2015's calculator anniversary - the SR-52

[Gene]

This year on September 16, the 40th anniversary of the SR-52 will arrive. In many ways, this large brick of a machine was groundbreaking and under-appreciated today (IMO).

Consider the value brought by this machine at the the time compared to the HP 65 (which I also love)... for 1/2 the price:

1) Line numbers - try editing on the HP 65 - it is painful!
2) 20+ memories - compared to 9. More memories in fact (STO 98 and 99 work). In total, you can have up to 50 STO and RCL friendly memories.
3) Indirect references - with 20+ memories, this finally makes sense.
4) Insert editing - made things much easier than on the HP 65.
5) Available printer.

Yes, I am aware that it has many shortcomings:

1) Mechanical - have one with a working card reader? I don't!
2) Functionality - No FRC or INT, no OCT or DEC, etc. Some things are missing.
3) Size. This thing is a giant!
4) Mechanical - see #1

But, it did bring to market some great advantages at the time for budget minded people. It was my FIRST programmable. I bought one used for $70 in 1977.

Three cheers (ok perhaps just two) for the SR-52. Without it, who knows if the HP 67/97 would have been as good as they were. :-)

More info on undocumented features can be found here:

SR-52 features

[jebem]

Thanks for bringing this Texas landmark into the light, Gene.

I have been looking to acquire a SR-52 since at least one year ago, but failed so far.
As a result, I have acquired the easier to find TI-59 and a few ones more from this TI-5x series.
Well, I can't quit this quest, so let me go back to the auction sites...

[isanchez]

HI,

I have a sr52 with all the stuff: box, case, standard pac, maths pac, stat pac, ee pac, books, programming forms, etc.

I repaired the card reader and it works fine but it took quite a lot of efforts, as I didn't find the proper o-rings (standard size), like for the hp41, 67, 65, 97...I had to use a dremmel to "resize" the new tracking roll inside the reader to the proper diametre...(by the way, I don't know the English term, but if it is of help, I used one of those solid rubber 10mm diam. rings used in the faucets)

About the hardware quality, it cannot be compared with HP at all (I don't talk about the software).
Inside everything looks fragile (card reader quality is an example), battery terminals, the cards are very sensitive to humidity, etc.

...but I love it!

My best wishes for all you this New Year.

Ignacio

[jebem]

Hi, folks!

One year later I got my first Texas SR-52 too

Great beast of a machine! Fantastic! And of course it is not working.

My questions to the community:
Where or who can provide me a service manual?
What about spare parts providers?
Or any repair tips that you might have or know about would be highly appreciated.

I have repaired a number of Texas machines in my time, including the 59, but I have no knowledge of this one so far.

This machine was serviced at least once by the previous owner(s) and at least one unwanted mark was left inside.

Machine was swiftly and easily opened, as usual on Texas machines. After a brief initial visual inspection:
- The battery leaking corrosion have attacked the PCA, starting from the battery terminals.
- The battery electrolytic is visible in the card reader pinch roller and rubber wheel; also the card insertion switch sensor have small signals of corrosion as well.
- The BP-1A battery pack was rebuilt but is dead and leaked badly.
- The plastic frame that goes between the keyboard and PCA is broken and a piece of it is missing, typically of a careless dismantling by the previous owners or service.

The BP-1A battery pack was in fact nicely and professionally rebuilt with some red Ni-Cd cells "Made in Hong Kong", so it probably was done during the mid 80's or so.
However these batteries have leaked as well, corroding the PCA.
After opening the case, I have found one of the battery terminals lose inside.

The card reader looks like it may work after a good restoration.
The roller rubber is dry but still has traction, despite the horrible motor noise.
The R/W head four coils seems to be all fine:
L1 Orange - 742 Ohm, 0.15 Henry
L2 Green - 750 Ohm, 0.17 Henry
L3 White - 754 Ohm, 0.17 Henry
L4 Yellow - 756 Ohm, 0.18 Henry

Applying 3.6V from a power supply for just a few seconds gave me a satisfying "0" on the display, but as soon as I press any operators it goes blinking.
So the machine is faulty.
I will do a full cleaning on the PCA as usual and see if it helps, but I'm afraid that one or more IC's can also be faulty

More news to come one of these days.

[jebem]

Ok, just for fun, I'm posting some pictures of the dismantling.
And a few more notes as well.

Did I mentioned that opening a Texas is extremely easy? Yes, it is. And safe, too, unless one's absolutely careless.




Now the good surprise. The motor assembly (motor plus integrated gearbox) was "Swiss made".
Switzerland, the land of precision mechanisms among other good things (I like their milk chocolates too).

Card Reader motor engraved details:
Model: escap 15
Ref: C 11 110 16
Date code: 1 76
Gearbox Ref.: G 16.1 84:1

How cool is this?

[Geoff Quickfall]

nicely done!

Switzerland, the land of chocolate-clockwork-motorized-watches. There covered it all!

[jebem]

I realize now that I have hijacked Gene's thread.
Should I create a new one?

Small update on my SR-52 repair:
- After PCA cleaning the issue remains the same, despite some variations, but in the end I allways get a crash (the two minus indicators are lit and requires a power cycle to recover).
- I'm using a 4VDC external power source connected to the battery terminals:+Vss (common) and -Vbat.
The Internal DC-DC converter seems to be good (Vdd = 10.5V, Vgg = 16.3V).
Although I still have to check the power lines noise.

I started my quest to get an SR-52 Maintenace Guide. I will get one, no matter what.
I know that we can use the TI-59 and the SR-50 service guides publicly available as a reference, but I like to have the correct documentation whenever possible.

After a few mails sent to Gene Wright, Joerg Woerner and Viktor T. Toth, I have a few more ideas.
Thank you all for your fast answers with tips and experience sharing on these machines.

Basically, the cleaning must be deeper than what I have done so far.
Most probably I will take extreme measures by removing all the IC's to clean under them.
At this point, I have nothing to lose. If an IC is defective, I would have to replace it anyway.
Problem is that I can't find any spares for these machines anywhere.

[Gene]

The best "spares" may be buying another one or two SR-52 models, regardless of condition, off that auction location, but I'm sure you thought of that. :-)

[jebem]

(01-25-2016 11:25 PM)Geir Isene Wrote:  I have a spare...

Thanks, Geir.
Let me first try to eliminate the usual causes before accepting that a IC is faulty.
Then I will get back to you and see what's your offer.

[Don Shepherd]

Thanks Gene. What a good-looking old boy!

[jebem]

Small update.
OK. Time for a deeper dismantling here.

Keyboard removal.
Found two corroded wires of the "flat cable" assembly, and one third broke when removing it. Not sure if it was broken already, but my guess is that it was about to break.

These corroded solder joints are the most difficult to desolder, as the iron heat is not easily transferred to the joint to melt it in the required in short time. The longer the iron stays on the joint, the bigger the danger of damaging the copper traces and components.



I did check the power supply lines for noise. Zero instability or noise.
Even so, I removed the caps for a off circuit checking.
These vintage Siemens electrolytic capacitors are of the best quality at the time.
All of them are fine. And all made between 7602 and 7604 (1976 week 02 and 04).

So this is curious: An American brand, USA made machine, using Swiss precision motors and German Siemens capacitors.



And then it was the turn for the IC's extraction.
The task is easy although laborious and delicate.
This time I had to extract the chips without destroying them and keep the PCB copper pads in good shape as well, for the next chip insertion, only because I do not have spares and I need to reuse the chips.
Otherwise I would just cut out the chips and desolder the legs remains from the PCB.

So each IC took me about 5 minutes to extract. As I said, this is a delicate job, considering the vintage state of the PCB and its, let us say, average quality.



The two piggyback PRAM chips looks a little suspicious.
One leg is missing there. Or maybe not.
That's why I need the service guide (or another SR-52 to compare with).



The bigger DIP28 SCOM and ARITH chips are easier to extract.
I used a very old dog trick from the 70's, when we had to extract the chips to be tested off circuit: After desoldering the pad, used a small screwdriver to carefully rock the chip legs until they get loose in the pad hole.
In this way the chip comes out easily.




I decided to extract the card reader controller as well.



Partial PCA dismantling general overview.
I left the power supply oscillator and the 4 NAND chips in the circuit, as they don't seems to be related to this calculator problem.




Next steps:

I found a couple of solder joints with marks of battery corrosion at the keyboard flat cable and on the ARITH chip, so the new soldering will fix it.

I'm planing to install DIP sockets and see if the extracted chips can work in this way.
The issue here is that the chips legs were cut off at the manufacturing plant and I still have to check if they can be used in a socket.

[jebem]

Small update on Keyboard dismantling.

Well, the IC legs are too short to be inserted on sockets, so they will have to go back to their original positions in the PCB.

Meanwhile, I saw some marks of corrosion at one of the edges of the keyboard.
Also one of the previous owners had the keyboard connector re-soldered, and in the process they broke the supporting plastic frame.

Removing the adhesive plastic film revealed the PCB routes.




It was an opportunity to draw the KB schematics, by using direct observation and with the help of an ohmmeter to check the connections.

The circuit traces presents about 90 Ohm per cm. The longer circuit traces measured more than 1KOhm.



After some head scratching trying to make sense of the above circuit, finally the Texas design was revealed.
The KB matrix uses 6 Rows per 10 Columns for a total of 60 keys, where 45 key actually were used.

[jebem]

Saturday morning, 6h00 in the morning.

Time to reinsert the IC's to their original positions.
While I'm at it, have re-soldered most of the PCB pad joints. Some had marks of battery electrolytic corrosion.

Then have checked all resistors, capacitors, diodes, transistors and coils.
All were tested good, with exception of the battery charging limiter resistor, and that doesn't contribute to the current calculator fault.

This limiting resistor has signals of past overheating and its value is way out of the tolerance. It is marked Brown/Orange/Black /Gold (13Ohm 5%) at 1/2W, and the current value is 17.2Ohm.
Despite its value is not that critical, it must be stable. So I replaced it with a 15Ohm that I had at hand.




Details of the Card Reader photocoupler sensor.
Contrary to the so many bad critics found in the forums on this card reader system (which I respect) I like its simplicity and easy of access.
I'm not saying it will be easy to replace the motor roller wheel, though.
That is the weak part of this design imho. The roller pressure is not adjustable (as we find on the HP card readers) and this alone creates a huge issue when restoring this mechanism.

So the photocoupler is a classic diode / photo-transistor combo. Both are in good condition.
The photo-transistor presents about 56KOhm under fluorescent light and about 190KOhm in the dark (measured in circuit).




14-digits 7-segment LED display details.
Printed labels: "UNIV SR" "EL SALVADOR 7551" "234 G"
The date code suggested it was made in 1975 week 51

[jebem]

Reassembling the keyboard PCA takes a steady hand and lots of patience.

I had to remove the metal domes from the original auto adhesive plastic film.
Then drop and align the domes one at a time, row by row, on the KB PCA.

To fix the domes in place, have just used regular (1.5cm wide) transparent adhesive tape.
After finishing with the 5 rows, have cut the excess tape around the lateral sides and applied a second layer of a wider (5cm) adhesive tape to seal the keyboard contacts.

A brief electrical test confirmed that all rows and columns were isolated between them.
Testing a couple of key presses confirmed good electrical contact as well.




After soldering back the KB PCA, did some more testing.
I'm using an external regulated power supply at 3.7V.

The initial random issues are gone, probably due to the cleaning and re-soldering.

After powering on from a cold state, this is what the SR52 can correctly calculate:
- All single X register functions (lnX, logX, x!, 1/x, √x, x2, sin, cos, tan, asin, acos, atan);
- D.MS and D/R both ways (using INV key);
- Degrees and Radians switch afects the calculations accordingly;
- P/R (polar/rectangular) works as well - in this machine it takes Memory Register 00 in addition to Register X;
- The four basic arithmetic operations are fine when using the PROD (x and /) and SUM (+ and -) functions for direct register arithmetic on any of the Memory Registers 00-19.
- Storing and recalling numbers to/from Memory Registers 00 to 19 and 98 and 99.

However, two issues remains to be fixed:
1.- The machine stops responding to keystrokes about 15 minutes after power on.
2.- After powering on from cold, this is what the SR52 is failing to do:
- All operations requiring more than the single X register (˟√y, y˟ , x, /, +, -);
- In LRN mode (programming), any key-codes appears as "00";
- Trying to store numbers in the Memory Registers 60 and 70 will just recall "00" as well.

The usual forensics test: 9 sin cos tan atan acos asin returned: 9.000004661




69! testing:



Degrees selected. 45 sin looks to be OK.
(I also tried sin (Pi/2) in Radians mode and the result was 1)




Idle current consumption around 160mA @ 3.7V.

[jjohnson873]

Jebem. Wonderful explanation and excellent pictures! I thoroughly enjoy the troubleshooting and the diagnosis you're doing on the TI calculator. Keep up the dialogue and let us know if you get it back to 100% working condition. While I only work on vintage HP calculators, any analysis of a vintage calculator is fascinating.
Cheers! ~ Jim ~

[d b]

Jose;
A final refined version of this would be both valueable and easeir to find as an Article.

[jebem]

(01-31-2016 01:34 PM)jjohnson873 Wrote:  Keep up the dialogue and let us know if you get it back to 100% working condition. While I only work on vintage HP calculators, any analysis of a vintage calculator is fascinating.

Thanks for your kind words, Jim
(Loved your NP-25 "joint venture" with Chris; today I was watching again some HHC 2015 movies on YouTube, including your Fritzing/NP-25 project presentation, after a search on SR52 subjects)

I'm waiting for some "spares" in the form of a SR-51A bought locally two weeks ago for 5 Euro only (the seller lives in another small village and keeps telling me that he is very busy to go to the post office). Guess that I will have to wait or drive there and collect it.

According the DATAMATH site information on the different models, the SR-51A shares the same ARITH chip with the SR-52, but nothing else concerning the SCOM/ROM chips.
I believe the SR52 symptoms can be explained by a faulty ARITH, but this is really a guessing game right now.

(01-31-2016 02:01 PM)Den Belillo (Martinez Ca.) Wrote:  A final refined version of this would be both valuable and easier to find as an Article.

Hi, Den,
Thanks for your idea.
Well, this restoration job may take longer than I initially anticipated.

Somehow, I knew this would not be a easy task due to the fact that this model seems to be difficult to find, even "untested", and then the rocket asking prices are a real issue for my bank account.

[jebem]

Card Reader motor / gearbox / roller wheel.

The roller wheel rubber was dry and falling apart (not gummy as we see on HP card readers).

Before removing the rubber from the wheel/shaft combo, I took some readings.
Here is the motor schematic drawing. Dimensions in mm.
Not really sure about the correct outer diameter of the rubber wheel, as it was deformed.

[jebem]

Today I have replaced the the ARITH/ALU chip from a donor SR-51A.

The problem remains the same, so my next suspect are the RAM chips.
In fact the RAM chips are usually the first ones to fail when over-voltage is applied as a result of using external adapter with bad or absent batteries.
I found this issue on a TI-58C some time ago, but was lucky to find RAM memory spares on TAS and could fix it.

But for this SR-52 finding spare RAM chip seems to be hopeless.

The good news is that the suspect ARITH/ALU TMC0501NL from the SR-52 is working fine when installed on the SR-51A

Also, I have to say that these Texas PCB's from the 70's are much better than I already knew. This PCB have resisted to two chip extractions and insertions without a single pad being destroyed.

Meanwhile, our good friend Etienne Victoria came to rescue. Thank you!
Etienne has offered me a SR-52 working chassis that is coming my way.

I have been reading some of Etienne's past work posted here in the MoHPC in the old forums and have been enjoying it.

With two machines I hope to build one working calculator. The card reader will be another story, though.

[Etienne Victoria]

Thanks for your kind words :-)

For the card reader, Fsoft has actually a repair procedure way better documented than mine here.

While I'm at it, I opened my SR-52.



Just like you, I measure the motor axis at around mm 4,3 which should be fine to fit a rubber roller between mm 3,0 & 3,8 without having to glue it.



I use a washer ajusted to inner mm 3,2 as seen below (no glue). Fsoft uses mm 3,0.



As for the outer diameter, Fsoft used mm 9,0 which is IMHO way too large. I think this would put excessive stress on the motor.

I use mm 8,2 which works.

But I'll check on my other 52s before posting back.

All the best for the electronics repair before you get to the card reader. This one is a monster :-)



Etienne