Proportional vs. Digital Control (graphs)

Mike T., Swampeast MO

These are from my office.

"Proportional" graph is an oversized iron radiator with TRV; constant circulation; mechanical reset and simple standing-pilot boiler.

"Digital" is electric resistance forced air. I use this for "shoulder" seasons. We recently had a warm spell and today was the first time the forced air system had to cycle since I got this measuring/logging equipment.

Same period of time for each--same range in the graph.

Even in the two-week graph I usually view I noticed this afternoon that I was getting lots of squiggles that I'd never noticed when the radiator was heating the space.

Have had this system since late December (actually began logging at the solstice) and have found it quite interesting that when viewed in a chunk of a few weeks or even a couple of months that the graph was amazingly sine-wave like with the peaks very frequently occurring with amazing regularity BUT seeming to progress slightly later each day... The proportional graph was my best idea of a representitive sample. Some have a bit more variance, others (when unoccupied) have MUCH less--around 1° daily.

I find this EXTREMELY interesting.

Objective notes:

I have three thermometers in this space.

One is a digital thermometer/clock on a partition wall by the operating side of a frequently used door. The temp on the other side of that partition is ALWAYS lower--usually about 8°. It is essentially "invisible" to the radiator.

The 2nd is the sensor for this data unit. It is at exactly the same height as the other but located on another partition in full view of the radiator but AGAIN, the temp on the other side is lower--generally about 6°.

The 3rd is a Galileo (the floating balls) thermometer with 2° temp resolution.

I have verified that the clock/thermometer is within a couple 10ths of a degree of accuracy with the datalogging by putting them extremely close together. Likewise the Galileo is quite accurate given its resolution. It is on top of my desk hutch about 1' higher than the other two thermometers.

With the radiator heating the space, the clock/thermometer is always lower than the data-logging--usually about 0.8 - 1.2°. The Galileo always agrees within its accuracy. Generally I run the ceiling fan at lowest speed 24-7 but this year the outdoor air has been unusually dry and the air movement seems to have bothered me much of this heating season...

With the digital forced air however, I notice that the clock/thermometer is reading from 2-4° lower than the datalogger and that the Galileo is consistently 2° higher. The ceiling fan was RUNNING at low speed.

Subjective notes:

Notice that the general temperature of the two is different. In years past I have had the TRV a touch LOWER and WITH the ceiling fan running. I would consider my comfort level in these two situations similar BUT I HAD to turn on the fan with the forced air running lest my head feel too warm.

The day warmed and the heat hasn't kicked on in quite a while--you can see the beginning of this trend at the end of the graph. Space temp is 71.9 ° at datalogger; 70.4° at clock/thermostat and 74° at Galilelo. It feels a touch cool--something I rarely if ever notice at this time of day when the radiator is heating the space.

--------------------------------

Notice how the digitally controlled graph seems to have ZERO to do with what is going on outside at the time? Compare that to the proportional graph. Add proportionality, use a decent amount of radiation and let the "automatic" radiative control do its job.

R. Kalia

Two comments:

a) while the radiator heat is more even over short time periods, it has a wide swing over a long period (e.g. the entire period of your graph). That's the disadvantage of this high-mass heating method.

b) If you are using electricity to heat the air, there is no reason for it to be on/off; it is very easy to do proportional (or its equivalent, very rapid on/off) with electric heat. The air would then blow all the time (like constant water circulation) but it would be warmed to varying degrees based on the thermostat.

I have fond memories of our former (nat gas forced air) house, where we could come home in the middle of a setback period and just turn the thermostat up and it would be warm in 10 minutes. Hot water takes some getting used to, as I am sure driving a Mack truck does too.

Mike T., Swampeast MO

I intentionally chose a similar temp variance

and a similar "type" of day--workday with cool morning, sunny day with light wind. Also notice that the temp variance is still within ±1.5° on both with the proportional a bit lower.

Attached is graph of the same space when unoccupied for two days. Note that temp variance is extremely low. Lo/hi as logged 67.1° and 68.7° or ±0.8°

Also it's quite important to note that the spaces surrounding this area are intentionally and significantly cooler at all times.


The Galileo thermometer is probably in the best place (and also of the best type) to read the "true" air temperature in the space. I've noticed the balls moving FREQUENTLY today--only VERY rarely with the radiator doing the heating. The clock/thermometer is in the best place to be influenced by the temperature outside the space and the activity caused by opening/closing doors. The logging sensor is in the best place to be influenced by the radiative output of the radiator. NONE are in a position to be influenced by the winter sun--and intentionally poorly on the first "bouce" as well.

What I [seem] to be observing with a radiant source under proportional control is that the space temperature moves in concert with both the solar and occupancy cycles with HIGH regularity while maintaining the same level of human comfort. (My hours are highly irregular and I can either work extremely late or start extremely early--I sincerely know the space very well.)

A similar period under digital convective control [seems] to loose connection to both the solar cycle and the occupancy BUT requiring a higher air temperature and still (despite intentional air movement to "even" the temperature) less CONSTANT comfort.

---------------------

I suppose it exists somewhere but I've never seen the type of electic system you mention. I'd hate to imagine the kind of electrical "noise" generated by switching a few KW of power in the manner of a "dimmer", and while it would seem to be easy to vary the fan speed with the resistance output, the type of thermostat required to provide this control information...certainly not a typical residential device.

R. Kalia

I've seen PID thermostats

> I suppose it

> exists somewhere but I've never seen the type of

> electic system you mention. I'd hate to imagine

> the kind of electrical "noise" generated by

> switching a few KW of power in the manner of a

> "dimmer"


I'll have to look for it, but I have seen PID thermostats on the web for residential electric heat use. The trick to avoiding noise is to do it slowly; unlike a light dimmer, you don't need to switch it many times a second, but only once every few seconds, so the noise frequencies are very low.