DIYHiFi.org

Common guys 83 views of the circuit, and not one opinion on it.

Cheers George

_________________
Can't play any more, I think I hear my mother calling me.

Pin4 goes to GND if you have'nt noticed. Or least can be used as star ground?

Here is a better alternative power supply:
http://www.amb.org/audio/sigma22/

Way to go, dude. Just what our boy needs more choices

[quote="singa Pin4 goes to GND if you have'nt noticed. Or least can be used as star ground?

Here is a better alternative power supply:
http://www.amb.org/audio/sigma22/[/quote]


Duh, maybe I should have showed this, the bubble no. 4 gets bridged, for what?

_________________
Can't play any more, I think I hear my mother calling me.

[quote="georgehifi]

Duh, maybe I should have showed this, the bubble no. 4 gets bridged, for what? [/quote]

George,I get what you mean so please read the setup instructions for the shorting links that you require.If they say to bridge pin4 then do it.I did not design the pcb. Most probably they connect the zapfilter ground or dac ground.A ground reference is needed for balanced circuit.

http://www.lcaudio.com/index.php?page=6

Let's get off this pin 4 bubble, I've already conceded that there is some track missing on the circuit diagram.
Now back to the main question what do you current I/V experts think of this circuit hooked up to my PCM1738E dac, (Jocko do you dislike this one too)? Pedja, Carlos?

Cheers George

_________________
Can't play any more, I think I hear my mother calling me.

George,

I'm no expert here and as it maybe with anything we all have our prefaces.

That being said here a few things to consider. The supply rails are a little on the low side for a design like this.

My biggest beef would be the zener diode and 300 ohm resistors loading the input stage, yuk. A current mirror here would be more my style. I am a little confused about the loading of the next stage connecting the collectors to the emmitters of the next stage????

I think you could implement something with less stages than this and get better performance.

Thanks
Gordon

_________________
Gordon
Chief Scientist
Wavelength Audio

On the topic of opamps and running temperature, somewhat discussed earlier in this thread (here relating to current-feedback (CFB) models such as AD844 and THS3001) …



...I am now noticing heat issues with non-CFB models (e.g. AD797 and especially AD843), and specifically used in the MF A324 I/V stage, as noted in this schematic:


(Note the use of the 3300pF capacitor C(L), and 620 ohm resistor in the IV feedback circuit.)

The MF A324 I/V stage was designed pretty much based on the PCM1738's datasheet suggestions (page 23 here: http://focus.ti.com/lit/ds/symlink/pcm1738.pdf).

When using AD797, I inserted a 100 ohm resistor in series with C(L), as recommended in the AD797 datasheet …


Using this config, the maximum running temp I recorded was 97F (36C). Note that C(L) was kept at 3300pF per orig. design, not 20pF to 120pF as in schematic above; the sheet says: "The value of C(L) depends on the DAC and again, if C(L) is greater than 33pF, a 100 ohm series resistor is required."). Note: the AD797 is a single-channel opamp so the application requires two per stereo channel.

When using AD843 (no changes to stock I/V in first image above) the maximum running temp I recorded was 112F (44.4C). Although this is puzzlingly "hot", it is still within spec (max datasheet temp: 70C). I heard no obvious distortion during listening tests. Note: the AD843 is a single-channel opamp so the application requires two per stereo channel.

I've also used AD825 (single) and AD827 (dual) -- both are cool-running models.

My Questions/Comments:

Is the stock IV circuit (show in first image above) not "optimized" for AD797 and especially AD843? If so, is that causing the heat "issue"?

If not, and if these opamps are behaving normally, should one heat-sink them?

While the running temps may be high, both these opamps are the best-sounding models I've used in this I/V circuit.

Hollow_man

The opamps might be oscillating and that is why they are getting hot, have you checked for this.

Regards
Arthur

They may be, but I don't own or have access to a 'scope. Not sure if there is another method to determine oscillation??

Uh........measure the current draw at "idle". Compare it to the value in the data sheet.........

Jocko

Some current and temperature measurements taken for three experimental opamps in the stock Musical Fidelity A324 I/V circuit (shown again below):

AD843 (two DIL-8 Brown Dogged pairs per stereo channel): 23.5 mA / 112F (44.4C)

AD797 (two DIL-8 Brown Dogged pairs per stereo channel): 15.8 mA / 97F (36C)

AD825 (two SOIC-8 Brown Dogged pairs per stereo channel): 12 mA / "cool" (room temp.)

I also spoke with two separate Analog Devices "opamp engineers" who both concurred that the AD843 is probably the better choice here (despite it running "hotter") because it is unity gain and FET.

Both also concurred that the AD797, a bipolar design, is trickier to implement in this I/V appl. because it is not unity gain, susceptible to current noise and, hence, may oscillate.

Curiously, and as noted in my prev. post (see fig. 39 below), pg. 14 of the AD797's datasheet specifically details I/V DAC use for this opamp.

Sound-wise…

I'll hold off a bit on a comparative "review" of the three opamps noted above. This is because the AD843s I have haven't fully been broken in for 100 hours as the others have. But based on about 20 hrs of break-in time, the 843s are more forward-sounding than the both the 797s and 825s. This may give one the impression that 797s and 825s "image" better. Both the 843s and 797s are better-resolving then the 825s wrt detail. Not surprising given the $ diff.

I've incorporated CarlosFMs suggested mods (including snubbers) -- alas w/o oscilloscope analysis/tweaking. IAC, this did improve sound. However, I think MF's stock I/V circuit (which is almost purely based on the PCM1738 datasheet suggestions for circuit design and component values; see below) probably needs to be tweaked for each opamp change. I've already noted one for the AD797 above (addit. 100R in series with stock C(L), shown again below).

Question:

For the AD797, is it best to use the MF A324's stock 3300pF C(L) or change it to a smaller value (20pF to 120pF) per the 797's datasheet (see fig. 39 below)?

Fig 39:


Stock MF A324 I/V circuit:

Hi,Carlosfm and everybody
My first post here.Could it be done only with 1x THS4032 for both left and right channels?
Before reading lots of posts about THS,I want to give it a try with my DIR9001 + PCM1798 DAC for I/V stage and direct output.
Being use 4x OPA132 for I/V and OPA134 for buffer,any suggestion and comment would be appreiated,thanks!

Some opamps don't like 3300pF feedback cap while working as IV-converter and start oscillating
You can try a snubber at the output then, or lower the cap.
Any DIL8 opamp that gets hot is oscillating, except AD811 that gets hot anyway.

I'm limited for space in a portable headphone amp. This amp uses a separate buffered ground channel. Is there anything wrong with JUST using the "very small cap across the rails (1nF MKT)".Carlos/Anyone: Have you ever calculated bypass values for the AD825?