Monroe 650 Calculator repair and cleanup

Welp, anoth­er cal­cu­la­tor tale of woe. This time it is a Monroe 650 cal­cu­la­tor found on eBay. The 650 is the top-end mod­el of Monroe’s 600 series cal­cu­la­tors. It is a 16 dig­it Nixie tube dis­play cal­cu­la­tor with addi­tion, sub­trac­tion, mul­ti­pli­ca­tion, divi­sion, square root func­tion, and two mem­o­ries. It oper­ates in fixed dec­i­mal or float­ing point modes, and has 5 chang­ing indi­ca­tors spaced every incre­ment of the base-10 expo­nent of three from the dec­i­mal point.
The price was right and includ­ed free shipping.

What is it with some eBay ship­pers that think their job is done once the item is sold.
Once again a sell­er packed a del­i­cate item in a thin 29 lbs./inch rat­ed card­board box that was too small for the cal­cu­la­tor, and then use a sin­gle lay­er of medi­um-den­si­ty foam which is much too stiff to pro­vide much of any shock pre­ven­tion.
This is what hap­pens when that is com­bined with FedEx Ground shipping.

The oth­er side was not in any bet­ter con­di­tion. And then to top it off com­mu­ni­cat­ing with the sell­er was frus­trat­ing and total­ly use­less.
It’s great that I like fix­ing things, because I def­i­nite­ly had quite a bit to fix on this project.

The first step of this project was to open up the unit and check for any dam­age from it’s recent ship­ping expe­ri­ence. The two case halves are held togeth­er by four machine screws, one on each cor­ner. With the unit fac­ing for­ward, the upper case folds up from the back and is hinged at the front edge by numer­ous cables.
In ear­li­er revi­sions of the 650 cal­cu­la­tor, the top case and key­pad could be eas­i­ly sep­a­rat­ed using the pro­vid­ed con­nec­tors. This revi­sion has the small daugh­ter board with three sol­dered wires con­nect­ing the board to the key­pad assem­bly. I found it eas­i­est to unsol­der the wires from the key­pad to sep­a­rate the two halves.

Other than being a bit dirty inside, I did­n’t see any phys­i­cal dam­age to the tubes and boards. I did pull out about two dozen small plas­tic shards from the inside. Most of the shards were from the epoxy used to orig­i­nal­ly join the case pieces togeth­er.
All the capac­i­tors checked out fair but not great when test­ed in-cir­cuit. After 51 years of use they were def­i­nite­ly due for replace­ment, so a list was made and an order placed. The pow­er sup­ply uses main­ly radi­al for­mat capac­i­tors, while the oth­er boards use only axi­al for­mat capacitors.

The cal­cu­la­tor uses Hitachi CD81 minia­ture Nixie tubes along with a strik­ing sky blue wire sep­a­ra­tor and tube hold­er fix­ture. There are 16 Nixie tubes and 21 Neon lamp indi­ca­tor dis­plays.
While wait­ing for parts and mate­ri­als to arrive it was now time to dis­as­sem­ble the unit and clean it inside and out.

The cal­cu­la­tor mea­sures 12 ⅜ x 12 x 3 ⅟₄ ” (31.4 x 30.5 x 8.26 cm) which is much larg­er than the cal­cu­la­tors that I have pre­vi­ous­ly worked on.

So I had to make a deal with our house cat for some cat­nip in exchange for bor­row­ing her kit­ty-lit­ter con­tain­er. Yes, I did clean it out well before using it for the retro­bright pro­ce­dure, which required two gal­lons of 12% hydro­gen per­ox­ide solu­tion. Also the bat­tery is not ful­ly in the solu­tion, and was used to dis­place some of the liq­uid to raise the over­all lev­el. Heavy chromed sock­ets were used to keep the case from float­ing due to all the bub­bles. In full sun­light each case halve took around 4 hours to complete.

The key­pad assem­bly took the longest to refur­bish. The foam pads below the keys had com­plete­ly dis­in­te­grat­ed leav­ing a gooey adhe­sive tape attached to the key­pad frame. All the keys were removed to facil­i­tate cleanup.

The key­pad is a mag­net / reed-switch type and has a smooth feel when cleaned up. Of spe­cial note, all of the keys had a light glue applied when orig­i­nal­ly assem­bled and take quite a bit of force to begin to break them free. Once the ini­tial move­ment to break the glue had start­ed, I switched to a small flat blade screw­driv­er to sep­a­rate the key­cap from the met­al posts by alter­nat­ing front to back posi­tions to gen­tly sep­a­rate the cap. The met­al posts and mag­net assem­blies will gen­er­al­ly fall out the back side of the frame once the cap and spring are removed. Use cau­tion not to let the assem­bly drop too far, as this may dis­lodge the mag­nets, which I man­aged to do just once.
The key­cap for the con­stant “K” key is not nec­es­sary to remove, and I end­ed up break­ing the cap from the glued post, but it was easy to epoxy back on.

All the slide switch­es were cleaned with DeoxIT D5 con­tact clean­er before reassem­bly. The key­press counter and dec­i­mal point switch were both in great con­di­tion and just need­ed some exter­nal cleaning.

Instead of using a foam strip for the key­cap stops again, I end­ed up using some 0.040″ 40 durom­e­ter (extra soft) Sorbothane, Polyurethane Rubber Sheet. My ini­tial test­ing showed that this mate­r­i­al did an excel­lent job of cush­ion­ing the key­cap press and was near­ly noise free. The only unknown is how long will it last.

After a good scrub­bing with my favorite cit­rus based clean­er and a tooth­brush, the keys were then soaked in some pre­vi­ous­ly used 12% hydro­gen per­ox­ide solu­tion overnight before re-installation.

The replace­ment capac­i­tors arrived ahead of sched­ule, so it was time to tack­le the pow­er sup­ply sec­tion of the calculator.

The DC reg­u­la­tor board is a sin­gle sided phe­no­lic com­po­si­tion with the com­po­nent leads bent over on the traces and then hand sol­dered. The replace­ment elec­trolyt­ic capac­i­tors went well and were most­ly small­er than the originals.

On the pow­er sup­ply AC rec­ti­fi­er board I replaced the orig­i­nal ceram­ic 1 nF AC line to ground capac­i­tors with some new X1/Y2 safe­ty capac­i­tors of the same rat­ing. I did find one cracked trace that looked to have hap­pened recent­ly in ship­ping, which was quick­ly fixed by reflow­ing some new solder.

Now that all the elec­trolyt­ic capac­i­tors had been replaced, I took the time to re-lace two of the wiring har­ness­es before start­ing reassembly.

The Monroe 650 cal­cu­la­tor is inter­nal­ly a slight­ly dif­fer­ent fea­tured ver­sion of the Canon Canola L163 cal­cu­la­tor. The 650 was built specif­i­cal­ly for Litton Monroe by Canon of Japan, with just enough changes in fea­tures to dif­fer­en­ti­ate the two mod­els. The 650 lacks the per­cent­age func­tion of the L163, but gains addi­tion­al mem­o­ry func­tions, and is faster.
The 650 uses a Texas Instrument chipset is com­prised of sev­en IC’s using 10-micron process, 1‑metal P chan­nel MOS tran­sis­tors.
The chipset includes the TMC1766 key­pad inter­face, TMC1763 dec­i­mal point selec­tion inter­face and out­put to the TMC1761 dis­play shift reg­is­ter, TMC1764 dis­play anode cath­ode and indi­ca­tor out­put, the TMC1765 arith­metic chip, TMC1768 and TMC1793 pro­gram ROM’s for stor­ing plu­ral­i­ty of microprograms.

Earlier revi­sions of the 650 did not have the small daugh­ter board locat­ed above the key­pad con­nec­tor. This addi­tion­al seems to be relat­ed to the mem­o­ry func­tions of the cal­cu­la­tor and has a set of three wires attached to the key­pad diode matrix. The wires attach to the (=-1), (=-2), and the (=-) diode matrix pads. There are five Toshiba TM43xx log­ic IC’s on the board along with twen­ty two 22K Ohm pull-up resis­tors, two 56k Ohm resis­tors, eight diodes, and a dozen wires that con­nect to the main cir­cuit board.

Based on the elec­tron­ic com­po­nent date codes and man­u­fac­tur­ing stamps, this cal­cu­la­tor was built in April of 1972. Hopefully Google Translate did a good job on the city / region / process trans­la­tion. Any cor­rec­tions would be appreciated.

The Nixie tube dis­play assem­bly is con­nect­ed to the main board by an edge con­nec­tor cable, and to the pow­er sup­ply with three indi­vid­ual pow­er wires that plug-in to ter­mi­nals on the high volt­age reg­u­la­tor sec­tion.
The key­pad diode matrix rests on the black foam pad locat­ed on the front edge of the dis­play assem­bly. At first I was con­cerned that the mate­r­i­al may have dete­ri­o­rat­ed from age, but the foam is still very resilient and I per­formed sev­er­al high resis­tance checks on dif­fer­ent areas of the foam with no issues, so no need to replace it.

It was final­ly time to pow­er up the unit, so I attached sev­er­al of my mul­ti­me­ters to the dif­fer­ent DC out­puts of the pow­er sup­ply before turn­ing it on.
Most of the volt­ages were very close to where they need­ed to be with the excep­tion of the 215 volt VH‑1 ter­mi­nal which I could not adjust down to the cor­rect volt­age. The cul­prit end­ed up being the 130K Ohm resis­tor in the adjust­ment sec­tion of the HV reg­u­la­tor.
After replac­ing the resis­tor all the volt­ages were nom­i­nal, and I had a dis­play of 0 on the cal­cu­la­tor.
After a thor­ough check of all the keys, switch­es, and func­tions every­thing was work­ing great.
Here are some pow­er read­ings I mea­sured at 121 volts AC input:
16.7 Watts at 0.85 Power fac­tor — Full load 16 dig­its dis­play­ing 8’s and most indi­ca­tors on.
12.5 Watts at 0.81 Power fac­tor — Idle, only the least sig­nif­i­cant dig­it “0.” and no extra indi­ca­tors on.

I spent a good part of an after­noon repair­ing the dam­aged upper case. I used some J‑B Weld 50139 black plas­tic bon­der epoxy, which was a good match to the orig­i­nal epoxy used by Cannon. I cleaned up the excess epoxy that squeezed out on view­able sec­tions using 99.9% iso­propyl alco­hol.
After sit­ting overnight I used some of my Plexus plas­tic clean­er, pro­tec­tant, and pol­ish on all the out­side plas­tic parts of the case, includ­ing the keys, and dis­play window. 

A web­site that was a great help on the elec­tron­ics side of the restora­tion has been the Madrona Grove Website of Brent Hilpert. He has reverse engi­neered an ear­ly ver­sion of the Monroe 650 cal­cu­la­tor, and has pub­lished his schemat­ic for this cal­cu­la­tor, and many oth­ers also.

So now this is my favorite cal­cu­la­tor in my small col­lec­tion. It has tak­en over the posi­tion from the Burroughs C 3207, 14 dig­it Nixie cal­cu­la­tor.
What’s not to like? It has two more beau­ti­ful orange Nixie dig­its, an extra mem­o­ry, square root func­tion, float­ing point mode, an entry counter, and a back­space key. Now I just need to make a bit more room on my desk.