Bill,

Thank you for the response.

What I am designing is a transformer based DAA.  The transfomer based DAA
I design use an active 2W to 4W hybrid.  All previous designs using single
5V supply and rail to rail opamps function very well.  THD is very close
to transformer THD.  The only difference in is this DAA is the 3.3V opamp
supply.  I'm using Analog Devices AD824 opamps.  They are spec'd to
function with a +2.7V supply.  I have tried other vendor's r to r opamps
with the same results.  I have a feeling the problem is caused by the
carrier peak to rms delta.  Some others would call it rms to crest.  The
delta is +12 to +14 dB depending on who you ask.  According to my
calculations and crude measurements it's closer to 12 dB. 

Although opamps are called rail to rail the THD starts to increase
somewhat before the signal runs into the rails.  It increases
exponentially the closer the signal gets to the rails. Due to complexity
of the V.32bis and V.34 carrier it is impossible to see or measure the 
THD.  I can measure it with steady state signals, but that is not the 
same as a scrambled carrier.

I may have to use a charge pump to generate a 5V supply.  The design is 
already very tight now.

Regards,

Duane


On Wed, 30 Apr 1997, Bill Pedersen wrote: 

> 
> Duane,
> 
> 
> 
> I have experienced the same problem with the linear optocoupler design.  
> We currently use this design for international products for voice 
> applications, however the problem was detected at the compliance stage. 
> We were powering the opto coupler and the op amps that drive the opto's 
> off the -48 for safety compliance, which measured around 9 volts at the 
> opto and op amp at 80mA loop current.  Your distortion is probally being 
> caused by low loop current. The major problem is that we found was to 
> many voltage drop in the DAA, (ring detect, loop detect, and bridge) 
> which decreases the voltage level down to 3 volts to drive the op amp and 
> the opto.  This gave us horrible THD levels along with other problems.  
> We have now redesigned the DAA and have removed the Op Amp and Opto 
> Coupler and returned to the transformer in design.  I have heard through 
> the grape vine that were not the only ones caught.  I hope this has been 
> helpful.
> 
> 
> 
> Good Luck,
> 
> 
> 
> Bill Pedersen
> 
> GENTNER COMMUNICATIONS CORP.
> 
> 
> 
> E-mail: [email protected]
> 
> 
> 
> 

Reply via email to