Anotherpower.com Forum
Project Journals => User Journals => Steve => Topic started by: MadScientist267 on April 18, 2014, 12:47:08 am
-
Greetings Earthlings ;D
As some may know there's a truck in the works... Well, on paper at least. And if all goes to plan, this will be one of the more significant projects in building it.
I have a 5 cu ft chest freezer, and I'm looking into converting it into a refrigerator, complete with a water based thermal mass, fan controls, and a sensor I'm working on developing that directly detects frozen/liquid states of the water in the mass.
Below are the conceptual drawings for the conversion itself, showing my initial thoughts on the location the of baffles, water bottles, and fans. The sensor is still very much in the "hmm I wonder" stage.
There will be a continuously variable fan supply based on the return air temperature from the food compartment, and a door switch that will kill those fans when it is opened. The freeze loop fans will run at full throttle with the compressor, the set being controlled on a supply vs demand basis as reported by the ice sensor.
Any ideas on the location and number of fans (yes there are a total of 6 in the drawings) would be appreciated, as well as any other glaring issues that I might be missing. I'd like to reduce the number of fans obviously, but they'll be relatively small (essentially 80mm computer case fans) and cold air is heavy.
The front view, highlighting the food cooling loop:
[attachimg=1]
The top view, showing the thermal mass freeze loop:
[attachimg=2]
... and a side view for whatever clarification it might provide:
[attachimg=3]
The bottles in this version are roughly of the 2 liter variety, and would be filled between 2/3 and 3/4 full. The block separating the thermal mass and the food compartment is 2" styro. Or at least that's the idea. :)
I am not opposed to using a larger quantity of smaller bottles, but wonder about the differences it might make in the proportions of the entire box. Which, as of the moment, I don't have any exact dimensions on. The aspect is approximate, so until that's known specifically, I can't lock down a definite design.
Steve
-
Interesting---approx. 50 litres of water can absorb a lot of heat when melting. My gut feeling is that
you won't need as many fans, but many unknowns will determine that.
As an aside, have you picked out a truck yet? Always suspected you might need more room. :)
-
Yep, looking at either an old 17 or 20 ft uhaul... It's a balance mostly between space and heat/cool load, with cost being a really close third, so in the end, availability and cash will decide between those two. Mostly what I want space-wise is vertically and the real estate on the roof anyway, so once those are met, anything else is just icing until it becomes too pricey to feed the wheels or keep it warm.
Hoping you're right on the fans... It does seem like a bit much, but roughly looking at flow, steering it thru the bottles for an even flow might be an issue. In the drawings, I tried to use long narrow gaps as much as possible to even out the flow, but with the freeze loop passages being so narrow, I wondered about drag in addition to density. How well theory tracks reality, as you said, still needs to be determined.
Steve
-
Don't know how things work there, but here, school busses can only be used for 10 years. I have seen/bought/used
a couple in that size that I got for the $1000 mark. Fair rubber etc, needed minor front steering work.
Not as good on vertical space, and would need insulation to make it practical though.
Will be reading with interest to see how it unfolds.
-
Update time...
[attachimg=1]
There's been a lot of progress with this, and after several obnoxious challenges and rather strange observations, I think I'm finally getting really close. Let's hope I don't jinx it :-\
Anyhow, one thing about working with things that are supposed to hold the status quo, is that it takes forever for little tweaks and changes to take effect and reveal whatever it is they're going to do. This is no exception... with a "time constant" in the few days range, patience is a virtue, but something I'm short on. Let's just say it's been frustrating, but in the end, it should bear it's fruit.
The ice storage idea is working, but it's difficult with the limited space inside. The previous drawings were a little on the optimistic side in the size department, but I managed to cram most of what needs to be present into the space available. The most notable difference between the conceptual drawings and what came to actually be is in the amount of storage mass. There's not nearly as much of it as originally envisioned.
The Thermal Mass
The thermal mass is a hybrid mix of brine and fresh water. This took probably the longest to determine the best ratio and configuration. There's a total of 19L of water, just over 15L of it being brine. In total, there are 11 bottles of water. Seven 2 liter bottles of brine, two 1 liter bottles of brine, and two 2 liter bottles of fresh.
[attachimg=2]
Determining the performance of the brine vs fresh bottles, after 1 hour elapsed. As you can see, the brine bottle (left) is clearly much colder than the fresh. But this comes at a price...
[attachimg=3]
... The brine doesn't last nearly as long as the fresh. This is at hour 6, and the brine is nearly completely melted, with the fresh nearly still completely frozen.
[attachimg=4]
The 4L of fresh water is being used as a sort of "guard" for the colder but lower capacity brine. They are situated at the return so as to pre-chill as much air as possible before it hits the brine bottles. The brine was calculated and mixed to be at -5C (23F), tho there is some potential discrepancy in the actual solid mass temp. I attribute this to likely being caused by an increasing salinity as the ice forms, causing fractional distillation which leaves the remaining water more concentrated with salt. This may become clearer as I experiment more with the placement of various temperature probes to see exactly what's going on. It's not high on the priority list to determine this, as the temperature range the bottles provide is more or less right where it needs to be, but the curious cat in me want's to know exactly what's happening there in the event I do need to make an adjustment.
[attachimg=5]
Above is one of the 2L brine bottles in it's present state after about 36 hours or so of stable operation. As you can see, it's 95% or so frozen. This is about as good as I could ask for given everything that has to be considered.
Temperature Monitoring
For development and data collection, there are 5 temperature sensors inside and 1 out, if you don't include those that are just inherently present and picking up ambient (not really part of the design, so not worth mentioning other than that). Some are not strictly needed anymore, but simply haven't been removed yet.
[attachimg=6]
Above are the off the shelf cheapie thermometers I used to reel everything in.
The sensor list is as follows:
1. Ice chamber air temp (for computer control).
2. Ice chamber air temp (for observation before the computer control was put in place and to calibrate against)
3. Fridge chamber "mass" temp. This is the most recently experimented with probe. It was changed from reading free air temp to being placed on the side of a small bottle of water so as to pick up the effective temp of an "average" item in the fridge. This is the primary control for the compressor.
4. Fridge chamber air temp for blower control. Provides the feedback for the buck converter driving the cooling blower.
5. Fridge chamber air temp for observation. This is used for development to confirm that under a given circumstance the air temp is on one side or the other of the mass temp. It essentially tells the temp of the blower control probe.
6. Compressor temp. Controls the compressor cooling fan.
Air Circulation
There are now 3 fans:
1. Cooling blower - Small brushless squirrel cage blower that pulls the air between the two chambers. This is not controlled by the computer, but is observed by it, the info being used to help determine if the compressor needs to start. It is continuously variable between 0 and full speed, although "full speed" is limited by a resistor at the moment. Turns out there's a such thing as this fan running too fast. Eventually, the speed limit will be imposed directly by the buck converter driving it, but for now, the resistor is a temp hack that just basically stuck.
2. Circulation fan - Brushless PC case type fan that keeps air moving within the refrigeration chamber. There's no speed control on this, it's on or it's not. While it's currently controlled by the computer, I'm finding that it's probably best that it simply run 24/7.
3. Compressor fan - AC induction fan to pull heat away from the compressor (and likely the condenser coils eventually as well). Due to where the compressor is located, this was needed to reduce heat ingress into the ice chamber. Once a structure is built around the unit to duct the cooling of the condenser coils, this will probably be changed to a brushless DC variety and possibly even set up so the air flow can be reversed depending on the season. That part is still on the table. For now, it's an induction motor because I had a suitable one around and it was trivial to control it. It is essentially on a simple bang-bang type thermostatic control, dependent on the compressor temp.
The Fridge Compartment
This is made of 2" styro, leftover from the insulation of the van. It's held together with silicone caulk, and was strategically carved to help optimize some of the air flow. How close I got to an actual ideal layout would involve more math than I know or even really have time to figure out, but much of it just fell into the logical reason realm. I had to commit with something to continue experimenting.
[attachimg=7]
The entire fridge structure is supported almost entirely by the ice bottles underneath, with the exception of where the step is formed within the original freezer where the compressor resides. There, I used an additional piece of foam placed on top of the step (it was determined that this was the only surface inside that didn't have evaporator coils attached to it). This actually accomplishes two things. It reinforces the insulation that's already present to help keep stray compressor heat out, and also forms part of the return duct for the chamber. On top of this on either side are small pieces of foam that were slimmed down so as to form supports for that side of the chamber, and define the edges of the return duct.
[attachimg=8]
The blower is recessed into the foam on one side, with the intake directly facing outward toward the left inside wall of the freezer. The discharge is sent up through two channels that attempted to split the air flow before it goes into the chamber. This wasn't as successful as I envisioned, although doesn't seem to be having any serious negative effect, so I haven't really looked into doing anything to modify it further. The outside edges are sealed with gorilla tape and form the 4th wall of the duct, up and over, into the chamber.
An intake downtube was formed for the development using just a couple of oatmeal cream pie boxes. They were handy, and easy enough to shape quickly into something that would work. Without it, the air that the blower pulls into the chamber simply isn't cool enough when running without the compressor. It is visible in the core shot at the beginning of this post.
[attachimg=9]
The return side was formed by extending the front and back walls all the way to the right most freezer wall, forming a gap between the right chamber wall and the freezer. These surfaces are rounded somewhat to help reduce restrictions (part of the optimization previously mentioned). The return air from the chamber flows over the right side wall, down along the edge, and underneath into the duct formed by the supports, and into the freezer where it meets the fresh water bottles first.
(continued below . . .)
-
(. . . continued from above)
[attachimg=1]
Here is the box loaded with a 2L bottle of soda. In this pic, from left to right, top to bottom:
1. The heater (since removed, see below)
2. Modest thermal mass bottle (fresh water) to "anchor" the compressor control thermister
3. Bottle of soda as a load
4. Week old egg (and counting) as a subjective performance measure (eventually it will be cracked open and the aroma "gauged" ;D)
5. The fan speed control and cheapie thermometer sensors
6. The internal chamber recirculation fan
Computer Control
The computer being used for development is sadly using right around 5 times the power that the actual cooling system uses. Eventually, the entire operation will be shrunk into a single PICAXE chip, and placed on the back of the freezer. For now however, it's a frankensteinish conglomeration of sensors, wires, boards, power supply bricks (yes, more than one LOL) and a laptop running some rather clumsy software that runs the show. What I learn here will be converted into the PICAXE program for the final product.
[attachimg=2]
Because it's temporary, I won't bother really describing much of the specific ins and outs of this particular part as it sits. In the pics however you can see that it is very prototypical and delicately (if not precariously) put in place to get things done.
[attachimg=3]
Power Factor Correction
This was the simplest of all to implement, but involved a little bit of hairy math to figure out. It works beautifully, as I started with a PF of ~0.84 and ended at ~0.98 and it only requires a single component to make it come to be; a capacitor.
[attachimg=4]
I'll spare you the math, but in the end, the cap that it called for with this particular compressor was calculated at 12uF. I had a 10uF, so I used that. The current is still slightly lagging, but the peak rides along the top of a sine wave, so it didn't make that much difference. Within a couple hours from having the number, I had it connected and running. I couldn't have been happier with the results, as it will mean a significant savings when the compressor needs to run from battery.
[attachimg=5]
The draw varies slightly depending on compartment temperatures between 95W (at coldest) to about 115W (all at room temp from a "black start").
[attachimg=6]
Significant Problems Encountered
Three of the more mentionable issues during this endeavor:
1. Establishing/Maintaining the temperature difference between the two chambers.
Mostly because of space, keeping the two chambers thermally isolated has proven to be a problem. I've found a relatively decent workaround I believe at this point, but there were several attempts at coming up with an end result that were less than ideal at best, and downright useless at worst. With the blower long stopped at a much higher temperature, there was no way of actively preventing the air flow. Natural convection eventually takes over, and the cold air creeps into the chamber regardless. I tried several different things, it seemed nothing was working.
[attachimg=7]
One involved actually adding small amounts of controlled heat to the fridge chamber. While the power range was minimal (typically between 5 and 10W), these added up significantly over time, caused wild variations in the air temps while the blower/heat came to an equilibrium whenever it was cycled, and of course increased the workload on the compressor. Whatever was put in was roughly doubled once the compressor got involved.
The increased power consumption wasn't originally viewed as an issue, as it was supposed to be power that the charging system would have otherwise been throwing away anyway. The problem became that it was determined that at times, it would even be necessary to run it at night when the battery was the only thing available. At a cost of just shy of 400Wh/day, I desperately sought to eliminate the need to do this. Trial after trial led eventually to the third fan that keeps circulating air in the chamber, along with other tweaks and modifications to the air flows. The experimentation was (and kind of still is) rather extensive, but I think it's under control at this point. More on this later as I confirm my findings.
2. The buck converter from hell
Quite possibly the most annoying part of this entire thing was the buck converter that controls the blower intermittently and randomly flipping out. This took a long time to narrow it down to the actual control chip itself. I suspected all kinds of things along the way, and without being able to reproduce any behavior, I had to simply wait until it did it again and hope that I caught it in the act.
Bad sensor, condensation in splices, failing trim pots, noise from the supply, flux on the board? All of these things came to mind and were dealt with one by one until finally, it began to have a predictable behavior. The random jumping eventually stopped after I cleaned the board, but it still was malfunctioning. It would then "go to sleep" and not wake back up when called upon after sitting for a while at 0 throttle.
It came to be that the only predictable behavior involved flexing the board. So, with nothing to lose and no real danger of doing serious damage to anything, I realized the actual problem quite by accident one night when I got really irritated at it and went between the pins with a tweaker (yes, with the power on). If there was something to get that was left behind, I was gonna find it. By accident, I briefly shorted the switching transistor pins, and found that it would wake up and run properly until the next time it was to shut back down for a bit. A few kicks in it's head later, it was clear: This was repeatable. In the pic below you can see where I had carved canyons in the board between the pins trying to figure out what in the world was going on... The transistor is between the top two pins between the diode and the red wire toward the right hand side. ::)
[attachimg=8]
Finally, after weeks of frustration, I had an answer. The chip was bad. I guess flexing the board stressed something in the chip just enough to modify it's behavior, and shorting the pins jolted it enough to maybe heat something to the point it made good contact within and it would continue until the down time let it cool off. Either way, I dug another one out and put it into place, all has been well since. What a PITA.
3. The ice sensor
I was intent on having a device in place that would sense the amount of ice present and adjust power use accordingly, but while I did end up with a functioning device, it was extremely difficult to get it to behave exactly as envisioned. It was capacitive, and senses the ice directly. As it turns out, it will be better to "push" power when it's available, to the extent that the temperature of the fridge compartment will allow it, rather than determine the needs and pull it. It was a novel approach, but in the end, would have likely proved to be a royal nightmare to implement. I abandoned it fairly early on when I realized that reliability simply had to trump power consumption, as near conflicting as that sounds. As it turns out, the method I ended up with is much more predictable, so that's what I'm going with. If someone wants to know more about how this device works, let me know and I'll put up some details. It might have a use somewhere, but I don't think the inside of a fridge or freezer is the place.
[attachimg=9]
The ice sensor, laying on it's side in the bottle without any water, posing for it's first pic.
There will likely be many more gotchas, but this is the basic list of what's been encountered so far.
Till next time...
Steve
-
Did a little rearranging and optimizing in hopes of characterizing just a little better before I make the plunge into the chip control.
Hurricane Arthur, while being a "disappointment" for our area (in terms of actually getting much of anything from the storm itself), turned out to carry a hidden lesson about the overall design of the fridge.
In preparation for the storm's arrival, I went through and neatened some things up, added a battery backup for the blower/recirc fan, and manually pulled it down as hard as I could as any power failure would have been useful in terms of an excuse to see what this thing was really made of.
In the process, because I've been using thermisters and had only a rough calibration to go on, I missed the mark on the water mass in the fridge chamber. I was only off by a single degree F, but that was enough to cause a rather significant malfunction.
While I was aware that I was likely supercooling the mass, I was having trouble preventing it. As the water sits, the crystal seeding points dissipate and eventually disappear. I only briefly looked for a way to keep a persistent seed source as I needed to find out exactly where this point was, but nothing simple was coming up. My end solution involved periodically shaking the bottle up vigorously to get some air bubbles trapped in it so the ice would seed if it was going to form, right at 32F, and no supercooling would occur. The night of the hurricane, apparently, it all came together and I discovered ice the next morning.
The result was, because of the separate nature of the control for the blower, the air temp inside was beginning to warm up just enough to cause the blower to turn on, bringing cold air from the core up into the chamber and further perpetuating the ice in the sensor mass. Because of the ice, the compressor never started, and it was simply "draining" the core instead.
The good news in all of this is that I was able to pinpoint the exact freezing point and set the lower threshold to avoid it in the future. Had I been using digital sensors to begin with, this never would have been a problem, as I could have just set it slightly above from the beginning, but you go with what you've got sometimes, right?
While there was no actual power failure, there might as well have been. It ran for just shy of 12 hours without compressor, and the core held it's ground. A good sign ultimately. It would have easily run much longer than that, but I stopped everything to thaw out the frozen fridge mass bottle.
I decided that this point would also be a good time to take the center bottle out of the core and put the sensor on it so I could gauge the core's ice temp. I removed the bottle, and put it in the cab of the van, in the sun, to thaw it as quickly as possible. It took about 3 hours or so to completely get rid of the ice, during which time I worked on other aspects of the system.
[attachimg=1]
The core bottle with the sensor attached and insulated from the surrounding air.
The thaw was necessary to stop the formation of condensation which would have interfered with the adhesion of the tape. After the main thaw, I was running out of sun and still needed to raise the temp somewhat to get it up to room temp. Nothing a couple of minutes in the microwave didn't handle with ease. :)
[attachimg=2]
The bottle placed back in the center of the core, ready to resume it's purpose.
Up until this point, I had only been measuring air temps in the core, and suspected that the temperature plateau observed due to heat of fusion when fresh water freezes was skewed, and it turns out I was right. Monitoring the bottle directly shows that the temperature falls relatively quickly as it approaches the freezing point, but doesn't then simply level off as the ice forms. Instead, the freezing point slowly, but progressively drops for the remainder of the brine due to fractional distillation and increasing saline concentrations. The result is a slight slope during freezing/thawing, which turns out to be a useful characteristic that can be used something like a "fuel gauge" for remaining cooling store. I was skeptical that this wouldn't be possible when first venturing into this because of the well established behavior of fresh water as it changes states. Now I know, and ultimately, even if I were using the capacitive ice sensor in the design, this would supercede it. Live and learn.
Another note about thawing out just a single bottle...
I have been pondering from early on how I would go about defrosting this beast when it gets so clogged up that air can't efficiently flow through it and the cooling power diminishes to the point of hopeless. I can say as of now that the "one bottle at a time" approach will be a highly unlikely method.
The thawed bottle reminds me of what happens when there's a dead cell in a battery... It provides a constant strain on the remaining bottles, and throws the performance off, and a lot of energy can be wasted trying to pull the one bottle down. The heat doesn't propagate quickly from bottle to bottle in there, and so there's a lot left simply to time to sort out. I'm still pulling down on this one bottle, and it's still nowhere near the rest of the core yet, and I've dumped about half (so far) of the total energy required to pull the entire set down from room temp.
One possible solution I'm considering is yet another fan to circulate air down there, but that comes with it's own set of problems. Aside from adding to the parasitic power load, I've noticed that the bearings on these small brushless fans do not like the extreme cold and tend to chatter readily. I've been able to mitigate this somewhat so far with speed limits on the other two fans, but a fan in the core would be exposed to some really extreme temps, particularly during hard pulldowns. At times, the core air can go below -10F. I've heard the blower bearings chatter under the right circumstances at a paltry +10F, so the lifespan and reliability of a fan in such conditions may be rather limited at best.
One thing is certain, eventually, I will have to thaw it out, either partially or completely, and it's an energy intensive process either way, not to mention must be prepared for ahead of time (consume all perishables etc) and produces a lot of runoff that must be dealt with in the process.
[attachimg=3]
I also moved the core air sensor to just inside the downtube for the blower intake. So far this looks like a good location. Whether or not this sensor remains here when it's all over and done with is yet to be determined, but so far is helping draw the big picture by providing confirmation of things like heat load and air flow etc. Time will tell.
[attachimg=4]
A shot of the back side of the freezer, not included in the original update because of pic restrictions. It's just as well, because I just recently added the door switch that kills the blower/recirc fan. You can see the wires going up into the right hand side hinge. Currently, they simply interrupt power going to the two buck converters that power the motors, but will eventually just tell the PICAXE that the door is open and software will deal with controlling power to the fans. The compressor cooling fan is down on the lower left.
[attachimg=5]
Detail of the door switch. There's only one spring loaded opening assist mechanism, and it's on the other hinge, leaving this one open with plenty of room to put a leaf type microswitch in.
This was the only "permanent" modification made to the freezer, two holes drilled in the side of the hinge arm to hold the switch in place. Even the PFC cap is just plugged in.
[attachimg=6]
Last but not least, the final version of the top gasket. Up until this point, and not visible in any of the other pics, I was using just a full piece of foam with a gap on the left side to make it easier to get out and to give the wiring a place to go that wasn't restrained so I could move them around if I needed to.
With this in place, I can now stand 2L bottles of soda upright inside and close the door properly. The lack of extra insulation in the access port also lends nicely to a little extra heat leakage that helps keep the differential between the two chambers under control (when I'm not foolishly trying to manually push things to the limits, that is ::)).
The two temp sensors along the bottom of the fridge chamber are still there and the wiring still runs along the surface, which interferes with the physical placement of objects to be cooled. In the final design, I'd like to move these, but I'm not sure of the best location for them. I've experimented with moving the control sensors, recirc fan, etc around and the results aren't necessarily "better" in one location or another, just different. But controlling the blower speed has the greatest effect when it is running without power, and so I'm having trouble deciding if there would be a better location for them, and if so, where? If not, should I just trench the wiring in the bottom of the chamber?
Any input on this concept (or the defrost process as well) would be greatly appreciated. I'm weighing the pros and cons of everything on this, so I'm open to whatever ideas anyone might have.
Till next time...
Steve
-
-10F *extreme temps*
Wow man, spend a winter up here in frozen hell... err uhh,
I mean Minnesota.
THEN let me see what your definition of extreme temps are.
-
Naaa, I'm good LOL
They don't like it on the other side of zero tho, and neither do I :P
-
Jeez, walk away from a prototype for a few... uh, eons...
[attachimg=1]
Not sure what this says... I guess one of two things:
"Soon I'll need a prybar to disconnect all of it to build the real thing", or
"It works so well it hasn't been touched"
The latter is certainly true... this hasn't been abandoned (regardless of what the spiders might think)... the truck itself has simply taken precedence.
I very recently pulled all the plots from various stages and conditions, will be digging thru them and finding some key ones that "show it off" so to speak... Luckily, the cobwebs seem to have stayed outside the computer; this thing is a solid performer, even though I still want to try the core mods. If I have time, maybe, if I have time...
Until next time...
Steve
-
Steve
What are you, some kind of madsientest or something?
cheers
gww
-
LOL People keep saying that... I don't believe them... I mean, don't mad scientists do really off the wall experiments and build strange things? :o ;D
Steve
-
Wow, uhh, ok... this has come a long way since I posted last... even got the 180 day warning from the forum...
I'll give more details about what's going on here shortly with this... those following the scratch pad (http://www.anotherpower.com/board/index.php/topic,890.msg10246.html#msg10246) thread have had a preview already of what's going on behind the scenes. I apologize for that type of posting, but I'm using a phone for all this, and I also am trying to keep the "end result" threads reasonably clean. I digress...
[attachimg=1]
I present to you PiFridge 1.0 ;D
The nutshell: It controls the compressor and blower... Blower turbo and disable, and otherwise continuously variable based on feedback from a thermister. I went this route because in the event of a failure, it is set up so that with no other inputs, the buck converter tries to maintain the fridge temp feeding from the core.
You'll note as well that there is no heat or stir fans involved. I don't believe at this time that they're necessary, but can easily be piggy-backed onto this layout if required.
More to come soon...
Steve
-
Had a few glitches with the temp probe bus, and found that it really was desirable to have a fan on the compressor, so there's some deadbug and grown-in-place happening here.
[attachimg=1]
2 buck converters, an interface board for isolating the Pi, and a GP A-to-B type deal (mostly just for the RJ45 that heads to the fridge, but there's a couple other components on it). Not the best layout, and I wish I could rebuild it and make it all pretty, but that's not gonna happen...
[attachimg=2]
It's a solid performer, the whole thing more or less meets expectations, aside from some miscalculation on my part... LOL No accounting for the human factor (other than to say yep, it's gonna be weird if it comes from me) 8)
I have some more pics of the fridge itself in its more or less completed form... I'll get to them shortly...
Till then...
Steve
-
Controller up on the wall and dressed in... finally starting to really clean up and look like something in here for a change...
[attachimg=1]
More to come...
Steve
-
Been meaning to get these up, as well as the stray remaining pics... For now, graphs...
When I leave it alone and let it do it's own thing, it does rather well... The past week looks like this:
[attachimg=1]
[attachimg=2]
The upper graph is pretty self explanatory... The only things to really mention is that the water sensor is near the supply, and the air sensor is right at the return. This appears to be a decent arrangement after lots of tinkering and what not... The other mention is yes, it operates centered on 0C, a normal household fridge typically runs higher. This is due to the difficulty of maintaining the two relatively close chambers at two rather different temps using nothing more than blower speed.
The "score" graph, simply put, is it's "EEG", a peek directly into it's brain. There are several influences on it's behavior, each of which gets normalized and then averaged. The result is the red line, which becomes one of the two master controls to start/stop the compressor.
The grey lines represent the "soft" limits, the upper being start, the lower being stop. Likewise in the main fridge plot, the water temp has hard limits; it is the only sensor that can override the rest. All 4 trip points are regularly hit during normal operation.
Explanation of the monitored data:
Scr.ToD: Time of Day. Derived solely from a simple sine function, it is used as a generic predictor for ambient heat, and tries to turn the compressor off in the afternoon, during peak heat of the day. Presently centered around 4PM, as there's additional lag caused by the thermal mass of the living space itself.
Scr.F_Water: Corresponds to the water bottle in the fridge chamber. Easiest to think of this as a zoomed in view of the space between the lines on the regular fridge plot. While this sensor has ultimate control, it also helps convince the brain if it's on the fence about starting or stopping.
Scr.Core: Based on the ice core temp. There is some extra math applied to this value that causes it to "speed up" as the core temp gets lower, essentially forming a knee in the score that suddenly tries to stop the compressor as the temp crosses the -12C mark. IOW, not a linear relationship like the rest of the inputs.
Scr.Power: Based on the battery SoC, currently set as upper=100%, lower=70%. I'm considering letting this one extend "beyond zero" so that when the batteries are lower than they should ever be, this value can ultimately stop compressor operation altogether by driving the average too low to ever start it. This value is also slated to be forced to 100 when the truck is connected to the grid or otherwise detects a non-renewable power source. Simple as it is, the "sensor" involved in this (a tricked out cell phone charger and an optocoupler) still hasn't made the finished list. I could have assembled and installed it in the time it took to explain this... I digress... ::)
Scr.Amb: This design may be one of the few (only?) fridges ever to take user comfort into account when making run/stop decisions. Most of the time now this value stays at 0, but once in a while pops up when the room temp is low enough. The idea is simple, all else considered, if the room is on the warm side but the conditions in the fridge aren't desperate, put off running until it's cooler. It worked well during spring when the temps are all over the map, adding a little heat to the room when it's chilly, and holding off when it's warmer. Ultimately, the other sensors win out of course, and the value effectively becomes meaningless during summer.
Scr.Pyr: Pyranometer, telling the brain what the theoretically available power is from the sun at the moment. This input is presently represented by a simple "rectified" sine wave, and essentially "every day is sunny". This will be replaced by the input obtained from a "poor man's pyranometer"; a small PV string (think "large" yard light) dumped into an appropriate value resistor. The voltage that appears across the resistor can be used to approximate what can be expected of the main array. It's not perfect by the definition of "pyranometer" but will serve the requirements well enough for these purposes.
Scr.Boost: Last but not least, this too is derived from a sine wave, and can be thought of as the "master sync pulse" that tries to make the compressor start early in the morning. There is one pro and one major con to it's attributes... It helps dodge peak heat of the day by starting before the sun starts really warming things up, but must be balanced with the idea that the compressor is starting when the batteries are theoretically at their most dire point. Its amplitude is such that on its own, it can't trigger a start unless it's already approaching it anyway, and essentially works in tandem with the pyranometer input to sway the actual start time.
This input is also associated with the transfer blower being disabled, causing the fridge chamber to begin to rise, further convincing it that it should start early in the morning if it's ok to do so. The blower remains disabled until first shut down. It then uses analog feedback to "pull coolth" into the fridge chamber as needed. In the event there is more than one compressor run in a day, the others aim at running as little as possible by turning the blower on full tilt while the compressor is on, so as to satisfy the needs while dumping as little heat as possible into an already hot living space.
[attachimg=3]
[attachimg=4]
During boost mode, it will toggle the blower on and off as necessary if additional load presents itself. These are today's plots, and actually, there was some manual override today as I actually wanted some of the extra heat for testing the battery cooling system, but the behavior is identical regardless. I had artificially set boost mode late in the day after having locked the compressor out until about noon. When it was subsequently unlocked, and later heavily loaded, the blower was modulated over the course of a few hours. The result is clearly visible here, as it "pulled" from the core in order to maintain regulation.
As you can see, it indeed has decent regulation when it's left alone (and even when it isn't, to a point... can anybody spot where else I screwed with it? LOL). It typically runs once a day, coasts the rest on the ice, and there's always a cold soda (when I remember to load it that is!)
Loading plays the biggest role in ultimately determining what it does when. How the load is placed also heavily affects operation. As a general rule, with 12oz soda cans only as a load, ~75% full is about optimum.
There are two "acceptable" methods to load it. One is what I refer to as "the twofer"... Take one can out, put 2 in. Obviously this requires observation of the total loading and attempting to maintain 75%. Until ASIMO comes to live with me, this task is likely going to be up to me to keep up with ;D
The other is to "shock" it... allowing it to get down to about 50% load over the course of the day, and then replacing the "missing" 25% somewhere just before boost begins. When this method is used, it's generally best to disperse the new load as much as possible throughout the existing.
Also, a strategically placed warm can or two at one end and/or the other (supply vs return) can mean the difference between running or not, and requires some judgement.
All in all, it works very well... It doesn't perform quite as well as it did on "the bench", but then again, what does? LOL
I guess all that's left to be said about this one is that there are a few pics I want to post that haven't made it yet... Time is crunching, priorities shift constantly.
Oh, except one thing... I know someone will call me out on it LOL - There's plenty going on as far as the batteries go, however they are not only aging but are also only half the label capacity the finished design calls for... As a result, to minimize unnecessary stress on them, the fridge is being supplied from the grid at the moment. So if you thought the SoC score looked a bit off, it is indeed, and bonus points for catching that in that tangled mess 8)
Until next time...
Steve
-
Time for a very overdue update...
This is without a doubt the top of the list for most-everything in the sub-project category... It took a lot of time and patience, and as predicted, it absolutely had to be running on solar to be correctly fine tuned. Some things just really can't be properly emulated.
In the end, the entire score system had to be essentially completely rethought, which wouldn't be a first, but there really was no way to tell exactly what was needed while it was still bypassed to grid.
Just so happens that today catches just about every aspect of typical operation, so rather than snag and post individual plots, I just took screenies of the portion of the still rather primitive webapp that is used to observe/control it.
[attachimg=1]
They don't all fit on one screen, so here it is sectioned up starting at the top... In this shot, the vitals.
[attachimg=2]
This section is the final version of the brain, a bit different than originally envisioned...
[attachimg=3]
And here we have the "extras", for quick reference to assist in setting up the various triggers and factors that pertain to energy production, use, and storage.
I'm holding out on what's behind "button number manual"... the rest of the webapp being slated to be revealed in a special way, just for you guys... I did tell you, after all, that graph pool comes together in a sensible way... ;)
The vitals section hasn't really changed much, still reports the crucial info needed to know at a single glance what "state of charge" it's in, and the condenser temp thrown at the end of that because it was the best fit. It's used primarily to gauge thermal efficiency; deciphering such being more a black art than science really.
The decision system, as mentioned, was totally redone. My "score" idea was useful but I wasn't implementing it effectively to run on solar.
The score plot is still basically a set of weights and a balance, there's just less of them now, with most of the input being in the summing counterpart, "Pyro DSP". But before I go into that, a quick sentence or two about the remaining "score":
The two grey lines form the normal operation start and stop boundaries, start being on top. Green is an amplified and clipped version of "Pyro_Final_DSP" in the DSP plot. Black is a suppressed zero component representing the top 5% of battery SoC, and Red is the average of those two values. There's a third hidden value that never changes; it's the leftover remnants of the "user subjective" value... This is all delicate enough that for simplicity and expedience, I just tucked it away and hardcoded it in the software. This is what's responsible for the seemingly not-quite-average the astute eye might pick up on...
When red crosses the upper line, it initiates the compressor start sequence. Likewise, crossing the lower line shuts it down.
The source of the green line, the above mentioned "Pyranometer (DSP)" plot's final value, is somewhat complex in the way it's derived. First and foremost, the pyranometer is the basis of the entire signaling mechanism. This, in a nutshell, measures the available energy in the light hitting the panels, and while not a true pyro in the usual sense, it's close enough for all I needed it for, and that's strictly to cue the fridge. The light grey line ("Pyro_Raw") is the signal as the ADC sees it, and as you can see, can be very volatile. In order for this signal to be useful, it needs to be smoothed out somewhat, enter "Pyro_20_Min_Avg", in blue. This line is the "crystal ball" so to speak that attempts to give the algo a heads up about what the sky is likely to be doing next. It doesn't tell the future obviously, but uses the last 20 minutes (in 5 minute sample intervals) to pass trending light conditions to the decision system. This is a balance in the sense that if it's too short, the compressor cycles potentially unnecessarily, and if too long, can leave it running on battery, the primary condition the whole design is intended to avoid.
Next up is "Pyro_Boost", in green, which is kinda mislabeled, as it's function is to attempt to start the fridge early if the quality of the earliest light meets minimum criteria to do so. This is probably one of the most difficult entities to work out and set up... and only experimentation and observation over time with many different conditions presenting can properly configure it. This also isn't a static pulse. It is slid earlier and later based on a crucial but borrowed piece of code. Thanks goes out once again to Ross for helping me with this, math certainly isn't my strong suit and there's plenty of it going on in there. As mentioned earlier, this graph views primarily the summed weights, and this bump is added to the 20 minute average.
The last piece of magic in this puzzle is the magenta line, "SoC_DSP". Again, somewhat misnamed, it's also somewhat of a redundant input. It's purpose, when combined with the other battery SoC (in the score plot), is to encourage it to stay running when the battery has been recovered. The two weights serve slightly different purposes. This value effectively increases the perceived available energy, while the other is used primarily to keep the compressor running until it actually IS running on battery. This is essentially the boost pulse's counterpart for extending operation until the last available light of the day is gone.
As previously mentioned, all of the major features of the algo are visible in these two plots for today. The Boolean plots above the two control plots represent the various states of different aspects of the rest of the system. There's a lot happening there, but the keys that I'd like to really like to point out are the "Truck_Running" and "Fridge_Emergent"... the others are either low/no relevance to the fridge, or are fairly self explanatory.
The "Truck_Running" key is just what it says, the mystery for some is in the "why"... It might not have much actual effect on the compressor, but my thinking behind checking for this was to prevent excessive forces on the internal components of the compressor. I tried my best to reduce risks as much as I could given everything going on overall, this was one of them. I can't afford to replace the unit, even tho it wasn't that expensive really to start with (around 100 bucks at Lowe's). Nutshell version of "compressor 101"; the motor/compressor assembly is suspended by springs inside the black ball many of us are familiar with. I've personally seen that one mode of failure is the compressed side tubing leading back to the case breaking off, leading to a "running but not doing anything" unit. Bumps, gyro effects, etc. So the ultimate last piece of code in the software kills it no matter what (can't be manually overridden even) to protect it.
The other key worthy of mention at the moment is "Fridge_Emergent". Normally, the blower between chambers is disabled, and natural "pooling and spilling" is exploited to give the ice core the advantage during operation on solar. Aside from the truck running and manual operation, in auto mode it starts and stops by one of two sets of criteria. The above describes the normal cycle. It also is shut down when the fridge chamber gets too cold, presently set at -2.5C.
Fridge_Emergent however causes it to start when the fridge chamber gets too high (presently +5C), and runs the inter-chamber blower at full throttle, so as to drop the temp as quickly as possible and shut back down. It's very likely that if this condition occurs, everything is already short on energy supply, and it's just about certainly running on battery (again, this entire thing's point is to avoid that if at all possible). It keeps track of which trigger started it so that it knows which criteria should shut it down. This unfortunately induces an undesired behavior if the sun should come out, but it actually still remains an unverified function - it's never actually reached that point during operation since this was finalized. I've emulated it best I could without actually letting it "fall up" to temp. I may try it some day, as occasionally it does need defrosting, but in the mean time, I'll likely already be very aware that things have gotten to that point.
The blower, when the compressor is not running, operates via analog feedback, just as in the original design, attempting to hold the air/water temps just barely above freezing.
That's all I've got for now (Whew!). Until next time...
-
I was wondering what you were up to, and you've been busy. Busy is good though.
-
Yeppers busy for sure... tho the fridge has been pretty much matured now for a fair while... I haven't had time to really catch up on posting anything... I combed thru last night and looked, some of the stuff just completely left hanging, but this one had the most to report I guess. The scratch thread and chatting in IRC has given the subconscious illusion of updating for me apparently as well LOL.. I'm gonna try to play catch up, nothing else really has had major changes since January.
Steve
-
Hey Steve Howdy!
Just read this thread top to bottom, nice! Love your troubleshooting techniques, we must be long lost brothers or something. HVAC and ACR can throw some weird crap my way sometimes.
3 questions:
1. How would glycol, simple as polypropylene 3350(available in the pharmacy), work out as an alternative to brine. maybe you could throw that experiment up here.
2. Ever thought about re-routing to a cold plate inside, sail boats use them without fans or controls of any kind (no controls, what fun would that be right?)
3. You substituted a 10mf in place of a 12mf that's a lot more than +/- 5% I assume it's humming right along just fine. most small appliances I repair use a 5mf, 3mf or nothing at all just relying on start relay. hmmm (the things that make you say hmmmm)
regards,
SN
-
Hey Steve Howdy!
Just read this thread top to bottom, nice! Love your troubleshooting techniques, we must be long lost brothers or something. HVAC and ACR can throw some weird crap my way sometimes.
3 questions:
1. How would glycol, simple as polypropylene 3350(available in the pharmacy), work out as an alternative to brine. maybe you could throw that experiment up here.
2. Ever thought about re-routing to a cold plate inside, sail boats use them without fans or controls of any kind (no controls, what fun would that be right?)
3. You substituted a 10mf in place of a 12mf that's a lot more than +/- 5% I assume it's humming right along just fine. most small appliances I repair use a 5mf, 3mf or nothing at all just relying on start relay. hmmm (the things that make you say hmmmm)
regards,
SN
SN -
1. I'm not sure how the glycol mix would behave in this. I'd imagine the right mix ratio and temps and it could be made to behave similarly. The intent with the brine is 2 fold... First, it lowers the freezing (and therefore thawing and operating point) to well below 0C. Remember that the phase change is the most prominent feature of the storage medium being exploited. I don't want it to stay a liquid all the way to bottomed out (mechanical thermostat limit cutoff).
The second is more of a side benefit, as I'm finishing up now a "reserve gauge" that is based on a twisted averaging calculation to give a rough estimate of the percentage of ice in the core. It won't be perfect because of things like thermal impedance and not (yet) aware of thermal loading (eg changes in ambient temp and sudden loading of the fridge chamber)... however, because the temp continuously falls (and rises) without hitting the plateau associated with pure water, the amount of ice can be estimated on the average core temperature. As it freezes, the remaining brine gets more concentrated, requiring the temp to go lower and lower to completely freeze it (which doesn't quite happen with this combination (see the bottle showing the inch or so of brine that still remains on the first page).
2. I'm not entirely sure what you're referring to with that, but I can address the "controls" aspect of it... the reason this thing performs so well is because it's "brute force micro managed", if that makes any sense :o
3. Power factor correction for a "purely inductive load" such as an induction motor as found in a compressor follow a sine wave when you give several values for capacitance. 10uF isn't perfect, no, but it's only just barely starting to work it's way off the peak of the wave (which represents unity PF). 14uF would have a similar PF, just the current leading instead of the voltage as with the 10. Either way, actual measured PF with a kill a watt sees 0.98~0.99 with the 10uF... Still a very nice looking number compared to the much lower high 0.6 to mid 0.8 (depending on load... it's lower until the Freon pressures normalize) I was reading without it.
Also, don't mistake a start or run cap as a PFC cap. PFC caps are not connected the same way as the other two. PFC effectively goes across the mains feed at the motor. Start and run caps are part of an auxiliary winding that's used to cause phase shifting that makes the motor start and or run with the correct amount of slip. They will *affect* power factor, but the PFC cap is placed in *addition* to any start or run caps, and has nothing to do with the way the motor runs. They strictly control reactive power that otherwise gets lost as heat in the motor and upstream supply chain.
Hope this helped answer at least one of the questions ;)
Steve
-
Hey Mad How are you!?
I see what your saying about the compressor cap and it's association with start and run. Most small btu compressors I deal with simply use the inductance relay to shift from start winding(some need a start cap for the kick ie., restaurant equipment I repair) to run winding without caps (small home appliance) and if I understand correctly from what your saying the cap on some I encounter is across the start and run terminals, Is that a phase shift or a compensation for power factor of the run windings? When I encounter a compressor that won't start and run and is tripping the external overload I will install what's called a 3-n-1 start run kit. It's electronic. has a start cap appropriate for the hp and leads that slide right on where the relay and overload normally plug in.
As far as a cold plate is concerned it may not be as fun. it's a stainless box with and evaporator tube coiled inside and filled with non-toxic glycol. In your case it would be bolted inside the lid or to one side or even bottom of the freezer and connected to the refrigeration components by flexible lines. I had one for awhile but a friend has it in his sailboat now. fewer bells and whistles involved but it would be sized and temp controlled for freezing or cooling. some boats have them mounted to the side with a platform or basket made from thick plexi midway down for cooling and underneath for freezing. The whole idea is that the cold plate stays frozen for a long period and relieves the need for long running periods of refrigeration etc etc.
Thanks for the great info!
SN
-
Hey Mad How are you!?
I see what your saying about the compressor cap and it's association with start and run. Most small btu compressors I deal with simply use the inductance relay to shift from start winding(some need a start cap for the kick ie., restaurant equipment I repair) to run winding without caps (small home appliance) and if I understand correctly from what your saying the cap on some I encounter is across the start and run terminals, Is that a phase shift or a compensation for power factor of the run windings? When I encounter a compressor that won't start and run and is tripping the external overload I will install what's called a 3-n-1 start run kit. It's electronic. has a start cap appropriate for the hp and leads that slide right on where the relay and overload normally plug in.
When I was in HVAC, we called those "hard starters"... tho they're possibly slightly different going on your description alone. I know of one model that did away with the start cap entirely, and shorted it directly to hot (via a set of normally open current relay contacts). With no power applied, the start terminal only had the run cap across it (which depending on which compressor it was intended for, may or may not have had an augmenting cap included). When the contactor closes, the LRA pulled the start relay in, and connected the start terminal directly to hot (run), and upon reaching speed, the current would fall below holding threshold and the contacts would open, letting it run normally (or augmented as the case may have been).
Usually we'd determine if a hard start would be useful by the good old fashioned screwdriver test... Hold the screwdriver you opened the access panel with across the start and run terminals, and manually actuate the contactor, taking the screwdriver away a split second later. If it started, a BRB was all that was needed to get it running for at least the moment ;)
There are several variations of induction motor out there, but they all have 1 thing in common (in this context); they require some means of generating a "rotating field" in the stator. Normally a compressor motor achieves this by an auxiliary winding that's physically misaligned with the main winding, as well as being phase shifted (the job of primarily the run cap). The start cap serves more as a current limiter for the start winding in a correctly functioning motor, and while there's some phase shift associated with it, during start, it's mostly about providing the extra power needed to get the rotor spinning. A "hard start" takes this to the extreme by essentially allowing "unlimited" current to flow (only limited by the relatively low inductance of the winding). As such, further stress is placed on the motor... the compressors we'd install them on got tagged with nicknames like "the running dead" and such, as they might last thru a season in a pinch, but invariably indicated it's days were numbered (as you probably are aware). We would get the seasonal rush calls, install them for people that couldn't afford new condenser units right at that moment, and as stop gaps to reduce backlog for those who were in position to replace the units. Those usually saw a few units before they couldn't be accounted for or turned up dead themselves LOL We'd often reuse them (with the customers approval), offered up as freebies with the upcoming system replacement in mind.
As far as a cold plate is concerned it may not be as fun. it's a stainless box with and evaporator tube coiled inside and filled with non-toxic glycol. In your case it would be bolted inside the lid or to one side or even bottom of the freezer and connected to the refrigeration components by flexible lines. I had one for awhile but a friend has it in his sailboat now. fewer bells and whistles involved but it would be sized and temp controlled for freezing or cooling. some boats have them mounted to the side with a platform or basket made from thick plexi midway down for cooling and underneath for freezing. The whole idea is that the cold plate stays frozen for a long period and relieves the need for long running periods of refrigeration etc etc.
Sounds entirely viable... I've come to the conclusion that there are several ways to strike back at thermodynamics... I'm all about the ones that win! ;)
Steve
-
Went to toss a quick "by the way" at this, and realized it's been forever and there's been yet another major revision done to the algo... but I'm gonna have to come back to that... LOL... and as of once again shifting priorities back in the winter direction, I see another tweak to that involving what's below. What's that saying? Finished but never done? :o ::) ;D
I'm not gonna make any bones about it, I drink a lot of friggin mt-du (I don't recommend it therefore I'm skipping shame-like plugging lol)... and in general, that's what you'll find the fridge keeping cool. There are reasons for that particular selection, most of which are well outside the scope of this thread, and even this one sorta is, but I figure since it's involved in a rather direct way, I'd put it in the fridge thread...
One of the things I use it for is auxiliary heat. And if I time things right, I can get a lot of bang for my buck. Last winter I burned ~25 gallons of propane total... This year I'm determined to get that number under 20, and 16 is the goal I'd just be tickled with.
Timing however is only part of it... there are several other ways to manipulate and optimize the fridge's role in heat supply during winter, and arrangement of the loading patterns is one of them.
[attachimg=1]
This go round, I let it mostly empty out, trying to coincide my consumption, running, and loading for "empty + cold front + 'depleted' core", so that I could fire it off in the middle of the night for a couple hours and have it pump a lot of heat very efficiently.
Pulled the last cold can out, and then fanned out the next case as seen above, combined with the transfer blower forced to full throttle, and that's exactly what she does.
Here in a bit I'll grab the ambient temp graph and show the results, but wanted to get this part of the story up and posted for the time being.
-
Ok it's hard to get an "all else equal" view of thermal data of course, but the run I was originally going to grab didn't illustrate it really well... I noticed that today that the effects of both the propane and the fridge are rather predominantly and independently visible, particularly in the change rate plot...
[attach=1]
I decided that's probably the best view because it shows the relative timing and influence of the sun, but it's obvious where the main propane burner was both started and stopped... And while the fridge doesn't have super obvious start and stop points visible there because it "fades in and out", it's pretty clear just how much effect it can have.
To that end, I'm including the condenser temp plot to show its timing more explicitly. Note this subsystem uses CV (continuously variable) fans, hence the steady operating temp and "fading" effect it has on the space heating.
[attach=2]
Not a bad system for winter... Just wish I'd have had time during the build to come up with a better way to deal with this extra heat in the summer... it's hard to cram all of this into a 14' truck... all I can say there lol
Til next time...
-
I'll call this an addendum to the above...
[attach=1]
Here's as close as I could get to "all else equal". Held off on running the fridge at all today to grab this so the differences would be more apparent.
I'll add that in both scenarios, the pilot for the heater was on the entire time (I've brushed on the effects of that elsewhere) - In the environment I've got going here, non-trivial effects in itself.
The blip on the outside at about 5PM is engine heat from a run up to the store.