seismo:

Guessing the peak consumption is as good as just reading the power rating off the manual and using it to calculate the energy consumption. Of course, it will give you *some* value - like, you would be able to guestimate the consumption based on number of starts as well as running time, but this can be used as a guide only - every start will be in fact different.

However, logging this data for further processing does really allows you to diagnose the problems that will not manifest themselves otherwise, you are right here.

Me, I'm just a log freak - logging everything The more data you have, the more conclusions you can draw, even though the value may not be obvious at the time. Plus, sometimes when you try to visualize the data you think you know the nature of, you become really surprised - like, I'd never guess that the temperature hysteresis for a cycle can easily reach 5-8F, or that the temperature difference in different rooms can reach 14F if the house is not balanced right, or that your unit starts to behave funky when the ambient temperature drops below the freezing point - you're sleeping at that time anyway and wouldn't otherwise notice.

__________________
--vt

seismo and vt, the data logging is done using a board sold in kit form at:

http://www.phanderson.com/t64.html

I then wrote a VB program to continually poll 7 of the digital I/O ports and when a change occurs, I log it to a file. I also log 4 different temperatures every half hour. Here is an actual sample from the log file (temps are in Celsius degrees):

Mon 04 Feb 02 10:28:19 : IN6 ON , Aux Elec Ht ON
Mon 04 Feb 02 10:30:02 : AT-TEMP >T000 18.31 Attic Temperature
Mon 04 Feb 02 10:30:03 : OU-TEMP >T001 -8.54 Outside Temperature
Mon 04 Feb 02 10:30:04 : PO-TEMP >T002 -7.31 Pool Temperature
Mon 04 Feb 02 10:35:00 : IN6 OFF, Aux Elec Ht OFF
Mon 04 Feb 02 11:00:02 : AT-TEMP >T000 18.31 Attic Temperature
Mon 04 Feb 02 11:00:03 : OU-TEMP >T001 -7.31 Outside Temperature
Mon 04 Feb 02 11:00:05 : PO-TEMP >T002 -2.75 Pool Temperature
Mon 04 Feb 02 11:30:01 : AT-TEMP >T000 18.31 Attic Temperature
Mon 04 Feb 02 11:30:03 : OU-TEMP >T001 -6.03 Outside Temperature
Mon 04 Feb 02 11:30:04 : PO-TEMP >T002 -2.25 Pool Temperature
Mon 04 Feb 02 11:54:02 : IN1 ON , Rev. Valve ON
Mon 04 Feb 02 11:54:03 : IN3 ON , Defrost ON
Mon 04 Feb 02 11:57:26 : IN2 OFF, Blower OFF
Mon 04 Feb 02 11:57:50 : IN3 OFF, Defrost OFF
Mon 04 Feb 02 11:57:53 : IN0 OFF, Compressor OFF
Mon 04 Feb 02 11:57:54 : IN1 OFF, Rev. Valve OFF
Mon 04 Feb 02 12:00:02 : AT-TEMP >T000 18.31 Attic Temperature
Mon 04 Feb 02 12:00:03 : OU-TEMP >T001 -6.25 Outside Temperature
Mon 04 Feb 02 12:00:04 : PO-TEMP >T002 -2.51 Pool Temperature
Mon 04 Feb 02 12:07:12 : IN0 ON , Compressor ON
Mon 04 Feb 02 12:07:13 : IN2 ON , Blower ON
Mon 04 Feb 02 12:11:18 : IN2 OFF, Blower OFF
Mon 04 Feb 02 12:11:22 : IN0 OFF, Compressor OFF


I find this useful to monitor any changing trends that may indicate a problem, like abnormally long defrosts, etc. I have about a year's worth of data so far. The latest temperatures are also saved in another file that I can then consult over the phone to know what they are (kind of nice to be at work on a hot summer day and call to find what the pool temperature is). As you can see, the logging activity is not related as such to the Ocelot or other home automation controllers, it's just another "service" running on the server. If the server were to crash for any reason (and it never has), nothing would be impacted in the home's operation. If one of you ever wants to make a logger from the same kit, I'd be more than happy to give you my VB program.
__________________
If you don't know what you're doing, do it neatly

Dependableairservice, tgrams:

Here's the diagram. It's a quick and dirty sketch, but the essence is captured:



The I/O is currently RS-232, but quite possibly could be any of the appropriate controllers that are able to run the thermostat core.

The system can be controlled either from local or remote computer, or from a conventional wall controller.
__________________
--vt

vt-

Wow! There are others out there who are thinking along the same lines. I'm currently ironing out my temperature sensing capabilities with dal-semi's i-button stuff. I just want to monitor how bad the problem is before I attempt to do anything to fix it.

I couldn't run my furnace off of a Microsoft OS, my thermostat does not have 3 buttons to hit control-alt-delete with.

Congrats - you are quite a-ways ahead of me actually controlling the air distribution via the pc.

I'm encouraged to see that the professionals here have not flamed you on your approach, just given you some things to consider and asking interested questions.


Cheers

Jim

Guy:

I'm using a different approach to logging. I prefer the data visualized, and though it may seem less flexible (think adding more data sources), it's sufficient from the practical standpoint.

This is how the raw data file looks like:

1012883318:22.75:23.9375:22.9375:24.375:22.25:11.5
1012883352:22.8125:23.9375:22.9375:24.375:22.1875:11.5625
1012883385:22.8125:23.9375:22.9375:24.375:22.1875:11.5
1012883421:22.8125:23.9375:22.875:24.375:22.1875:11.5625
1012883454:22.8125:23.9375:22.9375:24.4375:22.1875:11.375
1012883500:22.8125:23.875:22.875:24.375:22.125:11.375
1012883534:22.75:23.875:22.875:24.3125:22.125:11.1875
1012883564:22.75:23.875:22.8125:24.3125:22.125:11.0
1012883594:22.8125:23.875:22.875:24.3125:22.125:11.0625
1012883628:22.8125:23.875:22.875:24.375:22.125:10.9375

Then, it's fed to rrdtool (I keep the raw data anyway, the size of all the temperature logs that were mentioned here is about 20 megs since last August - peanuts), which handles storage, retrieval and graphing quite efficiently.

One more difference - I take the readings much more often. Currently, a sweep on the house takes about 6 seconds (non-optimized; it's probably possible to make it in 3) and once in 30 seconds the data is stored. by the way, 6 seconds delay on the temperature reading, coupled with the temperature sensor with non-existing inertia, is quite insufficient - the overshoot may quite possibly get to 0.5C.

__________________
--vt

jtm065:

Side note, I'm actually happy this stuff develops the way it does. First, there was a simulation model, then I got hold of the sensors, then installed them, made sure that what I was feeling was indeed happening, then collected some stats, updated the model, put together the servo stuff, installed it, now waiting for the final components to arrive.

How to eat the elephant?
One bite at a time.

That's what I'm doing.

As a teaser, here's one of the early model run results:



No latent heat in this model, but still proves the point. Ambient temperature is real (September 10 2000), the rest is simulated. The point was to make sure that the house with clearly insufficient A/C is still possible to cool down, if you get the airflow right. As you can see, it is possible, though this model is not optimal - it is much better now

Legend:

- horizontal lines on the top - setpoints, different for every room
- curves on top - simulated temperature in the rooms
- curves on bottom - controller output, ignore this
- rectangles on bottom appear when the registers in the corresponding rooms are open

btw, looking at this graph, the optimizations to the model seem obvious to me, I'd be curious to see what you guys can think of so we can compare

__________________
--vt

Fascinating....quite an excellent temperature control in each zone. I do have a couple of questions though: I notice by your chart that there are times when only one zone damper is calling, and the equipment is running in cooling. Have you installed some kind of bypass damper in the ductwork to decrease static pressure in the supply duct?(therefore cutting the possibility of indoor coil icing) Balancing airflow at the registers is a novel approach at zoning, most zone systems are zoned at the trunkline with dampers there, cuts down on velocity loss, if you could use the same control panel, and utilize dampers closer to the main trunkline the temperature swings would be less in my opinion, and you would get higher velocity out of the open ducts.One thing you could do is what carrier used to do, if one and only one zone is calling, open a zone damper in annother zone to dump some air...similar to a bypass damper. It would also be better if you had variable speed blower, you would have to interface your control system with the control board in the blower but it is possible...Hell, you could probably pull out the board in the blower and build yourself one to do what you want. It's very interesting work you're doing though, keep us informed.

tgrams:

Again, just to make it absolutely clear, that was not a real thing, that was a model. The only real thing there was an ambient temperature.

The static pressure is controlled by partially opening the registers - this is a cost cutting measure. Also, allows to balance the temperature better - less swings.

Registers and dampers are interchangeable from the model viewpoint, in real life, again, this is a cost cutting measure - cheaper to manufacture, cheaper to install. Velocity loss - thanks for mentioning that, I didn't think about it.

Variable speed blower (and multistage compressor) - yes, this system just begs for it

Pulling the board out - nope, can't do. There's FUD factor - fear, uncertainty, doubt - kills me. Hate to destroy the hardware. Back in old good times, had a chance to do a lot of that for free, not anymore

__________________
--vt

I totally agree with the FUD factor. I will *not* try to modify any existing system unless I either have a schematic diagram, or have figured it out completely by tracing the circuit on my own. I did not proceed with my electric aux heat addition until I had the schematic of the dual fuel relay setup and even then, I studied it for a couple of days before touching anything. There was an intriguing "plenum thermostat" in series with the compressor run signal, and I fnally figured out that it was there to protect the compressor from running at the same time (normally not possible due to relays) or right after the oil furnace , which would have caused the evaporator to be in a hot airflow and send back liquid to the compressor (which as I've learned here in this board, is a bad thing).

Tyeing into an existing control board can have it's rewards however, I've done that with some industrial sewing equipment to get a computer into the action, to automate certain functions. A variable speed fan would certainly be a prime canditate for such a project, if you need to alter airflow speed for your own purposes. But I imagine that schematics for these might be hard to obtain.
__________________
If you don't know what you're doing, do it neatly

By the way, speaking of variable speed fans, how do you control them? Thanks to you guys, I know how to control the conventional A/C unit, doesn't the variable speed unit have some sort of a standard interface?
__________________
--vt

VT,

So this was your thread. I like what you have done so far.

The information I have found out so far is that the variable speed units control their fans in two ways. First the initial setups for each stage for heat and cool plus continous fan along with any offsets for the HUM input.

The next thing they do is control the blower so the CFM will stay at that setting even with changes in back pressure.

All you have to do is feed W1, W2, Y1 and Y2. The Carrier 58MVP does not have a Y1 but uses the fan setting for the low cool setting. That's the one thing I did not like about the MVP. I wanted a slower CF setting the the low cool setting.

Most of what I have found is from the manuals plue GE's web site. Most of the units seem to be using the GE motors.

quote:

---

Originally posted by vt
By the way, speaking of variable speed fans, how do you control them? Thanks to you guys, I know how to control the conventional A/C unit, doesn't the variable speed unit have some sort of a standard interface?

---

n6ber:

What's the HUM input?

What exactly do you feed to W1, W2, Y1 and Y2?

Somehow I have impression that you're talking about a multistage setup, not variable speed (though Carrier 58MVP description clearly states that it's a variable speed). What am I missing? My initial assumption was that the variable speed can be controlled in a stepless manner from 0 to full speed, but unless there's a complex signal (like PWM, for example) at W1,2/Y1,2, I don't see a way to control it like the name "variable speed" implies. Unless it's the HUM input - see above.

Do you, by chance, still have any links?

Too bad I don't have a variable speed unit yet and have nothing to stick the oscilloscope into...
__________________
--vt

The following discussion is all theoretical...

When I see discussions about variable speed blowers (although I have never seen one myself) and think in terms of process control, I picture a device which has a desired speed setting which would probably be in CFM, and a feedback mechanism (airvane ?) to tell the control logic how fast it is actually going so that it can then compensate the power being delivered to the motor to maintain the speed (if this sounds like the cruise control on a car, you're in the right frame of mind). To tap into this for you own purposes of introducing zoning, you would want to be adjusting the desired speed setting (thus CFMs) to match the needed air for the number of zones you want to heat or cool at that moment. The ideal would be that the desired speed setting would be set by DIP switches in a binary sequence, or a control voltage input to an op-amp type circuit with a potentiometer. These would be the easiest types of controls to tap into with a microcontroller. If the variable speed controller is set by a "cruder" method, like a jumper that can be plugged onto one of 5 or 6 pins and that these settings provide a wide enough range of speeds for your purposes, then you could get away (assuming you are using an Ocelot for this example) with using something like a SECU-16 I/O module to select the speed you want using the Ocelot's logic (example: 1 or 2 zones, speed 1; 3 to 5 zones, speed 2, more then five zones, speed 3). It's very easy to do in ladder logic. If you have the jumper type setting but want finer control, then this is where it gets complicated, because you need to get beyond the jumper's settings and want to inject the signal where it gets analog. A schematic diagram of the variable speed controller would be if immense help here.
__________________
If you don't know what you're doing, do it neatly

In HVAC they don't give us wiring diagrams for controllers and boards, they just give us "flow charts" to diagnose them with. If a 4� resistor fails, we have to buy an entire new $200 board.

I can ship you a VS module, but here's the basis for how it works:

The motor itself is basically a DC brushless motor. (BTW, the 'official' term for VS motors is "Electronically commutated" which may mean more to you than it does to me) I have been to several classes put on about these motors, and the idea is explained thusly; the drive (which is built onto the bottom of the motor in most cases) steps up the A/C voltage, runs it through a doubler circuit (just like a microwave) and spits out a pure DC signal of approximately V x SQRT3 relative to the input voltage. An inverter then supplies the commutated signal at varying voltage and frequency to accomplish the desired speed.

The actual sped is measured by the shaft rotation and torque, through a PID feedback, and the drive (not to be confused with the controller) adjusts the voltage and frequency comiserate with that feedback loop.

So, all the stuff you guys are asking about is controlled "INSIDE" the motor and built-on drive. The controller is just a simple IO board which determines which speed is being called for based on the input from the DIP switches.

Also, pertinent to this discussion, the VS technology only allows for ramping from 25% of maximum to 105% of maximum, with 50% being the start up speed. Any input less than 50% will not overcome the motor's internal resistance to movement.

If you are interested in the limited knowledge to be gained from manufacturer's books and maybe even a dead drive, my e-mail is in my profile.
__________________
There are no stupid questions. . .just a lot of inquisitive idiots.

I was interrupted in my last post and forgot to finish the discussion about introducing your own speed setting...

If you want to make life more interesting and daring, you could also tap into the feedback of the controller instead of the setpoint. This would involve a deliberate introduction of positive feedback to get the reverse effect (ie: make the controller "believe" the blower is going too fast by sending back a voltage that indicates this, and it will react by actually slowing down). But doing this means that you will theoretically never hit the setpoint, but make it "hunt" all the time. Plus you're at the mercy of the controller's integral or diffential (gain) settings. I would avoid even thinking about this method unless there really is no other way, but it is technically doable.
__________________
If you don't know what you're doing, do it neatly

It seems like Bama and I posted at the same time...

What you say is interesting though. Just to see how it is designed, could you show me a list of the speed settings and the corresponding DIP switch positions? This would show if the input appears to be digital. Also, how does the "PID feedback" work and/or connect? Is this accessible from the outside? When these motors die, is it the controller or the drive that peters out?
__________________
If you don't know what you're doing, do it neatly

First here is a link to the ECM that most of the furnaces are using.

http://www.geindustrial.com/cwc/products?pnlid=4&id=ecm

W1/W2 are low heat/high heat and Y1/Y2 are low/high cool. The HUM input is for humidastat. When set true the fan speed is dropped by a think 10 to 15%. This way more water is pulled out of the air.

Now back to zoning.

Maybe it's over kill but it seems to me that if you could get total control over the dampers you could get much better control over the zoning. I've not had a chance to read all of you posts but in my research I have found that very few of the zoning system do any more than bang-bang the dampers open and closed. Carriers is the only one so far that I have found that does a fair job controlling the dampers.

Now in my way of thinking it would be great if you could control the CFM through each damper. The only way to do that is to measure it at the damper which means a flow sensor at each one. Then the zone controller could tell the damper to pass 300 CFM and the damper would then adjust it self to hold the flow to 300 CFM. As other dampers are commanded to change the pressure changes would cause the other damper maintan the required flow. If the dampers could not adjust it would notify the zone controller which would then adjust the staging up or down as required.

Building a damper like that is not hard. A "Stamp" microprocessor could do the job at very low cost. The whole unit would fit on the damper and you could feed it with a 3 wire cable, power and signal.

I may try this when I get my sytems up which is why I will most likely use the Carrier dampers. They are the best built one that I have found. Now if I could only find who Carrier gets them from if they don't make them.

Again maybe this is overkill but it's fun to think about.

quote:

---

Originally posted by vt
n6ber:

What's the HUM input?

What exactly do you feed to W1, W2, Y1 and Y2?

Somehow I have impression that you're talking about a multistage setup, not variable speed (though Carrier 58MVP description clearly states that it's a variable speed). What am I missing? My initial assumption was that the variable speed can be controlled in a stepless manner from 0 to full speed, but unless there's a complex signal (like PWM, for example) at W1,2/Y1,2, I don't see a way to control it like the name "variable speed" implies. Unless it's the HUM input - see above.

Do you, by chance, still have any links?

Too bad I don't have a variable speed unit yet and have nothing to stick the oscilloscope into...

---

Usually, the drive succumbs to a voltage spike. The fact that it's built into the motor limits the size of the choke you can put in it. The last time I was actively involved in the workings of Trane, they were about to begin a program where you could buy just the drive.

The PID is only a theoretical entity as far as I know. It's in the drive, which mounts to the back of the motor. Bunch of plastic coated elctronic parts is all I can tell you. I learned about PID in controls class, and I know the accuracy with which the VS motors respond to changes is only available in a PID control. It's an awesome technology.

The "user interface" for lack of a better word, is usually a block of dip switches, or some wires with little pins to stick them on. You have frequently read in this forum about proper setup of a VS system? Well, it's a matter of choosing the correct airflow for the application. Example: you have 11 pins and four wires. White wire is for the lowest speed you ever want the fan to operate at. This is usually 40-60% of cooling speed, which will be the highest speed. Let's say you have a TWE060. That's a Trane model number for a 5 ton fancoil unit. It has the capability of moving 5 x 450 CFM, or about 2250cfm. Now, you connect that to a four ton unit, to give a 12 or 13 SEER. But, that 4 ton unit never, ever needs more than 1600 CFM. So, we set our cooling speed wire for the pin representing 1600 CFM, set the "ventilation pin to 640 or 500CFM, and the heating pins to a comfortable setting for heatpump, and a required setting based on the amount of supplemental heat.

Now, add the "Comfort" setting. This is usually done in a series of dip switch settings, which are actually on the controller board, (which controls very little, and has nothing to do with your concept of a PID controller in a block diagram.) This setting allows the unit to maintain a certain outlet air temperature. As outdoor temps drop, the heat pump moves less heat into the house, and the fan slows down to maintain that nice, warm toasty feeling air.

I hope this helps some, but I'm gonna have to pass out now. I'm dieing here from lack of sleep.
__________________
There are no stupid questions. . .just a lot of inquisitive idiots.

VT, check out http://www.dayus.com as they have grills just for passing air through doors and walls.

quote:

---

Originally posted by vt
Diceman:

Good point However, the sound isolation in the most of the houses is so bad already, it doesn't really matter. As for visibility, the grills should take care of that.

---

BamaCracker:

Thanks for the offer, I'll drop you a mail when I'm ready (not gonna happen sooner than a couple of months, the equipment I need is still backordered)

Sheesh, I guess I'd really have to wait and see. Right now I'm tempted to just rip off the controller board and install a PWM controller, but then again I'll have to be practical when the unit actually arrives

Right now I'm actually pondering about proper matching of the blower speed, compressor stage and total percentage of open dampers. Do I get it right that the ducts are in fact a little bit oversized so the dampers/registers can be closed a little bit to balance the house? What is the tolerable percentage of open airflow that doesn't stress the compressor? 80%? 50%? In my house, I've had 70% of the registers closed at most, 40% at least - doesn't seem to have any negative effect on the unit - been running good (well, kinda) since September 2000.

At the beginning of this thread, Robin Boyd mentioned high temperature differentials due to low air flow - can someone please elaborate on that? I wasn't able to find a reasonable answer to that question.


__________________
--vt

BamaCracker:

OK, I get the picture. In the worst possible case I will have to have a stack of [solid state] relays to achieve a fine-grained control over a fan speed. I will need it, if the model I'm putting together is right.

n6ber:

Right now, the zone controller does indeed bang-bang the registers, however, I've put together another model that controls the register in an analog way, here's the result:



that, of course, is just a first take (PID autotuning would be a real help), but you can see that even with a subpar unit it is possible to achieve a comfortable environment. This one is much smoother than the old one based on hysteresis approach.

I don't analyze the CFM, though. The model is based on the temperature gradients, it is more adaptive and less quality-dependent this way.

__________________
--vt

What Robin was talking about is the airflow effects the temperature rise across the heatpump/ A/C coil. If you are closing damper you are changing that rise which will affect the refrigerant flow which will damage your compressor( which I am certain you already have.) You should put sensors in the return and supply at the unit and watch the rise. Also place sensors on the large and small copper lines leading to and from the indoor coil. Closing the damper raises the pressure on the high side triping the compessor on high pressure. I don't beleive I have ever seen a heatpump without a high pressure switch.

The manufactures want to see 350 To 450 airflow across the coil. I would recommend getting a static pressure guage installed across you coil. Get manuf. data and check airflow. Most systems are already experiencing low flow and you are making it worse. To yours benefit yours is already old and not much to loose.

The is a company in your area that will test the airflow and complete system operation. They will also check the infiltration rate of your home. They were known as Digital Air and should have changed to Enalasys. It will cost but you could be try to fix a problem that cannot be fixed with the approach you are taking.

Guy I would very much like the VB program you have developed. Will it run on the learning addtion.
How hard is it to understand for someone just learning VB.
I need to inteface with the sensors reading as well. Probably with the board you have recommended. I would also like to import data into Excel for analsys, with formulas I have.



[Edited by Lorne on 02-08-2002 at 11:45 PM]

Lorne, like I said, all this program does is send commands repeatedly to the "phanderson" kit and when it finds that an input has changed, logs it to a file. I don't know about a "learning edition". It only uses the MSCOMM device, which I think must be pretty basic. It is in VB5. Let me know if you want it and I'll e-mail it to you. It is written in a ladder logic model, so it may be a bit hard to follow if you're a beginner, but you're welcome.
__________________
If you don't know what you're doing, do it neatly

How small of a change? I will be using it in my duct sytem and also a outdoor sensor. Can I import to Excel. Is there a way I can simulate to try out the program for my needs without purchasing all hardware (probably not just hoping). My email is lornelambert@charter.net

Lorne:

What would be the symptoms of the damaged compressor?

I'm pretty paranoid about the behavior of all the gadgets that I have, and I pay extreme attention to all the telltale signs that I can discover, so far, it doesn't look to me like the compressor is damaged - none of the operating conditions have changed. Cycle is the same, sound is the same, temperature gradients are the same - what else is there?

350 to 450 airflow across the coil: I guess you're talking about CFM?

Since I'm talking about a prototype, I'll probably think about installing an airflow meter and static pressure gauge in my house, however, for the same reason (it's a prototype) it's definitely not going to make it into the final version - it's going to increase the complexity (thus damaging the reliability) and drive up production and installation cost. These things can be figured out analysing the byproducts - temperature gradients in particular. I think. I'll verify this.

__________________
--vt