I dont think this is a good idea, its best to leave the current mode signal as short, direct and untouched as possible or you will lose resolution/dynamic range quickly, this is when signal routing is at its most important, as its very vulnerable, far more vulnerable than voltage signal. besides i'm not sure why you would want to do such a thing?
also adding any resistance inline with the current will just convert it to a voltage.
also adding any resistance inline with the current will just convert it to a voltage.
Your sentences are right both.
For the firsth, I could say that anytime you put something in your signal path you loose resolution/dynamic. So this might be weight with what I can gain with this insertion.
For the second: no comment. Infact this is the cause I decided to post my request. I have no idea how to reduce a current of a current generator keeping the equivalent circuit still a current generator.
Any Idea ?
its just a little hard to shunt a controlled fraction of the current when the DAC chip is designed/speced for use with a I/V converter with the lowest practical impedance that can be synthesized at a virtual gnd
translinear circuits can be analysized as converting I to a (log) V in transistor Vbe junction and back to current at the collector - so no joy there on strictly "keeping the equivalent circuit still a current generator"
since the DAC itself is already a current modulator of impressive resolution, linearity, low noise the simple engineering solution is to just get system gain structure after the I/V correct and then attenuate in the Digital domain - with 24 bit resolution you don't even need dgital dither - the noise of the analog I/V will be quite adequate
it can be hard to even match the PCM1794 preformance in subsquent anaolg stages - certainly when you start specifying a "required" design principle for a circuit rather than a result in terms of signal properties like noise, linearity
translinear circuits can be analysized as converting I to a (log) V in transistor Vbe junction and back to current at the collector - so no joy there on strictly "keeping the equivalent circuit still a current generator"
since the DAC itself is already a current modulator of impressive resolution, linearity, low noise the simple engineering solution is to just get system gain structure after the I/V correct and then attenuate in the Digital domain - with 24 bit resolution you don't even need dgital dither - the noise of the analog I/V will be quite adequate
it can be hard to even match the PCM1794 preformance in subsquent anaolg stages - certainly when you start specifying a "required" design principle for a circuit rather than a result in terms of signal properties like noise, linearity
Last edited:
I'm having an idea for a nice amplifier. But if I can't solve this issue my thinking is useless.
Control the digital side. For example, some PC players will output variable digital signal - Foobar2000 for example.
Right, it will the latest solution. I prefear an analog solution instead a SW solution.
I would not recommend doing what you are asking, but if you feel you need to then one possibility is to use a pair of BJTs. Join their emitters, just like in an LTP, but then use them in common base mode - inject the DAC output current into the emitters. The current will divide between them according to the base bias. Take your output current from one collector, and throw away the current from the other. The signal current will divide in the same proportion as the quiescent current.
You will need well-matched transistors with high current gain - these are available as pairs on a single chip? This techniques relies on the exponential characteristics of BJT, so it will not work so well with FETs.
You will need well-matched transistors with high current gain - these are available as pairs on a single chip? This techniques relies on the exponential characteristics of BJT, so it will not work so well with FETs.
Too late. The DAC is wooking form more or less 2 years. But the DAC is build as several modules (receivers+ DAC +IV) and I wouldn't modify the digital parts but only the analog.
looking around I've found than it's possible to reduce the output current changing the IREF of the PCM1794 changing the 10k resitor to another value. Has anyone tryed to do this?
How much can I rise the resistor value and what will happen about the performance ?
Thanks
Aiace
There's no indication from the datasheet that the PCM1794 has the feature you'd like.
One similar part that does have that feature though is AD1853:
http://www.analog.com/static/imported-files/data_sheets/AD1853.pdf
On page 10 it says:
Analog Attenuation
The AD1853 also offers the choice of using IREF (Pin 10) to
attenuate by up to 50 dB in the analog domain. This feature can
be used as an analog volume control. It is also a convenient
place to add a compressor/limiter gain control signal
One similar part that does have that feature though is AD1853:
http://www.analog.com/static/imported-files/data_sheets/AD1853.pdf
On page 10 it says:
Analog Attenuation
The AD1853 also offers the choice of using IREF (Pin 10) to
attenuate by up to 50 dB in the analog domain. This feature can
be used as an analog volume control. It is also a convenient
place to add a compressor/limiter gain control signal
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.