SR-52 flaky display

[Colin Verrilli]

Hello All,

Please excuse the OT post but not sure where else to go.

I have an SR-52 that I got new in '76 (as a freshman in high school) that I'm trying to resurrect.
Besides the gummy wheel problem, it has a problem with the digits. I haven't characterized it exactly, but here are some things I notice:

- when I have the degree / radian switch on radians, I can not enter any digits.

- when I have the switch on degrees, I can enter all digits and the display seems to work fine.

- I can do operations on numbers that do not contain 8's or 9's<nl>
- whenever I enter a number with an 8 or 9 and then press an operation key, the display flakes out

- also if I do some math on smaller numbers and the answer contains an 8 or 9 the display flakes out

- flaky display adds a bunch of 8's before the number and / or adds some digits after the decimal point. Sometimes there are gaps between digits.

- The flakiness seems to be not just at the display, but also in the operand registers of the processor since I can not undo the display mistake by doing the reverse math operation.

- happens both on battery power and with charger

Wondering if anyone has any experience with problems like this or any ideas.

Thanks,

Colin
Apex, NC

[ClausB]

When I repaired my SR-52's display I referred to both the SR-50 Maintenance Manual and the TI-59 Service Manual, as the 52's display is sort of a cross between the two. The manuals are available here:
http://www.rskey.org/CMS/index.php/the-library/12

My 52 just had a blank digit, nothing as wild as yours. Some of your symptoms point to a problem with the high bit of the 4-bit digit, called I/O8 in the SR-50 schematic. Also note the deg/rad switch in that schematic - it's connected to D14 through a diode and there is another diode on the KS line. Those diodes allow the deg/rad switch to close without interfering with scanning the keyswitches. Check whether your 52's switch is wired similarly. If so maybe one of those diodes is bad.

Hope this gets you started.

[Colin Verrilli]

ClausB,

Thanks for the response and the good suggestions. 

I'll do some digging into those docs, dust off my EE skills and see what I can find. I'm wondering if I need to invest in an oscilloscope.  I'll post back here with any updates. 

[Colin Verrilli]

Here's an update.

I tried replacing three of the 1N914 diodes with 1N4148.


This seemed to solve the Degrees / Radians problem, but now the digits problem is worse. The display is always showing with a 80000. If I enter a '1' for example, I get 80001. in the display. Doing calculations produces random stuff.


I've looked at the US Patent describing the 501 ALU and my best guess now is that it's the 501 chip itself that is bad. I'm going to try replacing it from another unit. Can I use any SR-5x model to source the replacement?


Colin.

Edited: 15 July 2012, 12:54 p.m. after one or more responses were posted

[ClausB]

Interesting that the diode replacement made the digit problem worse. Could there be a bad connection on the circuit board? Was there battery corrosion at some time?

Yes, I believe the 501s are the same in all SR-5x models.

[Colin Verrilli]

There was slight battery seepage, but not onto the circuit board. I'm not seeing any obvious damage on the circuit board just by looking. I have noticed that depending upon the angle of the keyboard to the main board, the behavior can change slightly. But I'm not quite sure how to check for hairline cracks in the connection bus.


I have a sacrificial SR-50A so if I can manage to remove 2 501's and solder in one without doing more damage, we'll see what happens.

Edited: 15 July 2012, 12:52 p.m. after one or more responses were posted

[Marcus von Cube, Germany]

Colin, could you please reformat your messages so that they reflow according to the window width? Using PRE tags makes reading unnecessarily uncomfortable because readers need to scroll horizontally.

Edited: 15 July 2012, 9:59 a.m.

[Colin Verrilli]

Ok done. Sorry

[Marcus von Cube, Germany]

Much better now! :-)

[Colin Verrilli]

Update:

I replaced the 501 chip with one from a working SR-50A.

It still doesn't calculate properly, but the behavior has changed significantly. I no longer get the repeated 8's or 80000. I can do some basic calculations using all digits. But there is weird behavior with many functions. For example,

- sometimes when I press CLR, I get 0 00 (00 exponent) and sometimes I just get 0

- sometimes I get blinking display just pressing =

- COS, SIN, TAN does not work

- y**x does not work

- x**2, ln, log does work

- trying to enter programs does not record all key numbers properly. SST and BST do work.



I plan to check for soldering mistakes and also try to check for any high resistance in the keyboard to main board connections. If that doesn't find anything, I'm about to give up on this one.

Will post again if any change.

[ClausB]

Hmm. Maybe a bad ROM, or a bad connection to a ROM? There's ROM in the big SCOM chip and also in the two small BROM chips. Sadly, the only donors for those are other 52s. As long as you're soldering a lot already, why not touch up all those chips' pins?

Can you test the RAM? Remember that RCL 00-19 and 60-99 can address all RAM locations in the two chips (excluding register RAM in the 501 and SCOM).

[Colin Verrilli]

Another update.
No luck with salvaging the original board.
I replaced board and key matrix from a donor SR-52 calculator.

Now my problem is with the card reader. I'm trying the one from the donor calculator. It required extensive cleaning. I removed the remains of the old gummy wheel. I then took a rubber faucet washer with the correct inner diameter but outer diameter too big. I used a dremmel to lathe the outer diameter down to about 8.5mm. Then I used epoxy to fit it to the motor shaft.
Re-assembled, the motor will barely pull the card through and the motor makes a whining sound. Not smooth at all. Perhaps the O.D. is too large? I'm going to take it apart and try running the motor separate. Also I may try to reduce the diameter a bit.