EEVblog Electronics Community Forum
Electronics => Repair => Topic started by: m k on December 30, 2021, 06:14:49 pm
-
Two controllers
V503/604 working ok.
V504/605 no keyboard.
Swapping softwares and ok controller is still ok but other one turns to totally silent, blank screen, nothing.
With original software the keyboardless controller is measuring just fine.
What I can't see, minimum software version for that board number?
-
Dang, I've been sloppy and clearly forgotten something.
No magic.
Even that Isolator has all address lines in use those 2 highest ones are inverted.
(Manual 608-9, 2nd par, "Two inputs...")
So Isolator is off when front panel is on.
-
Do any of the boards actually have a minimum software requirement? - it seems to me that Fluke solved these kinds of issues by making all the boards compatible at the hardware level, but I may be wrong...
-
V506/607 software is also measuring.
I didn't have any 27C64 chips but 27C256 is also fine.
Just quadruple the code and drop it in.
Chips I found were TMS 12.5V/100us but I used 500us pulses, just in case.
My take for the software thing is that the new version just dropped some initial checks off to make space for added code.
So if machine is totally dark with old software the first easy check is software upgrade.
This device has also ohms and current problems but that's because DC Sgn Cond relays are not in order.
-
Since the last time some other things have happened but first thing first.
That address line was probably not dead, possibly my CRO just didn't show it.
The scope is standing under the table in a carriage so brightness is quite low.
After changing the chip I obviously fiddled with some knobs and the signal was definitely there.
Used digital voltages of the meter are of course a bit different than usual 5V so can't say for sure how the original measurement situation was.
Current measurement is clear though, digital ground is used, signal is correctly oriented and time(X) is so slow that output pulse is stable and clear when brightness is high enough.
Other stuff then.
When measuring something with the other meter its display started acting up.
One segment became bright and when certain number was visible elsewhere that bright one disappeared.
When continuing with it the display started acting up more and more and finally went completely dark, few times it came back after some cool down.
Front panel is a bit difficult to troubleshoot but voltage test points(TP1 & 2) are easily available.
Low and behold the voltage was not missing, it was way up and so I swapped the power from the other meter.
The bad power was already operated earlier, its 5V regulator was missing, so regulator was added and voltage adjusted.
Regulator is not what is advised, I didn't have one but had a lower current one so used it, it's still of course uninsulated TO-220, working fine and not even warm.
Somebody had marked 5V trimmer bad but it was just fine, I thought it was changed, can't remember how closely I checked its joints.
This time that trimmer wasn't fine at all and since I didn't have a replacement it was replaced with a single resistor.
When the trimmer was off the board it worked just fine, I even put it back and took it out again, same thing, somehow it lost its stiffness when soldered in.
After 5V value correction the display came back and the situation was reverted to its earlier state.
The problem meter has addendums in the mail so for now the situation is on hold, again.
-
So is the issue still that the meter doesn't work with the updated software, but works fine with the original / older EPROMs?
-
No, the opposite and originality is not an issue.
It's also a controller only thing, swapping controllers will change things and changing software versions to faulty controller will change things.
Changing software versions to ok controller has no effect.
Since the last time I've done very little with software versions.
I did two new 506/607 sets earlier, from which one stopped working pretty quickly, later I did a third 503/604 set to check that original set wasn't the reason for darkness.
Can't remember exactly what else I've done but original set 504/605 is off and newly programmed 503/604 also.
Now keyboardless controller has new 506/607 set and the machine is measuring, another machine is reporting 503, most likely including the original set.
Seems that I haven't done much more.
Since my real DSO is still in store I tried DSO138 for IC6 measuring.
Free running pulses were very clear although a bit round, probe was P6101/A.
And still I couldn't trust that pulses were real.
-
I have had a lot of problems on several different controllers that behaved truly strangely.... some functions of the meter worked, others didn't, sometimes the meter couldn't boot, sounds like what you are seeing!
I ended up buying a new programmer and all my problems went away when the EPROMS were reprogrammed with the new unit (same software version 506/607).
One could easily imagine a "borderline" EPROM working in one controller, but not another...
These old EPROMs are almost like analog devices, not digital! :D
-
I remember that, miniPro couldn't serv, needed voltage was too high.
Since the ancient history I've actively tried to avoid too old ROMs.
Means that I've thrown away currant chips based on program voltages and manufacturers.
Nowadays it feels that good approach would be a minimal MCU module with micro USB for programming.
Down side is that none of those modules seems to have their USB connectors mounted vertically.
Pin headers are better than nothing but then there are those which way and what protocol problems.
-
Yes, it would be cool to replace the entire controller board with an Aurduino, LOL! :D
Maybe the effort wouldn't be crazy hard...
-
Not crazy I guess but probably harder than a science project.
Controller is also a bit misleading, at least these days, since it is an actual active part of the ADC.
-
Yeah, the ADC (despite being on a separate PCB) needs 100% handholding from the controller...
But the cool thing is, one could use the ADC differently and maybe squeeze out more performance, since we can effectively change its microcode?
-
I've heard that HP have also used same A/D parts for 3478 and 3457.
Maybe 6.5 is already a squeezed level for non monoliths.
-
Fluke is probably the only company that used the Recirculating Remainder principle... a kind of analog memory used as a step in the A/D conversion. Given the technology at the time, it was a pretty cool idea... It can probably be squeezed to more digits with modern components.
Who knows, given the component shortages and crazy prices, it may make sense to do something like this again one day, LOL! :D
-
So the missing keyboard seems to be a software addressing issue.
Pure luck here, without that LED board from Ohms issue I'd never have tried different software versions like now.
V506 machine is measuring its bootup range voltages just fine.
V502 and 503 are never booting up and are randomly disordered.
Software is checking its own things before anything else happens, so ROM and RAM can be addressed and nothing there is stuck.
A15 is also good or nothing would be seen.
One set of v502 bootups had some styles, now doing more regularly so, where last main digit was blinking 8, no dots, and all extras were blinking with it.
Means also that exponent display was 1+8.
Pressing VDC and blinking '8's turn to 'U's.
Pressing VAC and blinking '8's turn to 'd's.
Extra LEDs acted also, as expected.
That last main digit is connected differently, its data goes with extras and address is joined with main display's sign digit, exclusively.
So the key here is that it is on, since it clearly shouldn't, with extras.
And since its address is not used by extras the moment of those blinking things is wrong.
So front panel selector triple NANDs are on when they shouldn't, and again, do not follow schematics lines, check numbers.
(3rd data D-flop from top of 2of2 of front panel schematics is different)
(U3 lines are not side by side)
Front panel selector NANDs from top
IC
x-1- --11 3. extras data, 1st
x-1- ---1 2. address, prev ID7 must be 1
x-1- --1- 1. digits data, prev ID7 must be 1
x1-- --11 4. keyboard data in
So the problem can't be with those IC lines, it must come from the inner side.
I remember somebodies have argued around U18 but can't find it.
Addresses
U24 0000-1FFF
U23 2000-3FFF
U22 4000-47FF
U20 4800-4FFF ?
RAM test is expecting 0x55 from 0x49FF.
Stack pointer is set to 0x4800.
So 512x8 EEPROM, and as expected then.
Some SALEAE Logic Analyzer 24MHz 8CH things for _ICENABLE and its gates.
The analyzer is very strictly grounded so USB GND and logger side GND are the same.
For Fluke that is a bit no-no.
Luckily I found a bag of two TLP281-4 modules, only one LED was broken so it'll do.
...unfortunately it didn't do, when connected the machine didn't boot.
Inner side of those 4001 IC drivers were ok but outer side didn't like that opto isolator at all.
Luckily that earlier LED board was available and outer side of 4001 accepted that one extra LS244.
But..._ICENABLE is missing, not really since scope can see it.
So adding a 4050B in the middle...and still not co-operating.
Fully functioning other controller has the almost same logic picture.
No RAM is accessed at all, that is not possible so the used opto module is hopelessly too slow.
I thought it will show something since less than 1us pulses were visible, but clearly not.
More stuff is coming, this time from local seller and from local storage so delay can be expected to be shorter, maybe.
_ICCLK picture should show the first blink of IC6 with IC5, chip is changed, only once though.
A15 picture has whole V504 and V506 boot ups.
-
The situation of 503/604 software is actually a bit different.
Original ROMs have weak bits and can't be taken as a reliable set.
I did a ROM copier where receiving end is EEPROM, so a simple Arduino UNO is enough as a controller.
It can also be done with a single clock source and few gates, only sort of an obstacle is that comparing to other things writing takes quite long.
Now the copied set of ROMs are booting and measuring.
So must take back all that later software versions and better fault tolerance.
A side note,
do not underestimate the weakness of USB as a power source.
And if soldering Cat5 wall cable wires, don't forget that the wire insulation is not very heat tolerant.
-
Weak bits on these PROMs cost me a lot of time too... would strongly recommend swapping them with known good ones early on in the fault finding process, if you find one of these instruments acting strange!
-
The other thread has some good ones.
Do you have the very early versions?
I'm interested in V501/602 and earlier.
-
No, I don't have the really early ones. Could be good to collect them all, obviously, so we can eventually note the actual changes made...
-
I had no luck with other versions either.
All 8500A and 8502A ROMs are evading me.
That 8502AT V300 would be sort of special.
Since there seems to be only one thermal module set it's possible that this software is the first double, like A and AT.
Assuming here that 8502AT was before 8506A.
Also wondering why not 8503A.
-
The other thread has some good ones.
Do you have the very early versions?
I'm interested in V501/602 and earlier.
Here you go...
-
Thank you.
So I think we have found the beginning.
This is 501/601 and next is 502/603.
Here greetings text is also the very last of the set.