Last night, when the ambient temperature fell to above 30F first time in this season, my A/C started to act really funny. Check this out:

http://freehold.crocodile.org/temp-day.gif

At about 4:10AM, it stopped cycling for almost an hour, then switched on and worked for more than 90 minutes, then repeated the same in about 20 more minutes. Obviously, there were no changes in the settings, and I don't see how the outside temperature drop could have affected the switch.

Anybody out there can give me any clues?

Other than that, the A/C is cycling way too often, but that's another story...

PS: the temperatures on the graph are in celsius
__________________
--vt

OK, let me see what I can sort out here. . .

We are obviously talking about a heat pump; no?

The coldest outdoor temperature corresponds to the unit cycling off compeltely?

Which also corresponded to a 2� drop in the house temp, and a different temperature profile in the house?

I can do nothing but guess, with the information I have so far and the assumptions I have made, but I guess you have a duel fuel system, your outdoor stat is set for about 32-34�F, when the outdoor temp dropped below the setpoint, heatpump cycled off, gas furnace supplemental took over, and the temp overshot and undershot (as furnaces often do) until the outdoor temp came back within heatpump range, at which time the heatpump kicked on and ran non-stop till it caught up.

What model and serial number do you have?

Is the supplemental heat actually gas, or electric?

Can you spread the abiscus (sp) so I can see shorter time encrements?

And, can you possibly lose the stinking C scale? (as a personal favor to me?) It reminds me of a recent terrible experience I suffered.
__________________
There are no stupid questions. . .just a lot of inquisitive idiots.

BamaCracker:

A short disclaimer I forgot to prepend: I'm not a HVAC professional. I should have provided that information ;/

Yes, we're talking about the heat pump. However, this is just a single stage electrical air conditioner/heat pump, no dual fuel. That's what made me scratch my head.

I'll have to get up on the roof to get the model and serial number, but not now So far, here's one more fact that might have been relevant: the thermostat is the Honeywell Chronotherm III.

Side question: what's abiscus? I can guess that you want to see just the relevant time interval, yes, I can do that in a couple of hours, as well as try to change the temp to F - the graph was a current snapshot, it's gone forever and I have to harvest the logs now.


__________________
--vt

That graph looks like my last EKG!! Or was it EEG??
__________________
Dumb sounding questions are better than Dumb mistakes

VT;

Sorry about my atrocious mispelling!

An abscissa is the horozontal axis of your graph. I would have liked to examine it in more detail. The data acquistion system we use here will allow us to go into history and zoom in to see more detail.

The electric heat may/can also be conjoined to an outdoor stat, but not usually to disable heatpump operation. So, with electric heat, I am more pone to believe you actually have some malfunction. Our confirmation of this will be repeatibility. Tonight, assuming the weather gets cold enough, we'll see if it does it again.

Once we know what kind of unit we are dealing with, we can give you some more specific responses.

Rayr; if your EEG or EKG looked like that, congratualtions on your miraculous recovery!
__________________
There are no stupid questions. . .just a lot of inquisitive idiots.

BamaCracker:

I don't have any industrial data acquisition system, just a bunch of sensors, however, I do have 1 minute increment logs.

Here's the narrower selection, yesterday night and this night:

http://freehold.crocodile.org/temp-cut.gif
http://freehold.crocodile.org/temp-cut2.gif

Yes, it did exhibit the same behavior, but to the lesser extent.

I don't think there's any outdoor stat - the system is about 30 years old.

What I'm thinking about - could it be that the switch got stuck with the frost and failed to switch the pump on, even though the stat commanded it to?
__________________
--vt

Give me some time to think about this, and try to find out for us what type system this is; i.e. model and serial number, brand name and any distinguishing characteristics.

It's not a water source is it?

Do you have the ability to track electric consumption as well? That would be a nice plot to look at.

Does the blower (indoor circulating fan) still run when the unit is down?

The frozen switch thing is highly unlikely, since the contactor coil would act as a heater when voltage was applied.
__________________
There are no stupid questions. . .just a lot of inquisitive idiots.

BamaCracker:

- no, it is not a water source (though to be honest I don't know what is a water source, I just think it isn't 'cause there's no water)

- I can't track the electric consumption

- the blower isn't on when the pump isn't, this case gives a distinctly different temperature change pattern

I wouldn't be able to get the model and S/N until tomorrow, though will get them definitely.
__________________
--vt

vt,

Can you tell us about the equipment and software you used to produce this chart. Way cool.

Gene
__________________
Just a house nut.
"Reality is that which, when you stop believing in it, doesn't go away." - Philip K. Dick

Gene:

Hardware is 1-Wire sensors - good resolution, negligible cost (about $2 a piece for a sensor, $10 for an adaptor), high reliability and flexibility. Check out http://www.dalsemi.com/.

Wiring is regular CAT5 cable.

Software is rrdtool (search on google) to graph data, and some code I wrote to collect data.

Actually, it's all a part of a bigger project - smart A/C. You see, I have a shitty tri-level house with a single A/C, and it's not possible to balance it at all - the daily temperature range is up to 50F (Arizona ), so I'm putting together the sensors and active zoning to balance the house. All the stuff is under $500, not including R&D

I've been looking over your plots, talking to some friends, and thinking. . .(it's hard to believe I actually get paid to be there ain't it?) So, here is where I am with these thoughts.

From what you have said so far, I assume this to be a package unit (one single piece of equipment, containing the entire system except the thermostat) is this correct?

You mentioned climbing up on the roof to find a model or serial number. Is this package unit mounted on the roof? If yes, how does the air enter the house?

It's 30 years old? Pretty dadblamed old for a compressor bearing anything.

Have you begun any of the interfacing to automate your damper system? I wondered because I noticed some rooms track different than others. You use a setback thermostat I take it? Or do you just turn the heat down manually every morning when you get dressed? Or is that permanent drop a symptom of the failure? If you do have a setback t-stat, what kind?

I am curious about whether you have supplemental heat or not, but that is not something I can expect you to tell me.

So, that's my questions and thoughts. Once I have that equipment type, I'll be able to fill in some gaps. If you were without heat, I would be telling you to call a pro, but at this point, you are only uncomfortable for an hour or so each night?

Are those two plots from sucessive nights?

[Edited by BamaCracker on 12-14-2001 at 08:18 PM]
__________________
There are no stupid questions. . .just a lot of inquisitive idiots.

quote:

---

From what you have said so far, I assume this to be a package unit (one single piece of equipment, containing the entire system except the thermostat) is this correct?

---

Yes, that's right.

quote:

---

You mentioned climbing up on the roof to find a model or serial number. Is this package unit mounted on the roof? If yes, how does the air enter the house?

---

Yes, it is a single unit on the roof. The return air goes up from the hallway and the ducts are descending down from the unit.

quote:

---

It's 30 years old? Pretty dadblamed old for a compressor bearing anything.

---

I've got the home warranty that is taking care of it. The air handler motor died last winter and was replaced, along with the condenser fan motor (the blades were unbalanced and they couldn't take them off, so they've just replaced the whole module). My guess is that it's been patched over a thousand times.

quote:

---

Have you begun any of the interfacing to automate your damper system?

---

Yes and no. I've solved the problem of controlling the dampers (though this subsystem is inactive yet), but didn't implement the A/C controller itself yet.

quote:

---

I wondered because I noticed some rooms track different than others.

---

Well, that's exactly the reason I started to work on the controller in the first place. See, the thermostat is installed in the room that is represented by the red line, and the sensor doesn't care about the temperature in the other rooms, and if it so happens that those rooms heat (or cool) faster (or slower), the difference appears. The lower the ambient temperature, the more is the offset between the red line and the others. The offset can reach 10F.

quote:

---

You use a setback thermostat I take it? Or do you just turn the heat down manually every morning when you get dressed? Or is that permanent drop a symptom of the failure? If you do have a setback t-stat, what kind?

---

Yes, it is a setback thermostat. Honeywell Chronotherm III. By specs, pretty smart, in reality, not quite so.

quote:

---

I am curious about whether you have supplemental heat or not, but that is not something I can expect you to tell me.

---

Since I don't know the exact answer, I'll try to put it this way: the roof mounted unit is the only source of heat, not counting the house's internal heat sources (fridge, TV, etc.)

quote:

---

So, that's my questions and thoughts. Once I have that equipment type, I'll be able to fill in some gaps. If you were without heat, I would be telling you to call a pro, but at this point, you are only uncomfortable for an hour or so each night?

---

I'm not even uncomfortable, I don't even notice it until I look at the graphs. However, since I've invested quite a bit of time into the understanding of how this stuff works, and it doesn't work as expected, my technical nature rebels

quote:

---

Are those two plots from sucessive nights?

---

Yes, that's right. On all the previous nights it was just cycling normally.

---

__________________
--vt

What has changed between the last night it cycled normally and the first night it cycled abnormally? Anything. Wife out of town, replaced thermostat batteries, took out the trash, absolutely anything, no matter how insignificant?

I ask because I notice on the graphs two correlations. One is "sort of" temperature related. The other, stronger correlation, is time related. The onset and duration are much more closely associated than the temperature relationship.

In a thermostat like the Chronotherm (about as smart as a stat gets, which ain't saying a whole lot) you have four possible time periods per day. They absolutely, positively HAVE to begin and end chronologically. If you established a time period for "away," then decided to delete it, you may have an overlap in the time, which sends the whole thing into a spin. It can be very hard to find sometimes. I went to a church about 25 times before I ever figured out why it wouldn't work. The times overlapped. But, you can get the same result from missing time periods. This diagnosis seems to have some validity from the standpoint that the end of the second anomoly is the begining of setback.

These are just thoughts in the process, and more data acquisition. I hope I'm not wearing you out with all these questions.

By the way, Guy is an expert in home automation, and he may have some pointers for you about what you are attempting.
__________________
There are no stupid questions. . .just a lot of inquisitive idiots.

BamaCracker:

OK, being completely paranoid:

On the first graph (temp-cut) we woke up at 8AM, and on the second (temp-cut2) at 7:40AM. In both cases immediately after that the dampers were readjusted, which explains the temperature change.

The only relevant fact that I can think of that explains the difference before that time for the blue graph is that I was fiddling with the damper, and on the first day it wasn't completely sealed (note the blue one is flat on the second graph). In other words, the A/C was more loaded on the first graph, that might explain why it went into the "soft failure" on the second.

The thermostat is set to hold the temperature at 69F (20.5C) - you can see that the red one floats around that value.

As for wearing me out with questions, it's me who has to apologize for dumping all this stuff onto your head

I do understand the importance of details and do appreciate your questions.

Speaking about details, I have recently fixed a software bug that depended on 10ms delay, and it's been undetected for a good four years before showing up once and disappearing again...

Keep asking, please. Thank you for your patience.
__________________
--vt

Looking at those graphs, it looks like it's going into a short setback period and then recovers 20 minutes later. Is there a way to definitely clear the program of the thermostat? Also, I would think that if it goes into supplementary heat, that heat source would usually be more powerful and result in shorter, not longer cycles. Those temperature swings also look quite large...
__________________
If you don't know what you're doing, do it neatly

Guy:

Nope, there can't be a setback period, unless the thermostat is malfunctioning. It's been set to hold 69F, and we adjust it a degree-two once in a while, but not on those graphs.

Take a look at the earlier graphs, you'll see the absolutely flat (waving around the setpoint) red line:

http://freehold.crocodile.org:8080/image_archive/

As for the swings, yes, I almost fell off the chair the first time I saw them. The sensors I use are in TO-92 cases, they're exposed (literally sticking out of the wall on the twisted pair) and the temperature inertia is nonexistent. Putting the sensor in a plastic box reduces the swings and should definitely be done later, but I can achieve the same result by applying the filter to the collected data, and I'd rather work with the raw data.
__________________
--vt

Bama,

What do you think???? 30 years old, Weathertron?????? That's my thought. Been following this thread. Not that this is the problem, but what happens on a Trane or older Weathertron if the T wire is left off or the Thermistor goes bad??? Must have installed over a thousand of them in the 70's and never came across that problem that I know of. Does it act like a whacko anticipator???????
__________________
Dumb sounding questions are better than Dumb mistakes

That was one of my thoughts Rayr, but the Chronotherm has no place to hook up the remote thermistor, it's heat anticipation is on board. I woke up this morning thinking it had to be a fluke in the set back, but VT says it's on HOLD TEMP, so there should not be a problem there.

I also thought about the automation system, but he doesn't have that connected yet.

It's beginning to baffle me, expecially since the malfunction seems more related to time than temperature. Once we find out what type it is, we'll be able to tell more about it. Or at least, guess more accurately.
__________________
There are no stupid questions. . .just a lot of inquisitive idiots.

Don't want to step on any toes here, but it seems to me the main thing that has changed here is the ductwork, and dampers, it seems to me that the problem of short cycling might lie in airflow problems or differences from the norm. From what I see, the system drifts from setpoint, Due to the thermostat probably getting alot of air in that zone. Is there a possibility there is too much airflow in the thermostat's zone now that the ductwork modifications have been made? Just a thought.

One of the things I was trying to determine was what changes had been made since the unit cycled normally. That was a thought I had, not specifically about too much air, but about changes in distribution, time, or setpoints. In looking at the old plots he posted, I don't see that 2� increase in setpoint I see in the new ones, right after the second anomoly.

And, even if that does explain the drifting about setpoint, it still doesn't solve that 20-40 minute period during the coldest morning hours when the unit just goes off completely. I'm stumped by that one. I can't shake the feeling we have another problem here that we're either not seeing, or don't have the right information to diagnosis.

Hey, VT; do us a favor, set the fan switch to on, so that the indoor bloower runs continous (independent of heat or cool) and then post the temp plot. If that's not too much trouble.
__________________
There are no stupid questions. . .just a lot of inquisitive idiots.

OK, here goes:

The A/C manufacturer is Goettl Brothers Metal Products, Phoenix, AZ.

The model is GAP 4-1B P, or something like that - it's completely faded.

S/N is 3438.

Now, other things.

tgrams:

There were no changes to the ductwork, and I should've been more exact, I've modified the registers, not the dampers. So far, I just open and close them manually when it is necessary. However, I didn't move them in neither of those periods in question.

Drift from the setpoint in any lines other than red is explained by the fact that the red one is the one where the thermostat is.

Drift of the apparent setpoint for the red line is (as I explain it) the artefact of the way this particular thermostat handles switching the A/C on and off - it seems to me that if they use the PID controller, as they claim, it's poorly tuned.

According to my reconstruction, it goes like this: the more the difference between the setpoint and the ambient, the longer it takes for the A/C to bring the temperature to the proper level before the thermostat shuts the A/C off. Since it seems that the thermostat shutoff is [almost] purely hysteresis, what it effectively means is the room temperature is at the level very close to the shutoff point, which makes it colder than necessary on hot days and warmer than necessary on the cold days. Good example is here:

http://freehold.crocodile.org:8080/image_archive/2001-09-19/

You can see the red line hanging down like a loose rope between 10AM and 10PM (22:00).

I'm not sure I'm right, would appreciate any other ideas.

BamaCracker:

It still seems to me that the anomalies were temperature related, not time related. Check this out, this is last night:

http://freehold.crocodile.org/temp-cut3.gif

At 12:30AM the setpoint was moved from 71F to 69F, and the ambient temperature didn't fall below 35F (unlike 31F the previous two nights) - the graph is completely normal.

Looks like something is freezing up there, but what is it?

Is there a theoretical limitation for a heat pump that affects its operation when the difference between ambient and output becomes more than X degrees?

Do I understand it right that there's no feedback from the rooftop unit back to thermostat?
__________________
--vt

Bama,ever hear of Honeywell T8200 set back thermostat???????
It was Honeywells worst nightmare. Without any exageration I changed hundreds of them. Honeywell extended the warranty to
24 months. I don't know what the cause was but I know that the heat would come on in July or Aug. when the A/c was set for Cooling and vice versa, Cooling would come on in Dec. when in Heat mode. They gave T8600 as a replacement. This was in the 80's. Just curious if you were around for that fiasco.
__________________
Dumb sounding questions are better than Dumb mistakes

On that unit, there is no feedback. The G.E. unit we were discussing was/is (Trane now) the only one that supplies feedback, and not the kind of feedback that would turn off your unit.

The definitive answer for the cause will come when the temp is below freezing at some time other than early AM.

However, there is nothing in a heatpump to freeze as long as it is not a water source unit.

Supplemental heat: because a heatpump is sized for cooling load, very few parts of the country can get by with just the heat produced by the ehatpump. The reason is two fold; first, in most areas, the delta T for cooling is 1/2 the delta T for heat. If I remember the numbers for Arizona (I designed a system for a friend who lives in Phoenix) cooling is designed for 100�OD and 75ID, and heating is designed for 20 OD and 70ID. Is that right somebody? So, you're designing for a 25� difference between indoor and outdoor in cool, but a 50� difference in heat. So, your heatpump needs some help. Interior load can handle some of it, people, lights etc. but we can't plan for that can we? So, we take the capacity of the heatpump at low temps (the other fold) subtract that capacity from the heat load, and apply supplemental heat of some sort to make up the difference. Supplemental heat can be from any source, but is usually; electric strips, gas furnace or oil burner. To expound on that second fold; the lower the outdoor temp, the less effective your heatpump is. Not less efficient, less effective. So, your unit (which I am assuming to be 4 tons) is probably about 2.5 tons at 17�F down to about 0�F.

So, if you do not have an oil storage tank, or a natural gas bill, you probably have electric strip heat.

I suspect that old unit has little to no electronics.

Did you install the Chronotherm? Or was it already there? Do you think the Chronotherm is the original stat?

Do me another favor. . .scroll through the program on that stat, and see if any of the times align with any of our anomolies. I know you have it set for HOLD TEMP, but it just seems to likely that the stat is at fault.

Rayr, I first got into HVAC in P-cola Florida in (about) 1986. The company I worked for was anti-setback stats, and I never installed or worked on one till the early 90s. So, I am fortunate to have missed that debacle.
__________________
There are no stupid questions. . .just a lot of inquisitive idiots.

We called it "The Thermostat From Hell"
__________________
Dumb sounding questions are better than Dumb mistakes

BamaCracker:

no storage tank, no oil/gas.

The Chronotherm was already there, however, it was not an original controller - the original one is right there next to it on the wall, masked with the Scotch tape Has one switch with the following markings on it:

Off, Low Vent, High Vent, Low Cool, High Cool, Pump Only.

I suppose it is inactive now, though it's still connected to some wires that go into the conduit; I didn't investigate further.

I doubt there are electric strips - this is my second winter in the house, and I guess I'd notice if there were any.

The program on the stat has borders at 6AM, 10AM, 6PM, 12AM, no correlation.

As for heat pump vs. cooling design, here's one fact that I can't find the explanation for: same house, same damper settings, the system cycles at least twice as often in heating mode than it does in cooling mode. Why?

Yes, the unit is 4 tons.

On a side note, I guess I'll be looking for professional help - when I was on the roof, I found out that the insulation from the air handler exhaust is almost gone, and the air is coming right out, heating Arizona instead of my house So if someone from Phoenix Metro is reading this, drop me a note.
__________________
--vt