Using the AD844 as an I/V

[abraxalito]

That 1kHz SQ wave has the pre-ringing of a linear phase filter on it - what's the origin of it? Also since your 20kHz sine doesn't show any steps that would tend to indicate you're using an OS filter.

[georgehifi]

I can switch from Cary's own dsp module which I believe has the DF1704 in it and it does upsampling to 96/24bit or via the remote go to the PMD200HDCD chip. Which stops the upsampling, I cannot remember which it was on when I took the pics.

Cheers George

[abraxalito]

I'm surprised that the output SR is as low as 96kHz - those 'spiral' oscillations on the 20kHz sine look to have an envelope frequency around 16X which to me suggests the PCM1704 is running at 384kHz (8X OS).

Have you tried running NOS? You'll notice an improvement in the dynamics.

[georgehifi]

I prefer the PMD200 on HDCD recordings. (is this nos as the upsamling light is not on just the hdcd light)
On standard redbook I just prefer the upsampling mode which the upsampling light comes on only.

As far as doing NOS I would not know where to start or if it is possible attached is a pic of the board.

There more info on the first page of the filtering on this Stereophile review here. Measurements came up good too
http://www.stereophile.com/cdplayers/504cary/index.html
Cheers George

[abraxalito]

I see - going NOS in that set up would be too great a challenge, you'd also lose HDCD decoding. You'll be wanting one of iancanada's boards to interface the PCM1704s.

[georgehifi]

Changed the TZ resistor/cap from 2.7k/500pf to 3k/1200pf to give me a touch more gain and filtering to see if it effects the spiral oscillation (you were right they were the steps as looked without any cap and the spiral turned into clear steps) and didn't change with either the DF1704 or the PMD200 switched in.
With the 3k and 1200pf (-3db @ 42.3khz) on TZ, all is clear now with the 20khz sine wave, just the smallest hint of spiral remains if zoomed in on. And I'm just a touch down on frequency response approx 1dB at 20khz. I'll have a listen later, probably won't hear any difference with this change, but it looks better on the scope and that's a good mental thingy.
While I was at it as you know I'm using the 844's buffer straight to the rca's which should be by the data sheet 15ohms output impedance, with no load just the scope a 1khz test wave droped 25% with 100ohm loaded, this tell me AD is a little off with their specs of 15ohm.

Cheers George

[georgehifi]

Abraxalito, please educate me some more, I listened last night (I will do some more tonight) to the above setup TZ with 3k and 1200pf which gave a touch more gain and 1st order filtering at 44khz.
The thing that I noticed again is the bass, seemed to be less defined still had power but less definition, more softened tube like. If I didn't know better it's as though the i/v stage cannot drive the buffer stage as well with this 3k and 1200pf.
The first time I listened to it Pedja's way with 1K and 1000pf ( which I must try it again) I remember I was impressed with the slam and bass definition even though I did not have enough gain with quiet recorded cd's, is it possible that with the the 1k and 1000pf that the I/V stage has better drivability into the buffer than with the 3k and 1200pf?
Maybe this needs to be simmed to see what's going on. But I will have another listen tonight to see if it is what I heard last night.

Cheers George

[abraxalito]

Hi George - I don't have any experience with open loop TZAs so I'm just guessing here. The grounding point of the parallel RC on this node will be very critical - if you have a groundfill then it will be very hard to decide on the right point to ground it. Why not experiment with different time constants - try with the 3k resistor but only 1000pF and see how that sounds. If this is like my experiments with LM6172 then smaller caps sound better.

I very much doubt that simming it will shed any light on why it sounds different - been there, tried that with my AD1955 DAC about 3 years ago.

[AudioLapDance]

Somebody needs to sim this beast!

SPICE Models | AD844 | Analog Devices

ad844.jpg

Dirkwright simmed an 844 in Tina (just as regular IV)

Originally Posted by dirkwright

Ok, I modeled this circuit in Tina using Analog's spice model for the AD844....

Hey George, do you wanna give learning Tina a try? It's a powerful tool!

[jcx]

LTspice is good too, some say better if you are looking for "an engineer's Spice"