Best opamp for I/V conversion? (DAC)

[DSP_Geek]

Originally posted by Charles Hansen
Elso,

Here is the link to an article by Walt Jung that touches on the problems of op-amps as I-V stages, although not in great detail:

http://www.elecdesign.com/Articles/A...7207/7207.html

The real information was in the articles by Barrie Gilbert that is referenced at the bottom of Jung's article, but these have unfortunately been deleted.

[...contexectomy...]

Interesting article, thanks for the link. In other news, has anyone looked at the AD9631? Open loop bandwidth is about 100kHz, so it's not an integrator at the frequencies of interest as mentioned by Jung. The thing does appear to be a bit finicky with respect to supplies, however, as PSRR is only about 65 to 75 dB. It also seems to be happiest with impedances below 1kohm.


Cheers,
Francois.

[ojg]

Originally posted by Pedja
Btw, you are misquoting Socrates.

Didn't mean to disrespect Socrates, maybe another of his qoutes are more fitting:

I know nothing except the fact of my ignorance - Socrates

[BrianL]

1. Operational amplifiers can be used as transimpedance amps.

2. Any amplifier circuit ultimately is an "integrator". Unless, of
course, you have some amp with infinite bandwidth. It
just depends on DC open loop gain and response corner.

3. Why make broad sweeping generalizations? Op-amp are
not inherently better or worse than discrete circuits. It depends
on the op-amp, just like it depends on the discrete circuit.

[jwb]

I'm kind of becoming fond of the AD844 part, personally.

[Jocko Homo]

The '846 was the genesis for the transimpedance amp that I came up with. The dealers did not like the sound of current feedback designs. Not just mine, but other manufacturers as well.

Yes, interesting part, but it somehow never lived up to its expectations.

Can't remember the company name.....out of Fairfield where all the TM goof-balls are.......wrote a "white paper" suggesting that CF was not good for I/V, as the settling times are too slow. Which resulted in their ability to only resolve 12-14 bits. Something like that. (They liked the '5534!)

Their claim, not mine.

Jocko

[Pedja]

Originally posted by ojg
Didn't mean to disrespect Socrates, maybe another of his qoutes are more fitting:

I know nothing except the fact of my ignorance - Socrates

I just wanted to tell that (else than the fact he has not wrote anything, so in that regard it is impossible to literally quote him), the guy was not agnostic. Actually, he was at the side of achievability of the truth. The point of his lore (related to the sentence in your signature) is that the knowledge is not given by itself nor contained in the common wisdom, but yet has to be achieved; this real knowledge starts by the discovery of the fact the truth is not given directly (plainly) but the effort is needed to come across it – as long as you do not start to (re)investigate, you know nothing.

Written as it was, it looks like the motto of the agnostic which Socrates certainly has not been. Even worse, it can be understood like the play with the words and from that same ancient time there are sophists that are known both for being prone to such things and for being on Socrates/Plato’s top lists of the bad guys.

Pedja

[Elso Kwak]

Originally posted by Charles Hansen


I'm a little unclear from your post if this is the circuit you listened to. If so, I would expect it to sound quite a bit better than any op-amp I-V stage.

Charles,
This is the circuit I listened to with the NON-OS TDA1543.
I changed it a bit for the TDA1543: replacing D1 by two red LED's in series and R3 =15k. Also a 10µF bypass cap to ground at the base of the transistor, BC550C. R8 is omitted.
I got sound, though pretty low in volume and some hum. I do not like this sound. The timbre of the musical instruments is not right. And the sound gets on my nerves. I know this circuit is very crude. What shall I do, adding a constant current source or build my first circuit without the Sziklai?

BTW with a squarewave and the OPA604 I got the same amplitude (10mV) band of signals on my scope. Could not trigger it well.

[Charles Hansen]

Originally posted by Elso Kwak
I know this circuit is very crude. What shall I do, adding a constant current source or build my first circuit without the Sziklai?

It sounds to me like that you are probably not providing the DAC with the proper conditions for proper operation. The data sheet for the TDA1543 holds the clues to your problems.

First of all when you are using the op-amp I-V, it will hold the audio outputs at Vref that is also supplied by the DAC. In this case it is 2.2 VDC +/-0.1 VDC. Your circuit (as described) will have the audio outputs held at around 3.2 VDC (depending on the characteristics of the LEDs used). The DC compliance is stated as between 1.8 and 3.8 VDC (assuming a 5 V supply), so this may be all right but maybe not.

The AC compliance is restricted to 25 mV, which means that the input impedance of the I-V circuit must be less than 10 ohms for proper operation. For the simple circuit you have drawn the input impedance should be around 2.5 ohms (assuming ~10 mA bias), which should be adequate. However I would double-check the actual bias current, and also measure the actual signal level at the DAC pin.

A Sziklai pair will help in this regard, but I would probably keep the single transistor and bump the bias current up to 25 mA to get the input impedance down to 1 ohm or so. (You may or may not need a different transistor to do this safely -- I'm not familiar with the European devices.)

Next, the PSRR of your circuit is dismal -- zero! Any hum or noise on the rail is directly coupled into the audio signal.

Finally, all the little details will matter. What brand of capacitors, what kind of PS regulators, et cetera, et cetera. In the end however, I think you will find the journey well worth it.

Good luck!

[Jocko Homo]

Er, ah.....I mean Elso.........

Yeah, I would not expect that circuit to sound good at all. No wonder that your initial impression was not good.

Jocko

[Elso Kwak]

Originally posted by Jocko Homo
Er, ah.....I mean Elso.........

Yeah, I would not expect that circuit to sound good at all. No wonder that your initial impression was not good.

Jocko

Yeah I should have listened.........to your initial suggestion having a common base stage between constant current sources.......right? Just wanted to do a quicky.......