Charlize 2

[dweekie]

New Charlize is out, version 2. The input caps have been removed, so the 2.4vdc bias must have been taken care of somehow. I wonder how much improvement comes from this change. It will definitely save some money on nice caps.....

http://www.diyparadise.com/charlize2.html

[theAnonymous1]

It's the same DC coupling tweak that has been discussed here before; nothing new. I have done it to a few Tripath amps.

You simply connect your source ground to the BIASCAP pin of the chip instead of the analog ground. This requires you to isolate the source ground and the power supply ground and Yeo does state that in the description.

[BWRX]

That sure isn't the best or most reliable way of getting rid of the coupling caps... but it is the cheapest!

[theAnonymous1]

Brian, would it be possible to power the chip with a +/-2.5v supply by connecting the supply ground to the BIASCAP pin? Could the 100nf cap then be taken out so the new BIASCAP ground can be connected to the supply ground?

[BWRX]

We don't know how the 2.5V bias is generated inside the chip. Most likely it is just a voltage divider with equivalent resistors, and if that's the case you could safely use a +/-2.5V supply with the ground connected to the biascap pin.

But if you're going to the trouble of making a separate +/- supply you could up the voltage to +/-12V and DC couple the source to an instrumentation amp whose output is referenced to the 2.5V bias. The inamp will also allow you to use balanced or unbalanced sources.

[theAnonymous1]

Originally posted by BWRX
We don't know how the 2.5V bias is generated inside the chip. Most likely it is just a voltage divider with equivalent resistors, and if that's the case you could safely use a +/-2.5V supply with the ground connected to the biascap pin.

I think I'm going to try this. I will cheat and use 2.5V regulators instead of adjustables.

Originally posted by BWRX
But if you're going to the trouble of making a separate +/- supply you could up the voltage to +/-12V and DC couple the source to an instrumentation amp whose output is referenced to the 2.5V bias. The inamp will also allow you to use balanced or unbalanced sources.

Say waaaah.

I'll leave the fancy stuff to you Brian.

[BWRX]

I'll draw up a schematic later. It's actually very simple, I just made it sound worse than it is.

[theAnonymous1]

No you explained it ok, my brain is just being lazy. I get it now.

I was planning on making a +/-12v supply for a buffer I am going to add anyway.

If I only have single ended inputs, would an instrumentation amp be worth the effort? If I did use one, would there be no point of trying a +/-2.5v supply?

EDIT: Nevermind about the 2.5v supply. Now that I think about what I was hoping to accomplish, it won't work. The negative of the supply is permanantly attached to the analog ground, so if the 100nf cap is removed there will still be a DC offset plus the fact it would short the negative supply to ground.

[BWRX]

The inamp will be worth it regardless of your source considering that you'll remove coupling caps from the signal path. But if you have a single ended source and are happy with the sound as it there's certainly not much reason to try this! If you use the inamp there is no need to use the +/-2.5V as well.

If you are going to try the +/-2.5V rails do it on an amp you wouldn't mind losing and with a dummy load or test speaker, just in case anything crazy happens. And don't forget to make sure this supply is floating.

If you're going to try the inamp you'll want to tie the ground of the inamp supply to the ground of the Tripath circuit.

One other thing... If the 2.5V bias is indeed derived from a simple voltage divider, it may be advantageous to use a unity gain buffer for the biascap voltage so you have a nice low impedance output driving the reference of the inamp.

[theAnonymous1]

Na, forget about the 2.5v supply idea.

I have some OPA1632 I can try this with. Is this what you mean for a reference voltage buffer?