< Go to Part 4.

13.1 Installing the cooling unit.
It quickly became apparent that it was impossible to lift the cooling unit onto the cabinet by hand. Apart from the weight, the cabinet is 1300mm tall and the bottom of the evaporator has to be lifted over the top. Then the cooling unit has to be delicately aligned so that the seal and high side tube is not caught as the unit is lowered.
GE did make a portable hoist specifically for Monitor Top installation but as this was only used by installers, such an item would be impossible to obtain now.

For those that want to attempt copying it, this is the GE crane for the cooling units.

.

An engine hoist is an ideal substitute and is what I used. It is important that the hoist can be raised at least 2m.

The cooling unit was supported on milk crates which in turn were placed on the feet of the hoist. Once the hoist was raised, the crates were removed so the hoist could be pushed in front of the cabinet.
Lowering the top into the cabinet is a tricky procedure and requires a lot of care. As the high side tube runs right against the cabinet opening, it is easily damaged. The seal is easily caught on the cooling unit as it is lowered, so this must be done slowly checking frequently that everything is in the right position.
As my cabinet was slightly out of square, the top had to be lowered with the back lip on first and then the cabinet pushed into square as the front was lowered. Once on, the top lip keeps it square.

The restored CA-2-B15.

14.1 First operation in cabinet.
As always when the cooling unit has not been powered, it is necessary to run the heater first, to evaporate the Methyl Formate from the oil. It does not take long for the Methyl Formate to start boiling once the power applied. This is audible by listening close to the compressor. The boiling gradually gets louder, and after a few hours it stops. This means it is ready to be used.
It was found that ice cubes could be frozen in a plastic tray sitting in the evaporator, so it would appear the temperature was low enough.

14.2 Calibrating the thermostat.

Despite the cabinet temperature appearing to be sufficiently low, the compressor rarely switched off, even on the No.1 setting. Looking through the AK service manual (Canadian version of the CK), it appeared that the thermostat knob might need to be reindexed. The fact that I had removed the bellows from the rear of the thermostat could have upset calibration. Also mentioned was the effect of atmospheric pressure on the bellows, and that at higher altitude the calibration would be different. This was plausible given the elevation at my location.

 

My assumption was that given the evaporator temperature should be no more than 25 degrees F, then that should correspond with the warmest setting; i.e. 1.
Removing the knob requires the centre disc to be carefully prised out. I used the smallest size jeweller’s screwdriver I had. With the knob initially set to 1, I took it off and moved it to 6. Effectively, it had 6 positions more in the warmer direction. This seemed a good guess with the evaporator settling at around 25 degrees F. Cycling of the compressor was improved, but not optimum. This I assumed would be losses through the door seal.

15.1 Door seal.
I was able to purchase a reproduction seal from Don at Rusty Metal Rescue.
This has to be contact glued into position as it does not have the barbs that fit the holes around the door edge.

Unfortunately, upon reinstalling the door, I found the bottom left corner still had a gap. Further investigation revealed the cabinet was slightly warped in this area. It appears from the dent in the foot below this area that the cabinet has been dragged with all the force on this leg, thus applying stress to the bottom of the cabinet immediately above. Given the impracticality of getting the cabinet perfectly square, I simply added a thin piece of neoprene seal to fill the gap.
Reproduction foot pads were purchased to replace the originals.

16.1 Non condensable gases again.
Over the course of running the fridge, it had been noticed that after a few hours performance would drop off, with a low frost line and continuous running, along with warm cabinet temperature.
Yet, if the fridge was rested for a few days all would be reasonable again for another few hours.
I had developed a theory that maybe non condensable gases were present in the compressor walls and the motor and being driven out when warmed up.
Given that the motor windings would still be likely to contain moisture (the motor was not heated during evacuation) this seemed a likely scenario.
So, I purged the non condensables again; it took a couple of hours before the methyl formate appeared.

Purging the non condensables requires oil to be placed in the charge valve to ascertain positive or negative pressure.

Performance was now excellent. Within a few hours the duty cycle was perfect; 2-3 mins on and 10-13 mins off. On the No. 9 setting the evaporator got down to 3 degrees F which was more than sufficient, especially as the bottom of the cabinet was below freezing.
With this excellent performance still at 11pm, I decided to leave it on over night. Alas, next morning was a disappointment with the compressor running continuously, a low frost line, and ice melting off the evaporator.
My theory at this point was that not all the moisture had been driven out of the motor during the previous purge. Thus it would seem necessary to repeat purging until it is.

16.2 Float valve blockages.
A further purge was required, but performance was still poor. As non condensable gases were not the problem, it had to be a blockage. Tapping on the side of the float chamber’s needle and seat area brought forth a good flow. Several times the fridge had been run with a gradual decrease in performance after a day. A pattern was emerging, in that if the fridge had been rested for a day or more, it would run perfectly but always start to deteriorate after a day’s running. By this time, power consumption was under 200W (including transformer and oil heater), and the frost line would be very low.
With non condensables eliminated as the problem, another theory was looking plausible. Observing power consumption over the time the fridge ran correctly showed an initial draw of about 240W. As the fridge cooled and was cycling normally, consumption went down to about 200-210W.  It would appear that the density of methyl formate decreased as it cooled because of less strain on the compressor, which in turn relates to less power consumption. So, with less pressure, there is greater difficulty in clearing the blockages. This would explain why the blockages largely cleared themselves on initial power up, but not so after the fridge was down to normal working temperature.
Running the fridge for 10 days continuously showed up an interesting pattern. Blockages would usually occur at around 5am which is the coldest part of the day inside the house. Late at night, the power consumption would start to drop and the cycle time reduce. With the float seat starting to block, a higher than normal vacuum would occur in the evaporator which in turn means the compressor has less vapour to suck in. Thus, the compressor runs at less of a load. When this happens, the power consumption drops to about 185-190W. After this,  the frost level falls because the incoming liquid is not being replenished as fast as it is being evaporated. Short cycling then starts to occur. This is presumably because with less frost on the evaporator, it is less able to store the cold temperature. Eventually, the frost line drops so that it is no longer near the thermostat bulb, and the compressor runs continuously.

16.3. Compressor timer.
Not always being at home meant that in the event of the compressor running continuously, I would not be able to turn it off and/or clear the blockage. So, I built up a timing circuit that sensed when the compressor was running. If the compressor ran for more than 25 minutes (or any other desired time), power would shut off and remain so until manually reset. This would protect the motor from an excessive run time. Importantly, it meant I could leave the fridge running unattended and not worry about it. Operation is sensed by a current transformer. This actuates an Omron H3CR timer set for 25 minutes.
If the timer is not reset within that time, a relay latches cutting power to the fridge.The relay remains latched while ever the 120V supply is present.

16.4 Improvement in operation.
Running the fridge for several months did improve things considerably. The blockages became much less frequent. In fact, instead of hours apart, it was now days. It would appear that whatever was causing the valve to stick may have been gradually dissolved. Alternatively, the constant flow of methyl formate and valve operation may have “machined” the needle and seat for a less sticky fit. There is no difficulty in freezing ice in plastic trays and the cycling times are good; 2-3 minutes on and at least 10 minutes off. Correct temperature is with the control in position “3”. This suggests the calibration is satisfactory, although for correctness the knob should be moved to “5” before screwing back in.

17. Shelves.
Only two of the three shelves were present. They are made of steel and either galvanised or zinc plated. The finish is too dull to be nickel or chrome. A previous owner has painted over this with white enamel.

The shelves had been painted in white enamel over the original plating. The black objects on the lower shelf are the original foot pads.

I had the shelves replated by Artarmon Electroplating. It appears that they were originally zinc plated and this is what was recommended. Nickel or chrome was advised against due to rusting in the moist environment.

Like new again, the shelves were the final part of the restoration.
 

18. Acknowledgements.
This was my first entry into the world of refrigeration and would have found it considerably more difficult without the help of others. There is a lot of negativity about working on Monitor Tops around the internet, and when I found the Flickr group all that changed. In particular, due acknowledgements must be given to the following group members for their help:

• jhigdon2 - the guru of monitor top repair.
• worldsfair39, Coldspaces61, and Allan 1955 - enthusiastic members always keen to see how the restoration was progressing.
• and of course, litlnemo - for establishing the group in the first place.

19. Suppliers.

• Heatcraft - for the silver solder and copper fittings, etc.
• Sigma Aldrich - for the methyl formate.
• keep-n-time - ebay seller for the footpads.
• Don at Rusty Metal Rescue - for the door seal.
• Artarmon Electroplating - who did the zinc plating for the shelves.

Monitor Top Refrigerator Forum (replacement for Yahoo and Flickr Groups).

Yahoo group - GE Monitor Top Fridges. (no longer active due to Yahoo changes).

Flickr group - Monitor Top Refrigerators.

Postscript.
Unfortunately, the reliability of the float valve in this machine turned out to be very problematic. It would block up frequently. After the success of the CA-1, the obvious solution was to do a capillary conversion.
See the details here.
This improved reliability considerably, but evidently the system had been badly contaminated because NCG build up was still frequent, requiring purging every month. The decision was made to confirm this was not due to leakage, so the unit was evacuated and filled with 50psi of R22. A leak detector could not find any evidence of leakage. Much to my surprise, the R22 must have had some kind of cleaning effect, because after evacuation of the R22 and refilling with methyl formate again, performance was really good. No purging was required for seven months. Cycling times were good; 4.5 mins run and up to 16 mins off with a cabinet temperature of about 34F.
In January 2015, a year after doing the capillary tube conversion, a high side blockage occurred. The inline strainer was completely clean which was good news - no more corrosion evident. However, it indicates the problem of the blocked evaporator remains. Again, I am led to believe this cooling unit has been left open to the atmosphere for quite a long time.

A replacement evaporator was subsequently installed. A check with nitrogen flow showed that the original was still partially blocked. A couple of years of methyl formate flowing through it had not dissolved the blockage as I had hoped. Since the new evaporator was installed, performance has been completely reliable. This is now my main fridge.