TDA1543 DC connect to TA2024 amp

[jkeny]

oon, I think you're talking about the TDA1545(1mA out) not the TDA1543 (2mA out). A resistor IV can be used on the TDA1543 but not the TDA1545.

I was hoping it would be possible to eliminate all caps & DC connect TDA1543 to TA2020 Tripath amp

[oon_the_kid]

You can use resistor IV in either. Actually my original pre built unit came with resistor IV. However, because the TDA series DAC are all current sink outputs, they will need a voltage at the audio outputs. And if you look at the datasheet and how the circuit is designed, you will realise the output is designed to be held at v-ref which is at 2.2V by the I-V converter.

This will pose the limit to how much your voltage can swing. Too much swing and you would change the voltage at the audio output current sinks of TDA1543 significantly and could affect the linearity of the DAC.

That is why the op-amp is present in all DAC datasheet for the I-V stage. other possibilities include a clever circuit that will maintain the audio output at the constant voltage while allowing the output voltage swing.

Oon

[jkeny]

Yes you can but the output from the TDA1545 needs an active IV stage after to get some decent volume. I was thinking of using ec-designs (really Henck's idea) of not connecting the Vref pin to anything but supplying Vref via a constant current source/sink into a 680 ohm IV R (see the SD Player thread). I think this ignores the voltage compliance of the output stage which is fine according to Henck AFAIR. Some chips can take this & the TDAXXXX can, apparently

[oon_the_kid]

Sorry but I am not familiar with that thread. You have to post to me the link.

If I am not mistaken, my guess would be that the output would have a voltage swing of about +-800mV. If you look inot the datasheet of the chip, you would find the output voltage is not recommended to fluctuate more than +-25mV. No doubt it would work, but I believe your linearity would be compromised. There is also a limit on the DC voltage according to datasheet.

There are I believe ways around it by using FET circuits etc. You are free to go ahead, but I do not believe a good cap is too sonicalloy degrading. I believe the limit is to be how many op-amps it has to go through. In which case it is the op-amp in the Tripath that would be a limit. BY the way avoiding doing I-V on an op amp doesn't quite work because of that stage.

You are still back to doing I-V than V to I and I to V again in that stage. (Do take note the input of the tripath is a inverting amp (not a non inverting amp), that is why there is another I-V there.

Oon

[ChuckT]

Originally Posted by oon_the_kid

Okay,

What I did is in the circuit.

In this circuit the ratio of the AC current between the 2 resistors is given by the ratio of the resistor between the two of them. So the volume cannot go completely to 0. I can't remember the full details between them. I think I used a 1Kohm for the VR and 100 ohm for R1.

In that way you can reduce the number of stages in between the D to A to the amp, by using the inverting stage of the amp as the I-V itself.

Oon

Hi Oon,
that is a very novel way of mating the 1543 and the t-amp. That must the shortest signal path other than a true power dac. Having u done this already?

[jkeny]

ecdesigns post here http://www.diyaudio.com/forums/showp...&postcount=184 Notice a couple of tricks on this schematic, all of which originated with Henck - biasing of TAD1543 input to 1.4V; Vref pin not connected; output IV through 680ohm biased with 3V reference voltage

[oon_the_kid]

Dear Chuck-T,

Yup, done that already. The downside is the amp now looks a lot like a high gain amp, and therefore you need to make sure the shielding is good. I tend to pick up RFs. Especially if the volume control has along wire. I've also done an alternate one (actually same one, before I modified it again), where I used a fix resistor to the I-V stage (about equal value to the one on the diodes constant voltage source), and another VR to a capacitor to the ground. The VR would serve to bleed some of the current coming through to ground, that means your actual signal would not even have pass the VR. Downside is after you used it for a while, the VR becomes a bit worn off, and as the volume is turned up or down, the volume suddenly gets really loud. A remote control relay based system should fix this nicely though...

Oon