The ES9022 is not like the rest of the ESS DAC range - it has an on-board op-amp output stage (stereo) so it's optimal load is determined by the op-amps output stage. I was hoping Dustin would have answered this one but the datasheet says 5K minimum so I guess that's the load to aim for? As I say though, there is no noticeable change in sound when a 10K R is put across the primaries & I believe this should then meet the DAc's load requirement. Now if there is no change in sound then it's not that I prefer the distortion, I just can't hear any difference!
Last edited:
And why would you? The 5 or 10k in parallel with the headphones won't change the loading in any meaningful way. So, if you indeed are using headphones, with or without the transformer the loading conditions are very similar and there should not be any obvious change in sound. Either way the opamps (if indeed there are opamps involved) will sweat bullets
Btw, it's not surprising you prefer the sound without the transformer - in this situation it is not doing anything useful other than limiting the bandwidth and adding distortion. A 70ohm resistor in series with the headphones will have the same loading effect on the dac.
Far from it. The headphones are already loading the output down to 70 ohms(+ transformer primary + secondary dcr) and any resistors in parallel will only reduce this.
Hold on, the 5 or 10K is across the trafo primary & the headphones are connected to the trafo secondary - are you saying that this will make no difference to the load that the DAc sees?
Yes, I am using headphones, do you doubt me? As to the op-amps - I'm guessing that is what is internal in the DAc!edit: I'm not actually guessing, I looked at the datasheet again & it states "2Vrms Op-amp driver"So, if you indeed are using headphones, with or without the transformer the loading conditions are very similar and there should not be any obvious change in sound. Either way the opamps (if indeed there are opamps involved) will sweat bullets
Last edited:
analog_sa,
I believe this statement of yours fails to take into account what I said about the reactance of the transformer rising sharply above about 10Hz where it is usually >10k (depending on transformer). So this is the load the DAC sees above 10Hz, below that it might sweat but not much, I believe - how much music is actually below 10Hz?
Putting a 5 or 10K R across the primaries would be seen by the DAC's output as loading & would satisfy the DAC's recommendation at all frequencies. In this case it makes absolutely no difference to the sound probably because what I said above is true?
So I'm not sure what you're saying - can you give some detail? Always willing to learn.
You should definitely do some reading on transformers. Apparently these 10Hz and 10k are really confusing you.
Briefly, due to the lack of primary inductance transformers do not transform so well below a given frequency and this determines the minimum source impedance which will allows a specified low frequency limit according to the formula Rs = 2*Pi*f*L. In some interstage transformers this limit is indeed 10k, in yours it is 600ohm. It is only important at really low frequencies, like the 10Hz you mention. At mid and high frequencies it is quite irrelevant. The driving impedance of an opamp is pretty much zero and thus your transformers will work fine. So, once they are working, they simply reflect the impedances between primary and secondary as 1:1 plus the two dc resistances. In this case you can completely forget about the 10k
Not sure if i can explain basic things well, i lack the patience. A textbook is really recommended.
Briefly, due to the lack of primary inductance transformers do not transform so well below a given frequency and this determines the minimum source impedance which will allows a specified low frequency limit according to the formula Rs = 2*Pi*f*L. In some interstage transformers this limit is indeed 10k, in yours it is 600ohm. It is only important at really low frequencies, like the 10Hz you mention. At mid and high frequencies it is quite irrelevant. The driving impedance of an opamp is pretty much zero and thus your transformers will work fine. So, once they are working, they simply reflect the impedances between primary and secondary as 1:1 plus the two dc resistances. In this case you can completely forget about the 10k
Not sure if i can explain basic things well, i lack the patience. A textbook is really recommended.
As a simple example about the practical difference between a 10k and a 600R transformer you can calculate the primary inductance for a low frequency corner of 10Hz. The 10k transformer needs 159H and the 600R - only about 9.55H. Apparently the 600R transformer will need a lot less turns and would probably be much easier to wind.
As I said I'm open to learning but I'm just repeating what I have been told so far about audio transformer operation in the thread I started "DAC output using transformer". Now I may have got this wrong or misunderstood what Revintage & others have said "At 20Hz your transformer will have a reactance of at least 10kohm, probably a lot higher, with unloaded secondaries." http://www.diyaudio.com/forums/digi...c-ouput-using-transformer-15.html#post1999272
Is the unloaded secondaries the critical part to this - when secondaries are loaded with 70ohm headphone does this all change? Am I misunderstanding what reactance is? Does it not look to the DAC like 10kohm resistance?
Is the unloaded secondaries the critical part to this - when secondaries are loaded with 70ohm headphone does this all change? Am I misunderstanding what reactance is? Does it not look to the DAC like 10kohm resistance?
Not sure in what context was this said but quite possibly it is a secondary load to reduce overshooting in the transformer. As your secondary load is 70ohm you need not worry. In any case a 1:1 transformer feeding headphones is not doing anything useful as far as i can see. If it was feeding a preamp it would at least offer some galvanic isolation.
Nope it wasn't about loading the secondaries to prevent ringing here's the post http://www.diyaudio.com/forums/digi...c-ouput-using-transformer-15.html#post1999285
& here's the context:
Q: " Now can you say what is the function of the parallel R across the primaries & do we aim for a 10:1 trafo DCR to DAC impedance ratio?"
A: "The parallell resistor is for loading the DAC in voltage mode. Should probably be somewhere between 2k and 10kohm. You don´t need to worry about DCR when using the 9018."
Now forget about the reference to 9018, I understand this - it's the ESS DAC & operates in either current out or voltage out depending on the load it's driving into!
This has me confused
& here's the context:
Q: " Now can you say what is the function of the parallel R across the primaries & do we aim for a 10:1 trafo DCR to DAC impedance ratio?"
A: "The parallell resistor is for loading the DAC in voltage mode. Should probably be somewhere between 2k and 10kohm. You don´t need to worry about DCR when using the 9018."
Now forget about the reference to 9018, I understand this - it's the ESS DAC & operates in either current out or voltage out depending on the load it's driving into!
This has me confused
Guys,
This may be premature but I think I've discovered a way to avoid the potentially dirty USB ground from feeding through I2S - just don't connect it
Here's how I happened upon it - I had wired up a I2S DIP8 socket the wrong way & the only way my ESS DAC would connect correctly to it with the correct I2S connections was to plug it in on one side of the DIP8 socket - see the pic. Anyway, I decided to try it this way rather than take it back to the work bench for another micro-surgery. To my surprise it worked & sounded just as good as with ground connected (I connected a gnd to hear if there was any difference).
So no ground connection through I2S - galvanic isolation? Now this may only apply to the ESS DAc as it uses a charge pump on it's output stage to generate a neg 3.3V supply from the single pos 3.3V feed. It may only be because I'm running the DAC on batteries. It may be a combination of these things, I don't know but it's probably worth trying on some other DACs. I don't believe it will kill them as I've connected up the I2S socket incorrectly a number of times in a number of different ways & nothing has died yet!
Maybe somebody with knowledge in this area could pitch in & outline what might be happening & what are the potential problems?
It's late here & I haven't applied this to the HiFace yet but will tomorrow
Report back on your results
This may be premature but I think I've discovered a way to avoid the potentially dirty USB ground from feeding through I2S - just don't connect it
Here's how I happened upon it - I had wired up a I2S DIP8 socket the wrong way & the only way my ESS DAC would connect correctly to it with the correct I2S connections was to plug it in on one side of the DIP8 socket - see the pic. Anyway, I decided to try it this way rather than take it back to the work bench for another micro-surgery. To my surprise it worked & sounded just as good as with ground connected (I connected a gnd to hear if there was any difference).
So no ground connection through I2S - galvanic isolation? Now this may only apply to the ESS DAc as it uses a charge pump on it's output stage to generate a neg 3.3V supply from the single pos 3.3V feed. It may only be because I'm running the DAC on batteries. It may be a combination of these things, I don't know but it's probably worth trying on some other DACs. I don't believe it will kill them as I've connected up the I2S socket incorrectly a number of times in a number of different ways & nothing has died yet!
Maybe somebody with knowledge in this area could pitch in & outline what might be happening & what are the potential problems?
It's late here & I haven't applied this to the HiFace yet but will tomorrow
Report back on your results
Yes, guys, I was aware that ground was needed to establish the correct potential for the USB signals & as I said this is working on a squeezebox receiver (SBR). The SBR's power plug has no ground. The ethernet cable to the laptop are galvanically isolated from the laptop's PC, the ESSDAC that plugs into I2S (without ground) runs on battery - so no ground loops in sight.
I assumed it would be the same for I2S! - if the potentials weren't correct then I2S wouldn't work - but it does in this case. Has it something to do with the voltage levels of I2S are low=0.8V & high=2V? Some more investigation is called for methinks.
I assumed it would be the same for I2S! - if the potentials weren't correct then I2S wouldn't work - but it does in this case. Has it something to do with the voltage levels of I2S are low=0.8V & high=2V? Some more investigation is called for methinks.
Sorry I re-read my original post & it appears in it's first line that I'm suggesting to not connect the USB ground BUT I didn't intend this meaning - I was just trying to be journalistic & interesting & lost the meaning - something that happens a lot in journalism I think. What I was saying was that not disconnecting the I2S ground, in my particular set-up, seems to work
I wonder if the I2S voltage requirements http://www.sparkfun.com/datasheets/B...rds/I2SBUS.pdf are more relaxed then USB USB in a NutShell - Chapter 2 - Hardware. From these links it can be seen that the signal levels are somewhat the same: low is <0.3V USB or <0.8V I2S; High is > 2.8V USB or >2V I2S. But USB operates via a differential voltage level between D+ & D- of 200mV whereas the the signal levels in I2S are as above. Does this allow it to work? More testing required.
I'm trying to tease out a cogent reason for it - maybe if the voltage potential between the intended grounds is measured beforehand & it's close then it's OK (& it won't change based on a fluctuating grid voltage !) I don't know the answer.
I wonder if the I2S voltage requirements http://www.sparkfun.com/datasheets/B...rds/I2SBUS.pdf are more relaxed then USB USB in a NutShell - Chapter 2 - Hardware. From these links it can be seen that the signal levels are somewhat the same: low is <0.3V USB or <0.8V I2S; High is > 2.8V USB or >2V I2S. But USB operates via a differential voltage level between D+ & D- of 200mV whereas the the signal levels in I2S are as above. Does this allow it to work? More testing required.
I'm trying to tease out a cogent reason for it - maybe if the voltage potential between the intended grounds is measured beforehand & it's close then it's OK (& it won't change based on a fluctuating grid voltage !) I don't know the answer.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.