Did some measurements....
Hi Ken, Got around to measuring a few Edcor transformers last night. Was rather interesting. Measuring just the primary without terminating the secondary gave weird results. Measured anywhere from 3 to 13 H. However when you terminate the secondary we (my Cousin and I) measured 30 mH. After some thought this makes sense. It also clearly shows why the secondary should be terminated so that the dac doesn't see any odd reflections from the secondary. We measured a XSM 600/600 and a WSM 2.7K/1.2K with the same results. So what the dac sees is 4.7 nF across about 30 mH. My Cousin did extensive sweep testing with the aforementioned WSM transformer with a CS4397 dac no peaking was seen. Given my druthers I'd pop for Cinemag transformers for my future projects, however the Edcors are great bang for the buck.
Dave
Hi Ken, Got around to measuring a few Edcor transformers last night. Was rather interesting. Measuring just the primary without terminating the secondary gave weird results. Measured anywhere from 3 to 13 H. However when you terminate the secondary we (my Cousin and I) measured 30 mH. After some thought this makes sense. It also clearly shows why the secondary should be terminated so that the dac doesn't see any odd reflections from the secondary. We measured a XSM 600/600 and a WSM 2.7K/1.2K with the same results. So what the dac sees is 4.7 nF across about 30 mH. My Cousin did extensive sweep testing with the aforementioned WSM transformer with a CS4397 dac no peaking was seen. Given my druthers I'd pop for Cinemag transformers for my future projects, however the Edcors are great bang for the buck.
Dave
An externally hosted image should be here but it was not working when we last tested it.
hi, sorry it's too amateurish
all good for 1541 but what about 1543? where would i connect the vref pin to?
Hi george. I did try 3 stacked AD844 in one of the channels. The 3-stack channel seems to have more depth in the sound but that could be my subjective impression. I cannot say for sure. I will order more Ad844 so i can evaluate properly both channels
I saw it on the old pedja thread that stacking does lower the input impedance. I wonder, tho, even with 50ohm halved by stacking how ad844 can compare to something like this that has zero? input impedance to begin with. currently i wish to build both to compare....
'Super common gate i/v converter for 1543' is the thread which i cant seem to link on my android phone...
'Super common gate i/v converter for 1543' is the thread which i cant seem to link on my android phone...
The datasheet for the 844 shows I/V converter with usual feedback resistor. In general, the input impedance of the virtual ground circuit would be the value of the fb resistor divided by the open-loop voltage gain.
The claim is that, since the instantaneous input voltage arising from a current step is much lower than that caused by an opamp with the more usual very high input impedance, the glitches will be much smaller.
PreSapian was wondering how the non-feedback use of the 844 might compare with a near-zero impedance circuit. One logical comparison would be with the 844 arranged to have near-zero input impedance.
Such low impedance can only be achieved with lots of gain somewhere, and gain always has a down side.
I've just stuck a couple of 844 in an Arcam Alpha and, once I'd stopped them fidgeting and corrected Arcam's strange choice of offset nulling current, it sounds and looks a bit better than it did. It takes me so long to do, however, that I can't be sure I remember what it sounded like before.
I've been scouring this thread to find the bit where it is demonstrated that the AD844 can suffer current starvation when used singly with, say, a TDA1541A. Help, please
The Arcam is quite unusual, btw, for having no cap in the signal path until after the I/V stage.
Last edited:
Here is a post by Supra with a TDA1541, as it has so much higher output than the PCM1704 (1.2mA) I'm using and it was much better with two than one, he found 3 stacked to be much better, looking at the output of the TDA1541 (4mA) I think 4 stacked may/maynot be even better.
http://www.diyaudio.com/forums/digital-source/227677-using-ad844-i-v-35.html#post3476672
Cheers George
As an experiment I just did a tripple stack per channel, with my(+-1.2mA) PCM1704K's CDP and while the difference was not as massive going from 1 to 2 stack, there is still an obvious difference with three.
I will be doing a direct A/B against an indentical cdp that I did a double stack on next week, I'll report on that comparison.
As for the subjective sound of the tripple stack without a direct a/b, it seems to have a bit bigger sound stage again, the bass seems more powerfull (just as tight), the mids seem more wholesome again (for want of a better word), and the top end is a touch less prominant (but just as detailed) this could be from the subjectivly more powerfull bass, and this bass could be the reason the mids ar more "wholesome".
It seems to have tilted the the frequency spectrum a bit towards maybe the "darker side" without being dark, if that makes any sense?
Another thought is that while there now is no current starving going on maybe what I'm hearing is the lower input impedance again (15ohm'ish) that the dacs output is being loaded down to.
Gain and frequency response is still identical using the same TZ resistor 4.7kohm and cap 470pf for all 3 configurations being 1, 2, or 3 , AD844's i/v stages.
Any thoughts from the brains trust on why three is different from two?
Cheers George
I will be doing a direct A/B against an indentical cdp that I did a double stack on next week, I'll report on that comparison.
As for the subjective sound of the tripple stack without a direct a/b, it seems to have a bit bigger sound stage again, the bass seems more powerfull (just as tight), the mids seem more wholesome again (for want of a better word), and the top end is a touch less prominant (but just as detailed) this could be from the subjectivly more powerfull bass, and this bass could be the reason the mids ar more "wholesome".
It seems to have tilted the the frequency spectrum a bit towards maybe the "darker side" without being dark, if that makes any sense?
Another thought is that while there now is no current starving going on maybe what I'm hearing is the lower input impedance again (15ohm'ish) that the dacs output is being loaded down to.
Gain and frequency response is still identical using the same TZ resistor 4.7kohm and cap 470pf for all 3 configurations being 1, 2, or 3 , AD844's i/v stages.
Any thoughts from the brains trust on why three is different from two?
Cheers George
The triple AD844 (I/V) is in place in my DAC and extensive music listening has been performed.
From Glenn Gould with classical piano music to Queen with rock music. The third chip will absolutely stay soldered in there. Glenn Gould´s piano is now in the room with me. Breathtaking. Playing on a pair of DIY Apogee Duetta copies
From Glenn Gould with classical piano music to Queen with rock music. The third chip will absolutely stay soldered in there. Glenn Gould´s piano is now in the room with me. Breathtaking. Playing on a pair of DIY Apogee Duetta copies
I should have looked a bit harder. The PCM63 is +-2mA that is .8mA higher that my PCM1704K. So you may like 4 stacked even better, if any current starvation is still going on. Or it maybe the lower input resistance the dac is seeing that making the difference, with 4 stacked if you do it your input impedance will be around 10ohms that the dac will see, just make sure it's not getting too warm.
Cheers George
Cheers George
success! -,+ of 1543 differential pair is connected to pin 2,3 of ad844 and it's making music. lowpass filter caps are omitted since i can't find 1nf, and still trying to figure out if i need c1 or not... and i'm pulling the signal staight out of pin6 with no resistor to ground.
in this half baked state, the sound is... well... i'm sorry but transformer coupled 1543 is still the best in my book. there's more slam and clarity with 844 but the harmonics aren't as rich. it sounds too hi-fi for my liking so far. will tweak further and report.
in this half baked state, the sound is... well... i'm sorry but transformer coupled 1543 is still the best in my book. there's more slam and clarity with 844 but the harmonics aren't as rich. it sounds too hi-fi for my liking so far. will tweak further and report.
...magnificent...
with 1nf in place, it's simply the best sound i have ever heard. before that it was grating.
thanks, george. this one is a keeper. it's so smooth yet resolving. purity at best.
edit: so from the first circuit you posted, george, i have omitted all the components except tz resistor and the cap. my amp is myrefc with input cap so i can dc couple 844 no prob.
what else should i implement in my build?
sorry i should really condense my posts but the excitement does make me impatient like a child... the sound is impressive, yes, but harmonics aren't as rich as passive output 1543. though more is gained from the trade-off, still i want to know if it can be remedied with 844 alone.
with 1nf in place, it's simply the best sound i have ever heard. before that it was grating.
thanks, george. this one is a keeper. it's so smooth yet resolving. purity at best.
edit: so from the first circuit you posted, george, i have omitted all the components except tz resistor and the cap. my amp is myrefc with input cap so i can dc couple 844 no prob.
what else should i implement in my build?
sorry i should really condense my posts but the excitement does make me impatient like a child... the sound is impressive, yes, but harmonics aren't as rich as passive output 1543. though more is gained from the trade-off, still i want to know if it can be remedied with 844 alone.
Last edited:
Yes you need that 1nf in place otherwise you sending a lot of very HF crap to whatever buffer you are using and on to the output.
Yes yes the TDA1543 has double the output of my 1704, so stacking 2 or even maybe 3 will be well worth doing, as it was so so much better than one with the 1704.
Cheers George
reading this post, i realize using both of 844's inputs is not the way to go. so i tried to remedy this and it doesn't work. i made another crude drawing of my setup. bottom 1543 that were feeding pin 3 with -L,-R were taken out and socket holes jumpered from pin 6,8 to ground. and.... i get no sound. what am i doing wrong? please help....
An externally hosted image should be here but it was not working when we last tested it.
Last edited: