cost much more than the board and would be very difficult to mount.
I have to say Peter is absolutely right about those polymer caps. I have never liked them and have tried to like them since they offer so many packaging advantages but they do something that is not attractive to my ears every time i have tried them.
After listening for many hours after writing I have to say this is a completely different device with those things replaced. It is more what I dreamed it would sound like. Of course, components NEVER sound like you would "dream" they sound but it very fine when they come close. I do not doubt getting that FET out of the way makes some difference but nothing compared to the caps.
I found there was no need to attempt to de-solder the FETs - just a pair of small dykes
cutting through the middle GENTLY and they pop right off. No need for any heat which is always a plus.
SOEKRIS, I know this stuff gets on your nerves and there is no question this is a fine product as shipped but I am grateful there are a few paths to making it really good. I do realize there are folks who may actually prefer the sound of the polymers. I am as baffled by the possibility of that as i am by the lure of "high resolution" digital audio with all of its unremovable artifacts and digital atmospherics that I guess some people think should be there.
I have looked at ALL of your posts many times. The first one is ambiguous at best! The next one makes it clear but then (without gong to look) I swear there was a photo of the board with only four of the BG .1uF caps somewhere.
I think TNT was writing about the 100R resistor I thought I would need to attenuate the DAC.
My SOEKRIS is driving a pair of xilica XP2040s not power amps.
I am going to use about 12 db attenuation at the amps and then split the difference between the DAC and the xilica for further attenuation. The amplifiers that handle the bass have pots so that part is easy. I am using the raw balanced output into the xilica's balanced input. I have found there is no need for the ground wire.
I would like to use Mr. Yazaki's SOSHIN silver micas - but eight would cost much more than the board and would be very difficult to mount.
I have to say Peter is absolutely right about those polymer caps. I have never liked them and have tried to like them since they offer so many packaging advantages but they do something that is not attractive to my ears every time i have tried them.
After listening for many hours after writing I have to say this is a completely different device with those things replaced. It is more what I dreamed it would sound like. Of course, components NEVER sound like you would "dream" they sound but it very fine when they come close. I do not doubt getting that FET out of the way makes some difference but nothing compared to the caps.
I found there was no need to attempt to de-solder the FETs - just a pair of small dykes
cutting through the middle GENTLY and they pop right off. No need for any heat which is always a plus.
SOEKRIS, I know this stuff gets on your nerves and there is no question this is a fine product as shipped but I am grateful there are a few paths to making it really good. I do realize there are folks who may actually prefer the sound of the polymers. I am as baffled by the possibility of that as i am by the lure of "high resolution" digital audio with all of its unremovable artifacts and digital atmospherics that I guess some people think should be there.
Last edited:
EQ - crossover - phase - the things are almost magical.
I have no idea how to make the SOEKRIS do those things and I like the idea of "the separation of powers" - not having one device handle too many tasks.
Plus more importantly I already had them before i got the SOEKRIS. After almost two years of tweaking the speakers with this i have no desire to start over again. I do not think there is enough capability within the SOEKRIS to do all that I need plus the xilica will drive four channels. It is quite a fine sounding thing. I was very skeptical of all of this DSP stuff - I bought one just to use for bass management and then hooked up one channel of loudspeaker system to the one and ordered a second one the next week. (I do like dual mono)
My test for whether a component is messing up the sound is how easy it is to detect polarity - with the xilicas I hear it as soon as the music starts playing. I do not doubt they are chock full of ICs and I had been as kooky as any in my disdain for them but I hear nothing but good things with this.
Many years of trying out all sorts of active and passive crossover schemes never got me here. Gentle EQ can make a world of difference in how much you can enjoy your system. What is surprising is what one can do with the all pass filters in fixing room problems in the high frequencies - above 1kHz.
Using REW I get phase plots at my listening chair (I know DON'T MOVE YOUR HEAD) that are quite good for a horn loudspeaker system. Beyond belief compared to what I would see before using the xilica. My system is too bulky to make nearfield measurements of the complete system.
I have no idea how to make the SOEKRIS do those things and I like the idea of "the separation of powers" - not having one device handle too many tasks.
Plus more importantly I already had them before i got the SOEKRIS. After almost two years of tweaking the speakers with this i have no desire to start over again. I do not think there is enough capability within the SOEKRIS to do all that I need plus the xilica will drive four channels. It is quite a fine sounding thing. I was very skeptical of all of this DSP stuff - I bought one just to use for bass management and then hooked up one channel of loudspeaker system to the one and ordered a second one the next week. (I do like dual mono)
My test for whether a component is messing up the sound is how easy it is to detect polarity - with the xilicas I hear it as soon as the music starts playing. I do not doubt they are chock full of ICs and I had been as kooky as any in my disdain for them but I hear nothing but good things with this.
Many years of trying out all sorts of active and passive crossover schemes never got me here. Gentle EQ can make a world of difference in how much you can enjoy your system. What is surprising is what one can do with the all pass filters in fixing room problems in the high frequencies - above 1kHz.
Using REW I get phase plots at my listening chair (I know DON'T MOVE YOUR HEAD) that are quite good for a horn loudspeaker system. Beyond belief compared to what I would see before using the xilica. My system is too bulky to make nearfield measurements of the complete system.
Since my balanced DAC is being used in single ended mode I’ve been using only 2 channels to experiment with the caps.
Well, this would be a perfect usage of Iańs battery supply...which I have lying around, but i was not brave enough as I am not sure if I will screw the dac completely....how many rails ? 8 ?
By the way: on the whole discussion around level etc: I am playing with the tube output stage as I want 16Vpp to drive my two stage power amp.
I had originally a differential stage, so Dam1941 feeds directly two 801A Dht (per channel) and those IT loaded with a LL1692A.
I changed it now to a SE stage with LL1660 to try the different distortion spectrum of a SE stage.
I compared SE output of the DAM1941, directly onto the grid of the 801A vs. the LL1692 to do the Balanced to SE conversion. To my big surprise, the transformer solution sounded much more crisp and transparent, while pure SE was a bit muddy and soft. Both had the same tube SE stage.
Another surprise: small microfone transformers like LL1674 or LL7903 did not sound open amd transparent enough, more like a cheap coupling cap.
So a question to Soekris: what would be the ideal transformer spec ? Currently the LL1692AM-PP is highly recommended...but maybe there is even a better alternative ?
As you may have read in the past, Soren preferred direct connection.
dam1941 - Next Gen Discrete R-2R Sign Magnitude 24 bit 384 Khz DAC module
Now I use direct SE connection to my 801 preamp, but I have Cinemag 15/15B and Jensen JT-11, so I'll try -sometime- its as BAL-SE transformers.
dam1941 - Next Gen Discrete R-2R Sign Magnitude 24 bit 384 Khz DAC module
Now I use direct SE connection to my 801 preamp, but I have Cinemag 15/15B and Jensen JT-11, so I'll try -sometime- its as BAL-SE transformers.
Why are you surprised? These little horrors are meant as opamp replacements and to my ears sound as cheap opamps only with inferior bass. And yes, I have both.
Technically any transformer would work with a 600R source, the question is which one would sound right. My feeling is our friend can wind something no LL can ever dream of, but he may need a couple of practice runs.
Do you really prefer the SE pre?
Well, I am about to remove those ceramic caps.
Looking around it seems the VISHAY 1837s at least give you some inkling of their abilities in the high frequencies so i will give those a try eventually - quicker if I feel like I need to. At least these have a reputation of decent sound.
I figure a combo of ,1/.01 uF per "nest" ... if they will fit.
Looking around it seems the VISHAY 1837s at least give you some inkling of their abilities in the high frequencies so i will give those a try eventually - quicker if I feel like I need to. At least these have a reputation of decent sound.
I figure a combo of ,1/.01 uF per "nest" ... if they will fit.
Well...I will be able to tell you in a couple of months....I just changed my power amp over to SE as well...and that gives me current a very natural, desirable tone...but overall a bit too much H2. It might be that I will change back to differential stage with LL1692A as OPT and no transformer between DAC and grids of the 801A.
Our friend 50AE is winding a nice IT for the poweramp (repöacing a LL1660)...PP to SE might be the next chapter...