Hello everybody.
I would like to know from the experts if it is possible to realize the following thing and what problems or advantages it entails.
current DAC output stage instead of the I / V converter, immediately following the passive line crossover filter driven by current instead of voltage, then current-current power amplifier and finally speakers which are consequently current driven. Starting from the DAC output, everything is driven in current.
bye thank you
I would like to know from the experts if it is possible to realize the following thing and what problems or advantages it entails.
current DAC output stage instead of the I / V converter, immediately following the passive line crossover filter driven by current instead of voltage, then current-current power amplifier and finally speakers which are consequently current driven. Starting from the DAC output, everything is driven in current.
bye thank you
Hello
I may understand your intended solution wrong, but allow me to come with some comments.
The DAC's I've looked into (and that does certainly not include all existing DACs🙂) delivers their output as a current difference between two outputs.
Both outputs has a DC bias at least half of the peak output signal.
Therefore I fail (lack of inspiration, possibly) to envision how to implement a passive anti-imaging filter that can convert the current difference from the DAC to a single output current while maintaining the DC output current.
Well, maybe using a transformer based topology?
It may also be very difficult to avoid inductors in the filter - and allthough I know opinions differs on this, my thinking is that using inductors for signal level filters is not recommended.
Why do you want to explore this "current all the way" topology, anyway?
Cheers,
Martin
I may understand your intended solution wrong, but allow me to come with some comments.
The DAC's I've looked into (and that does certainly not include all existing DACs🙂) delivers their output as a current difference between two outputs.
Both outputs has a DC bias at least half of the peak output signal.
Therefore I fail (lack of inspiration, possibly) to envision how to implement a passive anti-imaging filter that can convert the current difference from the DAC to a single output current while maintaining the DC output current.
Well, maybe using a transformer based topology?
It may also be very difficult to avoid inductors in the filter - and allthough I know opinions differs on this, my thinking is that using inductors for signal level filters is not recommended.
Why do you want to explore this "current all the way" topology, anyway?
Cheers,
Martin
I have given this some thought - the problem that I see is the PLLXO is driving into a zero impedance (typically an opamp acting as TIA thus generating a virtual short) and hence any shunt capacitor in the low-pass section of the XO needs to be of infinite value and the inductors of infinitesmal value. In practice of course we can settle for a very small terminating resistance - say 1mohm but even then the capacitors are huge and the inductors dauntingly small.
You can verify this for yourself by going here - RF Tools | LC Filter Design Tool. Try a 4th order Butterworth into a 1mohm impedance, the inductors come out in nH and the caps in tens of mF.
You can verify this for yourself by going here - RF Tools | LC Filter Design Tool. Try a 4th order Butterworth into a 1mohm impedance, the inductors come out in nH and the caps in tens of mF.
Yes, there is a current-mode equivalent for *any* voltage-mode network.
Is it practical? Often it becomes more work than the way we usually do it.
And all things corrected, the result is EQUIVALENT. V mode and I mode are mental abstractions and practical approximations, not magic.
Is it practical? Often it becomes more work than the way we usually do it.
And all things corrected, the result is EQUIVALENT. V mode and I mode are mental abstractions and practical approximations, not magic.
Hello and thanks.
I am attracted to two things combined together:
the driving of the loudspeakers in current and the absence of the crossover filter between the power amp and the loudspeakers.
there is also the possibility of obtaining the current driving starting from the transconductance amp and everything that is first driven in voltage but in this case you need a preamplifier with an output impedance tending to zero, very difficult to obtain.
there is also the possibility of using a DSP instead of a passive crossover but not in my case, I have never had to deal with it.
I am sorry, I can't give you any hints to reach your goal - I've never tried anything like that.
However: I don't see an anti-imaging (sometimes also called anti-aliasing) filter in your chain as the first step after the DAC - is that intentional?
The reason I ask is that I have seen at least one DAC implementations somehow having overlooked the fact that the application note used as reference was intended to go directly into a headphone - and not into some active electronics like an amplifier with unknown RF behaviour.
Best practice with a DAC is to have a lowpass filter with a cutoff frequency higher than the audible range and lower than half the sample frequency - to "smooth" the staircase steps from the DAC output so to speak.
Failing to do this will result in a very substatial amount of high frequency "aliases" of the original signal far beyond the audible range being delivered into the following stage in the chain.
Cheers,
Martin
However: I don't see an anti-imaging (sometimes also called anti-aliasing) filter in your chain as the first step after the DAC - is that intentional?
The reason I ask is that I have seen at least one DAC implementations somehow having overlooked the fact that the application note used as reference was intended to go directly into a headphone - and not into some active electronics like an amplifier with unknown RF behaviour.
Best practice with a DAC is to have a lowpass filter with a cutoff frequency higher than the audible range and lower than half the sample frequency - to "smooth" the staircase steps from the DAC output so to speak.
Failing to do this will result in a very substatial amount of high frequency "aliases" of the original signal far beyond the audible range being delivered into the following stage in the chain.
Cheers,
Martin
And what is the best way to engineer a highly linear current-current amplifier? My strong suspicion it is as an IV converter followed by a voltage controlled current amp... That's not to say you couldn't have a passive crossover network directly on the DAC outputs and IV convert after it (assuming you don't violate the voltage compliance specs of the DAC output).