|
|
Post by perryhahn on Sept 4, 2020 3:49:54 GMT
My refrigerator project is a DR1, so I have been having tunnel vision about high pressure domes. I wasn't even aware until recently that the CK units had a low side dome. It seems to make sense that they would not have as much dissolved refrigerant.
By my figuring, there is one shortcoming in all of the experimental data that I have seen for miscibility of various refrigerants in various oils. All of these measurements have been made at the saturation pressures of the refrigerant, probably because it is easier to set up the test equipment this way. However, with most of our refrigerants, we have superheated vapor in the dome under running conditions. Measuring under saturation conditions skews the data quite a bit, because pressure actually affects miscibility more than temperature does. And they generally don't even bother to give the pressures in the data logs or resulting graphs.
If we increase the temperature of a sample of oil with dissolved refrigerant, but keep the pressure constant, the miscibility goes down. However, if we take the same sample of oil and increase the temperature under conditions of saturated refrigerant, the miscibility goes up. This is because the pressure also goes up under these conditions. This is why we have to do some educated voodoo to guess the miscibility under conditions of refrigerant superheat. Does anyone else have a different take on this?
One of the best things I learned from David's videos was how the refrigerant port adapters work. I was unable to find one on E-bay, although I did learn that someone bought nine of them two years ago at an auction for a grand total of $50. Of course, this information doesn't help me any. Fortunately, David gave enough information that I was able to machine my own.
Perry Hahn
|
|
|
|
Post by turbokinetic on Sept 4, 2020 4:05:59 GMT
My refrigerator project is a DR1, so I have been having tunnel vision about high pressure domes. I wasn't even aware until recently that the CK units had a low side dome. It seems to make sense that they would not have as much dissolved refrigerant. By my figuring, there is one shortcoming in all of the experimental data that I have seen for miscibility of various refrigerants in various oils. All of these measurements have been made at the saturation pressures of the refrigerant, probably because it is easier to set up the test equipment this way. However, with most of our refrigerants, we have superheated vapor in the dome under running conditions. Measuring under saturation conditions skews the data quite a bit, because pressure actually affects miscibility more than temperature does. And they generally don't even bother to give the pressures in the data logs or resulting graphs. If we increase the temperature of a sample of oil with dissolved refrigerant, but keep the pressure constant, the miscibility goes down. However, if we take the same sample of oil and increase the temperature under conditions of saturated refrigerant, the miscibility goes up. This is because the pressure also goes up under these conditions. This is why we have to do some educated voodoo to guess the miscibility under conditions of refrigerant superheat. Does anyone else have a different take on this? One of the best things I learned from David's videos was how the refrigerant port adapters work. I was unable to find one on E-bay, although I did learn that someone bought nine of them two years ago at an auction for a grand total of $50. Of course, this information doesn't help me any. Fortunately, David gave enough information that I was able to machine my own. Perry Hahn Hi Perry. As for the different dome pressure designs; All of the antiques from back in the early days were designed in the Wild, Wild West so to speak! There is so little common design, and so many ways of doing different things it's truly amazing. One reason for the high-side dome was for gas density to cool the motor. Another was so the pressure differential could be used to circulate oil. Those were the early designs. One manufacturer had a split-phase start motor inside a hermetic dome. It had a conventional mechanical centrifugal start switch... inside the dome. The contacts were a mercury capsule which would tilt when the centrifugal mechanism actuated. No start relay at all. Centrifugal switch in a hermetic unit. Yeah; the Wild, Wild West. Then there's the Norge Rollator, with its cast iron rotary compressor, belt driven. Much hype around its merits by Norge. It was a high-side dome design with a mechanical shaft seal, exposed to high-side pressure. As you can imagine, some of the failures were epic, considering it used SO2! This sort of things are why I am so fascinated with the old fridges! You never stop learning and it's fascinating to bring back to life these old designs and see something that hasn't been used in many decades, and was a cast-off design, come back to life. As for oil miscibility; it's not important for the compressor that the refrigerant and oil are 100% fully miscible, or fully immiscible. The vast majority of these designs have a lot of splash and agitation in the oil, so some modest amount of undissolved refrigerant sitting in the bottom of the sump won't matter. What I have seen firsthand works horribly is a fully miscible combination in a low-side float Frigidaire. If the oil and refrigerant don't separate in the evaporator, it logs all the oil in that area. I had one with R12 and mineral oil come in last year and 99% of the oil was in the evaporator. Sincerely, David |
|
|
|
Post by perryhahn on Sept 5, 2020 6:44:08 GMT
Nevertheless, I would rather not ignore the effect of dissolved refrigerants on oil viscosity. The engineers at GE certainly didn't. The introduction of r12 required big changes in both refrigerator design and oil viscosity. R12 is highly miscible with mineral oil, and this would reduce the viscosity of the oil tremendously, especially in a high pressure dome. In their ads, Texaco recommended an oil for GE r12 refrigerators that had twice the viscosity of the oil that they recommended for GE SO2 refrigerators. This was in spite of the fact that their r12 refrigerators had a low side dome, and wouldn't get near as much dissolved refrigerant in the oil as an r12 refrigerator with a high side dome. (Am I correct on this?) Aaron showed me a useful post on this topic at monitortop.freeforums.net/thread/886/general-electric-refrigeration-oil which you have probably already seen. Which gets to one of the big advantages of r152a, your favorite replacement refrigerant for SO2. The viscosity of the oil would be higher using r152a than it would be using SO2, partly because of lower dome temperatures, but mostly because there is less dissolved refrigerant. Given the total amount of compressor parts wear on a 90 year old machine, this is probably exactly what we want. If it wasn't for the float switch issues, this might be a shoe-in for a replacement refrigerant. In your own refrigerators, you have solved the problem with TLC, but probably without much margin for error. Invasive surgery could solve the float issue, perhaps by fooling with the spring. Naturally, I'd rather not do this. Maybe some carefully arranged magnets outside of the float? This might make the float pull strongly to one side and get stuck. Just thinking out loud. Do you get a louder compressor noise by using r152a due to the increased compression? Perry Hahn |
|
|
|
Post by turbokinetic on Sept 5, 2020 9:13:03 GMT
Nevertheless, I would rather not ignore the effect of dissolved refrigerants on oil viscosity. The engineers at GE certainly didn't. The introduction of r12 required big changes in both refrigerator design and oil viscosity. R12 is highly miscible with mineral oil, and this would reduce the viscosity of the oil tremendously, especially in a high pressure dome. In their ads, Texaco recommended an oil for GE r12 refrigerators that had twice the viscosity of the oil that they recommended for GE SO2 refrigerators. This was in spite of the fact that their r12 refrigerators had a low side dome, and wouldn't get near as much dissolved refrigerant in the oil as an r12 refrigerator with a high side dome. (Am I correct on this?) Aaron showed me a useful post on this topic at monitortop.freeforums.net/thread/886/general-electric-refrigeration-oil which you have probably already seen. Which gets to one of the big advantages of r152a, your favorite replacement refrigerant for SO2. The viscosity of the oil would be higher using r152a than it would be using SO2, partly because of lower dome temperatures, but mostly because there is less dissolved refrigerant. Given the total amount of compressor parts wear on a 90 year old machine, this is probably exactly what we want. If it wasn't for the float switch issues, this might be a shoe-in for a replacement refrigerant. In your own refrigerators, you have solved the problem with TLC, but probably without much margin for error. Invasive surgery could solve the float issue, perhaps by fooling with the spring. Naturally, I'd rather not do this. Maybe some carefully arranged magnets outside of the float? This might make the float pull strongly to one side and get stuck. Just thinking out loud. Do you get a louder compressor noise by using r152a due to the increased compression? Perry Hahn As far as the sound of the machine, SO2 versus R152A; I haven't had the opportunity to directly and immediately compare the sound of a properly functioning SO2 unit, which had its refrigerant replaced with R152A. I just don't think it's OK to alter a perfectly working unit. All the ones I've recharged were either already empty, or they were not working properly for some reason. I can say that the DR1, running with a nonexistent SO2 charge was rattling, but with a proper R152A charge it seems in line with Travis's original DR1 that I have heard in person. The DR3 ran with SO2 at my shop, then broke on the road due to rough highway conditions. About a year later it came back and was repaired and recharged with R152A. It is noticeably quieter than I remember it before - but the weather was colder initially, and the run with SO2 was an initial wake-up after sitting for decades. There was some rattling at that time. Once the R152A charge was dialed in it sounds very nice. A lot of time passed between hearing it with SO2 then with R152A; so I don't trust my memory to say how much it changed. Oh and when I said that the compressors don't care much about oil miscibility, that was meant to be a follow-on to the newer low-side dome systems, with cap tube and heat exchanger. As said those have a very hot oil and low pressure environment, so that dissolved refrigerant is minimal. Not at all applicable to a DR or anything with high side dome. Was not clear about that. If one were afraid of the float situation; you could replace that with a capillary tube. Or cut the chamber open and install a spring, as you mentioned. I would tend to just be extra careful when recharging and be sure that air doesn't remain. If it's worked for a few days, I would see no reason it would stop working after that due to the float. There is either air in the system or not. It takes a certain time to collect in the float chamber, but once it's gone it's gone. For the low-side dome units such as the GE Spacemaker, Westinghouse, and Tecumseh; there is absolutely no change in sound versus R12. |
|
|
|
Post by ChrisJ on Sept 5, 2020 23:24:25 GMT
I personally don't like the cap tube setup I have.
The big problem is I have absolutely nothing to compare it to so my dislike could be 100% wrong.
But to me, in my head a properly working float valve CA should be more efficient and stronger under a wider range of conditions than my cap tube CA. It also should flush oil out of the evaporator better from the bursts of refrigerant vs the slow steady flow of the cap tube.
|
|
|
|
Post by cablehack on Sept 6, 2020 1:23:01 GMT
I personally don't like the cap tube setup I have. The big problem is I have absolutely nothing to compare it to so my dislike could be 100% wrong. But to me, in my head a properly working float valve CA should be more efficient and stronger under a wider range of conditions than my cap tube CA. It also should flush oil out of the evaporator better from the bursts of refrigerant vs the slow steady flow of the cap tube. I did have something to compare the cap tube with - the float valve in my CA-2. The cycling times and evaporator temperature which it provided were within specs. In comparison, the cap tube performance, after getting it right, has been equal to or slightly better than the float valve. The 'flushing of oil out of the evaporator' theory is an interesting one. My CA-1 has never suffered from an oil logged evaporator, and it's been 3 years since the CA-2 had that happen. |
|
|
|
Post by turbokinetic on Sept 6, 2020 2:32:51 GMT
I personally don't like the cap tube setup I have. The big problem is I have absolutely nothing to compare it to so my dislike could be 100% wrong. But to me, in my head a properly working float valve CA should be more efficient and stronger under a wider range of conditions than my cap tube CA. It also should flush oil out of the evaporator better from the bursts of refrigerant vs the slow steady flow of the cap tube. I did have something to compare the cap tube with - the float valve in my CA-2. The cycling times and evaporator temperature which it provided were within specs. In comparison, the cap tube performance, after getting it right, has been equal to or slightly better than the float valve. The 'flushing of oil out of the evaporator' theory is an interesting one. My CA-1 has never suffered from an oil logged evaporator, and it's been 3 years since the CA-2 had that happen. Don't forget that the CA compressor will send out very little oil to the evaporator due to the high side dome design. Also, the outlet from the inside of the dome is in the center, above a rotating fan blade. This will centrifugally expel droplets from the center, towards the outside of the dome, with the least dense area being at the center of the dome above the fan, inside the spinning gas vortex. So very little oil is expelled with the vapor. On the low-side dome machines such as the CK; a lot of oil is pumped out because of the oil sealed design of the compressor. The oil ends up in the evaporator header above the level of the liquid refrigerant. The return line is located in the upper part of the header, so it "skims off" the oil layer as it rises up to the level of the line opening. The compressor has a large enough sump capacity that it is unharmed even when a significant volume of oil is trapped in the evaporator, or suddenly returned from there. When looking at the frost line on the evaporator, most of the area above the frost line is filled with oil, or an oil foam. Things go wrong when the oil starts to change chemically, and the pour point temperature rises up, so it congeals. I have a theory that this is why some units start getting oil logged over and over; while others never have this happen. I bet if you were to drain and replace the oil from one which has a chronic oil logging issue, the issue would not return. I don't think we know the pour point of GE's original oil. I do happen to have the data sheet for the AB 150 oil used in my Frigidaire projects. It has a pour point of -49°F. That is well below the capability of the Monitor Top fridges. Assuming GE's oil had a similar pour point, that means even with a fully frosted evaporator, the oil layer floating on top of the SO2 would still be flowable and not congealed. You can see how, if something chemically were to happen to this oil and make it congeal easier; things could go wrong. |
|
|
|
Post by coldspaces on Sept 6, 2020 13:51:31 GMT
I did have mixed results with 152a. It seemed to lay in in the cold oil and not circulate right away. The main reason I use R124 is that it is the only class A-1 non flammable refrigerant I have found that lets me replace so2. I do refrigeration for a living and don't just don't want the liability of a flammable refrigerant I am not insured for.
So far the R124 has worked good. Only time will tell us for sure.
|
|
|
|
Post by birkie on Sept 6, 2020 17:03:21 GMT
Hi all, I've been a bit distracted at home and haven't been able to jump in, but this is some great discussion! The DR machines are particularly interesting, since they have a unique combination of high-side dome, pressurized oil system via piston, sliding/oscillating cylinder block with side valves, and an oil pressure-based unloader valve. Conventional wisdom just doesn't apply in the usual ways, which makes this a fascinating topic. If we increase the temperature of a sample of oil with dissolved refrigerant, but keep the pressure constant, the miscibility goes down. However, if we take the same sample of oil and increase the temperature under conditions of saturated refrigerant, the miscibility goes up. This is because the pressure also goes up under these conditions. This is why we have to do some educated voodoo to guess the miscibility under conditions of refrigerant superheat. Does anyone else have a different take on this? Yes, educated guess voodoo absent direct measurement. The thing is, nobody is measuring oil and refrigerant mixtures under the conditions encountered in the domes of DR units; what happens in the evaporator and suction lines seems to be the sole area of interest in the modern era. That's all the industry seems to care about, and that's all you can find. The sump heater changes the balance as well. This article is an interesting (but not directly applicable) read: www.rses.org/assets/serviceapplicationmanual/620-1071.pdfGenerally speaking, there is more refrigerant in the oil in a DR sump when the refrigerant and oil are miscible under those conditions, and the oil subsequently will have a lower viscosity. Concluding "good" or "bad" is nuanced, and there are open questions. For example, we all know about the famous DR rattle. It has been hypothesized to be the result of flashing refrigerant to vapor in the machine bearings. For every refrigerant, higher heater wattages strictly leads to smoother running, up to a point. Does the absence of rattle mean that the concentration of refrigerant in the sump is below a critical value (and therefore the "point of no rattle" has similar refrigerant concentrations, across refrigerants)? Likewise, here is a fascinating article from 2018 about the lubrication properties of oil and refrigerant combinations. It has been known for some time that CFCs and HCFCs result in a chlorinated layer on ferrous metals that provide a protective, lubricating barrier. The article mentions a study that found that an HFC refrigerant mixture requires 50% more lubricant viscosity to achieve similar wear to a CFC or HCFC. So compared to an HFC like R134a or R152a, R124 has a protective advantage due to this phenomenon. From this, I cannot conclude that R152a is strictly better due to a probably higher in-sump viscosity. I also cannot conclude that R124 is strictly better either. I think they both work well, but for slightly different reasons from a theoretical perspective. Absent hard numbers for the actual conditions in the sump, the best we can do is observe over time! (As an interesting aside, the same article mentions some protective effect for HFO refrigerants. Just as CFCs and HCFCs tend to relinquish their chlorine atom, HFOs will relinquish a fluorine atom to form a fluorinated layer on ferrous surfaces. It's not as strong an effect as CFCs or HCFCs, but it's definitely significant). Ranking refrigerants by miscibility in mineral oil, from least to most, we get r132a, r152a, SO2, r124, and finally r1234ze If I'm reading it right, I think the R1234ze paper you linked too doesn't say anything about pure R1234ze in mineral oil; all the results seem to have some fraction of dimethyl ether (RE170) mixed in, which would increase miscibility, unless I'm reading it wrong. Absent hard numbers otherwise, based on how it is described verbally in the literature, it sounds like it would be somewhere between R152a and R124. I do plan to try R1234ZE in a monitor top. I'm curious where you were able to get detailed info about R1234ZE oil miscibility. Honeywell does say it is "more soluble" in mineral oil than HFC and gases, but they don't say by how much. I may have said already but the R1234ZE is working well in at least 3 Frigidaire hermetic units, which have high side dome configuration. Yes, I'd love to see that. In fact, I'd love to see R124, R152a, and R1234ze in the same machine to make comparisons. Give it a year or two and I might get there. When life gets in the way, I live vicariously through what others are able to do  |
|
|
|
Post by ChrisJ on Sept 7, 2020 1:40:42 GMT
Ok I may come across as clueless with this one but, I assume a stock DR has mineral oil? I believe Nujol brand to be specific?
If that's the case R134 isn't miscible with it, is it?
What about 152?
|
|
|
|
Post by turbokinetic on Sept 7, 2020 5:38:03 GMT
I did have mixed results with 152a. It seemed to lay in in the cold oil and not circulate right away. The main reason I use R124 is that it is the only class A-1 non flammable refrigerant I have found that lets me replace so2. I do refrigeration for a living and don't just don't want the liability of a flammable refrigerant I am not insured for. So far the R124 has worked good. Only time will tell us for sure. Time will tell, with all of these. I was hesitant to try R12134ZE at first and didn't try it in someone else's fridge until I had one running with it for more than a couple months. It's working well but still I say it's not tried and true with only 3 or 4 units running. I did charge one unit which had the compressor opened once already, so we have pictures of it. If it dies, we can compare any possible poor lubrication damage or corrosion issues. As for R152A I completely respect your concern with it's flammability. I've used it for about 8 years now across a variety of projects. Never had any compressor failures while running it. Two of the Frigidaire Meter-Miser units have had quite a lot of run hours on them. One of my earliest uses was an already noisy compressor in a used car. I ran it in the southern summer heat for about 7 years and then did a teardown. It really was suffering no abnormal wear, the oil was clean, and had it not been for a faulty machining in the remanufactured compressor causing noise, it would still be in operation today. I did get a recovery unit from a flea market a while back, and dedicated it to this refrigerant. I'm using my DR1 as a shop fridge to keep paint related materials cool. I know, the horror! But it was given to me, missing shelves, not running, and given up for dead. It is better used this way than in a scrap yard. It is getting used in a hot environment as a stress test for the R152A conversion and so far has been doing very well. I know some think it's irresponsible to use an 80 year old unit in the heat; but in the interest of learning what works and what doesn't I can justify it. Time will tell; and I'm probably accelerating the effects of time a little bit LOL! Hi all, I've been a bit distracted at home and haven't been able to jump in, but this is some great discussion! The DR machines are particularly interesting, since they have a unique combination of high-side dome, pressurized oil system via piston, sliding/oscillating cylinder block with side valves, and an oil pressure-based unloader valve. Conventional wisdom just doesn't apply in the usual ways, which makes this a fascinating topic. If we increase the temperature of a sample of oil with dissolved refrigerant, but keep the pressure constant, the miscibility goes down. However, if we take the same sample of oil and increase the temperature under conditions of saturated refrigerant, the miscibility goes up. This is because the pressure also goes up under these conditions. This is why we have to do some educated voodoo to guess the miscibility under conditions of refrigerant superheat. Does anyone else have a different take on this? I think we can expect the Monitor Top units do well with a wider range of refrigerants than more modern systems operating at higher stress levels. This feeling is based on the degree to which they are overbuilt. As Aaron said, it's all conjecture what is happening inside the dome when they are running because we can't see inside. Those of us who have first-hand experience with these; and have heard many of them run correctly; can usually detect subtle differences in the way they sound. Also we can use power demand, temperatures, and cycle times to "estimate" how much stress the machine is under. Acoustic emissions are a very good indicator of the peak stresses applied to the moving parts. This is because clicking, rattling, and other sharp noises are caused by high peak amplitude vibrations. These are created by metal parts hitting one another and therefore rapidly transferring momentum from one part to the other, and ultimately to the machine casing. Any operating conditions which result in this type of acoustic emission are increasing stress on parts. To the breaking point? probably not in the short term; but longterm damage could accrue. For instance, if a DR lubrication pump is taking in refrigerant liquid which flashes to gas in the lube system, and disrupts the oil film, there is excess clearance and the parts are allowed to "hammer" without an oil film to cushion them. The parts are so massive that they seem not to be destroyed in short order by this; but I don't expect GE would have considered this OK for their new product at the time. Thanks for the article link to the Tribology in Industry article! It's a long read; but the key take-away I get from it is that the actual lubricative properties of the refrigerant were most significant when seizure loads, and dry operating conditions were simulated in the lab. These are boundary layer situations pushed to failure. When dealing with a real-world compressor; that would only be a factor if the lubricating oil film breaks down, leading to a boundary layer scenario. Thankfully GE enigneered these compressors with full pressure lubrication so that boundary layer lubrication is not depended on. Keeping the oil sump as free of liquid refrigerant as possible will help to ensure this is the case. I do have plans to experiment with oil miscibility an a "real" compressor with a viewing window. I got an Embraco sample a while back with a large sight glass window in it. The plan is to install an oil drain port in the bottom center of the housing, so that I can easily remove and replace the oil without having to take the whole system apart. I want to use it in a flooded evaporator system, such as the CK has. The experiment will have two goals. One; to see how much variability the compressor oil level has, as operating conditions change with the evaporator. And also, to see how it looks with different refrigerants such as R12, R152A, R1234ZE and the like. Seeing how it's a low-side housing unit, it won't help with the questions about the DR machines, but may be interesting none the less. Gosh I have too many project plans and not enough hours in the day!!! Ok I may come across as clueless with this one but, I assume a stock DR has mineral oil? I believe Nujol brand to be specific? If that's the case R134 isn't miscible with it, is it? What about 152? The HFC refrigerants such as R134A and R152A are only slightly soluble in mineral oil. I looked up Nujul oil and got some interesting search results! Seems it was used as a medical lubricant and laxative supplement. http://aconstantineblacklist.blogspot.com/2009/08/drug-story-by-morris.htmlBTW I really appreciate the good dialogue here. This is what makes it worth spending time on our forum! Sincerely, David |
|
|
|
Post by perryhahn on Sept 14, 2020 6:03:35 GMT
There are parts on a machine that don't achieve hydrodynamic lubrication, such as wrist pins, or a piston ring at top dead center. Any time a sliding part stops sliding, such as during an intermittent or oscillating motion, there is nothing to prevent metal on metal contact. Cylinder walls get worn mostly at the top and bottom of the stroke, due entirely to direct contact with piston rings. I don't think that the paper that Aaron quotes does any study of the wear on journal bearings in continuous motion, it being assumed that these bearings are immune from direct contact, and wear, under normal operating conditions
The DR's and some of the other refrigerators seem to have been designed to achieve as much hydrodynamic lubrication as possible. For instance, while the pistons do have an oscillating motion, there are no piston rings. The area of oil film between the piston and cylinder wall is so big, and the transverse forces are so low, that we probably don't have much metal to metal contact there. Similarly, while there isn't really a wrist pin in the piston, there is an oscillating journal bearing that takes more or less the same function. Once again, it has a much bigger bearing surface area than if we had used a wrist pin instead, and is easier to get high pressure oil into the joint. All of this is why we have 90 years and going.
Perry Hahn
|
|
