Progress!
So I went to try and test the DAC reference resistors in-circuit to try and determine which ones would be more suspect. To start I measured across the two 10.0K resistors, and one was measuring around 160K and the other was even higher around 300K I think it was, so they had to be bad. A good start would be to pull those out, one at a time. I pretty much routinely solder 1206 SMD parts by hand with a fine tip for one of my commercial projects that uses a few of these on the board, done with a T7 fine tip and a Weller 25W pencil iron "dimmed down" with a light dimmer to reduce somewhat the tip temperature. First I scraped a bit the dull solder on both ends of the resistors, and then added a bit of fresh solder, and after doing this on both ends the resistors would stand up on one end almost like it wanted to come off with the tip and without almost any effort at all. Luckily these parts don't seem to be held onto the the board other than with the two solder junctions, no other additional bonding stuff as it sometimes happens. So minimal stress to the board pads.
Went first to pull R2012 10.0K, but by mistake ended up pulling R2016. Ooops. Even tough it initially had some serious crud on the ends, turns out it appeared to be still ok, a quick check measured 1.001K, so not far off if anything but now that it is out will definitively replace it. Next I went to get R2012, one of the 10.0K resistors, which this time was the correct one, and came out very easily as well. But this one measured completely open. Same was the case of R2013, the other 10.0K resistor. Two for two, so far so good. So I figured now that I have these 3 resistors out of the network I could probably make some in-circuit tests of the remaining ones. R2011 a 4.75K resistor measured very close at 4.749K, so leave that one alone for now. R2015 the 82.5K one measured 60.6K, and R2014 221K measured 64.2K across it. Since both these resistors are connected together as well as to R2010 the 20K trimmer, it would be expected those readings to be way off their actual value. But where they still correct? So I took my previously "protoboarded" 5% resistor network simulator and pulled out the two 10K and the 1K resistor, and measured across the 220K, and it measured 60K, and the 82K resistor measured at 63K. That is very close to what I am getting on the A5 board so for now will not disturb them. BTW R2520 and R2521 which are 681-ohm 0.1% resistors used to convert the output current of the DAC to a voltage, in circuit measured 687 and 691 ohms respectively, so for now I will consider them good enough as well, and at least no opens.
So getting back to my project that uses 1206 resistors, by chance it happens to use some 10K and 1K units as well, and I had plenty of them in the parts drawers. Only thing these are 5% tolerance, but I though, what the heck, just for temporary testing purposes they are certainly better than an open resistor. So in went two 10K and one 1K 5% (I hope nobody pukes, probably the nice purist guys at Tekgroup probably would at least cringe at this, so maybe for now will keep it my little secret). Also reinstalled the temporary leaded elco filter caps (pulled from an old board, but checked for ESR and value) and back in went the A5 board. See attached picture of the reworked board with the temporary resistors and caps.
Before power up I held the tip of the DMM on the the +1.36V test point (I found a convenient location to be at the positive end of C2420). At initial power up it was still showing the same +0.26V as before, -oh crap no progress at all. But then 1-2 seconds later the screen readout came to life! I measured again and I now was getting +1.3848V, not exactly "on-spec" but close enough for now. And the -1.25V ref voltage (measured at pin13 of U2420) was at -1.2688. Wahooo!
On startup the scope cycles through all the test steps lighting up the front panel indicators in sequence, and at the end no errors are shown.
So, looks like I made some progress after all! Calibration will for now obviously be way off, but I am not going to mess with the +1.36V setting until I put in the proper tolerance resistors. No sense disturbing that setting at this point. On the last picture (ITS ALIVE!!! - remember the movie Young Frankenstein 1974?), for purely academic reasons I set the timing Alpha time cursors centered onto 2 divisions, and with the time base set at 500nS its obvious the resulting 973nS reading (which should have been exactly 1000nS) tells it all. But for now, today was a good day and I am happy with this progress. Will be placing the A5 parts order at Mouser tonight!