Entropy

Relax in southern comfort on the east bank of the Mississippi. You're just around the corner from Beale Street and Sun Records. Watch the ducks, throw back a few and tell us what's on your mind.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Entropy

Post by JR. »

Well in our onward march to final equilibrium and decay, my latest repair adventure was my 86yo neighbors electric wheelchair/scooter.
Image
We got about an inch and half rain storm the other day and the mobility scooter while covered in plastic got wet enough to stop working.

It was a bit of an adventure to get apart and troubleshooting codes gave me a bum steer.... The beep alarm (slow beep) suggested problem was with the wig-wag pot and/or speed control... (they call it wig-wag not me, but that the forward back control...). I wasted way too much time looking for the problem with it and traced the wiring all the way back to the processor, buried deep inside the guts.

At the end of the day I fixed it with a q-tip and some rubbing alcohol. Apparently the rain water and some dirt or corrosion products inside created a conductive path on the processor board where it reads the voltage of the wig-wag pot...making it appear bad.

I found a secret speed pot inside that sets the max speed so I turned that up. :D :D

Now my neighbors chair goes faster too...

I was glad I could fix it for him for free... I suspect some repair depot would sell him a new controller board.

JR

PS: It looks like there are some 15-20 different versions of these with the guts all made by the same company in China (Pihsiang). I hate to admit it but the design seems pretty solid and the troubleshooting help while not literally correct for me looks like it would be useful for most common repairs.
Cancel the "cancel culture", do not support mob hatred.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Entropy

Post by JR. »

My electric toothbrush developed a new annoying failure mode. The on/off switch stopped turning off, so the toothbrush was always on, it would stop running when sitting on the charger so I was reduced to putting the toothpaste in my mouth before grabbing the tooth brush, but it was messy when removing it from my mouth while still running, spraying toothpaste around my bathroom. :oops:

I found out they are not that hard to take apart. The top kind of unscrews. Inside mine had 10+ years of wet smutz inside, apparently some conductive smutz shorted the on/off switch.

I cleaned it out with some rubbing alcohol and one of those old fashioned manual tooth brushes. :lol:

Back in business...

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Entropy-electric blanket

Post by JR. »

My latest repair, is not really a repair but a work around. Last spring the on/off switch for my electric blanket stopped latching on. Since the cold weather was about done I was not motivated to fix it then. Now that the temps are dropping down at night it was time to get it working.

I don't even remember how old the electric blanket is, probably 30 years or so... I haven't even used the on/off switch for years, leaving ti turned on all winter and plugged into a timer so my bed is warm automatically and blanket is powered off in the morning.

The design is so old that the molded plastic clamshell is actually screwed together. The power switch was a very old school rocker switch with an wire spring that alternately latches it on or off depending on which end of the plastic button/actuator you press. The failure was that the plastic rocker had literally fatigued and split in half in the middle so when you pressed down the on side, the off side did not go up enough to snap the spring and latch. If i had some super glue or epoxy handy I might have glued the button back together, but since I didn't and only turn this on/off once per season, I decided to hard wire it on (solder the on switch contacts together) and I will unplug it next spring.

A pretty nice design, a bi-metalic switch contact, with a resistor that heats(?) the bi-metallic strip, so it responds to the air temperature in the room and average heat delivered to the blanket. This might be another candidate to automate, or I could just plug it into my microprocessor controlled baseboard heater... I do have a dual outlet on it, so this would be very easy.

For chuckles I could experiment with turning the temp control full up and let my room heater modulate the blanket on/off. I might try that. it would make the blanket a little smarter, but the thermostat works remarkably well for it's simplicity.

JR

[edit] Well I plugged my electric blanket into my smart thermostat and coincidentally the night-time temps warmed up enough that the heat hasn't come on since. The good news is that the electric blanket that would have been on a timer would have started coming on and wouldn't be needed. So far so good. [/edit]

[edit2] After plugging the electric blanket into the smart thermostat I could hear it toggling on and off too quickly (1 second sample rate). So I re-wrote my heat algorithm to slow it down so it can't dither on/off fast enough to be irritating. Last night was first cold night since the update and blanket was comfortable and I didn't hear it working. So all good... I may slow it down even more and add hysteresis but so far it isn't needed. I have a two to one bias making it easier to turn off than on, since there is lag in response. [/edit2]
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Entropy

Post by mediatechnology »

We had a lightning strike last June and I thought I'd repaired all the electronic damage...

That was until I set the thermostat to "heat" last week and found out that the draft/induction blower wasn't running.
That locked out the heater.
Looked like an SMT switching transistor fried.
So I had to buy a new controller board.

When I last used heat - it might have been as late as last April - the heater worked.
So it must have been the lightning in June...
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Entropy-SERVO

Post by JR. »

I mentioned that after my electric blanket on/off switch broke I hard wired it on and plugged it into my micro-controller thermostat. But then the plink plink plink of the blanket being turned on and off by my controller dithering on/off was irritating when trying to get to sleep.

My temperature control algorithm is not a simple on/off and I have experimented with adding hysteresis, sampling lag, measuring actual temp rate of change over time, etc all with varying degrees of success, it always seemed under damped with overshoot and then lag coming back on...

I finally came up with a simple approach that is very stable and has the added benefit of self reporting how much heat it is making when stable. I turn the triac on/off over a period of 5 half-cycles. That way any DC component from odd number of half cycles alternates every period for no net DC. I can command 20%-40%-60%-80%-100%

My new approach is to set the heat output rate based on how many degrees the room is below the temp target. For measured room temp more than about 1 degree low the heat turns on full... with 4 other increments approaching the temp target. This varying heat by temp below target is analogous to the R in a RC circuit, The rooms thermal mass is the C. Kind of like a SERVO loop 8-)

Now the thermostat control rarely overshoots and when continuously heating the room settles in on a stable heat rate... This morning when it was around 40' outside the heater was working at a steady 40% (roughly 400W) of heat output to keep the room stable. If the inside temp was 65' that suggests my heater can hold the room at temp down close to 0' outside and it rarely gets that cold here in MS...

Cool.. or should I say warm...

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Entropy

Post by JR. »

raf wrote:PI or PID loop is your friend and what every process control in the world depends on
Like so much in life I had to figure this out the hard way. :oops:

I had to google PID..... (proportional integral derivative).

I get the Integration factor, that is the thermal mass of my room acting like a capacitor, and proportional control is variable heat rate which is like a variable current charging my room's thermal mass cap. I have another unknown resistance/leakage term which is the heat loss to outside, like a resistance from the cap to ground.

I experimented with capturing rate of change to predict future target intercepts with and adjust the rate based on that prediction but it never settled down and behaved as nicely as my much simpler approach..

In an ideal world you wouldn't want even the 1' of slippage between max heat and min heat. I guess in theory I might trick it by using a real and virtual heat target where the virtual target adds windage for that fraction of a degree based on last actual heat rate. But for now I am more than happy to live with a 1' degree slip from min to max heat...

My electric blanket is no longer dithering on/off and it gets proportional voltage drive too.

My old in-wall heater with mechanical thermostat was just horrible making too much heat on warm days, and not enough on cold days.

JR

[edit] I am about ready to stick a fork in this.. I added a little hysteresis to prevent dithering at the temp target, I also added some windage for the slippage. I could make it more precise, by slowly adding in a specific windage offset based on actual heat rate, but it really doesn't need to be that precise. I would get more benefit (perhaps) from sampling temperature at multiple places around the room, but again this is soooo much better than before. [/edit]
Cancel the "cancel culture", do not support mob hatred.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Entropy-more bedroom heater updates

Post by JR. »

Now that I control the heat with 5 levels I now rate the weather as 1 dog nights, 2 dog nights, etc.

I do not want to complain too much about the cold weather since it is much colder north of me, but last night was a 3-4 dog night and tonight will be colder. I need to figure out when it runs out of heat. I estimate I can go another 10' colder outside. If that happens I just turn on the old in wall heater too.

I had a recent thought to maybe buy a canopy bed and wrap it inside a thermal tent. This way I could literally heat the bed area even cheaper... Might be nice to locally cool it for the summer. I don't want it too airtight since humidity and body odors would probably negatively effect a tight environment. For now just a mental exercise. A cheap canopy bed is probably $350.

======
I have spent a few months playing with my smart slow cooker figuring our what works and what doesn't. My biggest lesson learned from all this is the power of my thermal insulation boxes to trap heat normally lost while cooking. This only works with electric stoves and slow cooking, but makes a huge difference in wasted energy.

FWIW a lot of cooky is very energy wasteful and while cooking off water content you don't always want to trap that, like when frying foods. These days I can recall the last time I fried anything.

JR

PS I miss the good old days when nasty politics stopped between elections...
Cancel the "cancel culture", do not support mob hatred.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Entropy

Post by JR. »

last night in the 20's was near a 4 dog night... I only have one more dog left. :D

Tonight is forecast for low 20's ... we'll see.

If I see the heater control lose regulation and room temp start falling I can turn on the old heater to assist.

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Entropy

Post by JR. »

Thanx, I have invested a bunch of time into trying to write predictive code to estimate future temps from recent past heat rise etc, but none have worked as well as my much simpler just set the heat output rate based on how many degrees below target the temp is. I could add back in a fudge factor offset for this slippage but that could be a source of instability during transition, so I can live with a degree of slippage or so. This is sooo much better than old mechanical thermostat on the in-wall heater that was too warm in warm weather and not warm enough in cold weather.

I have some asymmetry in the thermostat response. To prevent underdamped over-shoot, i turn it off or reduce heat immediately when above target and only increase heat rate slowly after a few seconds.

The strategy of plugging the electric blanket into the heater seems less than wonderful. I will probably return to plugging the EB into a time of day timer so it just turns on and off at night. A warm bed is always comfortable, even when it isn't very cold outside and I can turn off the timer next spring.

JR

PS I may share some recipes for my slow cooker, I have been getting better results each week from tweaking the procedure. I basically cook everything but the beans and rice in one pot slow overnight. I could cook the beans in the same one pot if I cook them first (so I can rinse the beans), but easy enough to cook the beans separately first. I cook the rice the next day before serving and use gravy/juices from the vegetables and meat that cooked down overnight . . My mouth is watering just thinking about it. :D
Cancel the "cancel culture", do not support mob hatred.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Entropy

Post by JR. »

I didn't try pulsing it per se, but I tracked temperature rise over time for the given heat output to try to predict an estimated future temp if all else remained constant. This will be a moving target long term since the thermal mass of the air is relatively constant but the heat loss flowing through the walls is variable. Further the measurement appears to be influenced by wind currents in the room that disturb warm spots and cooler regions. When I get up in the middle of the night to pee I disturb the stable air patterns and the heater takes a few minutes to stabilize again.

The cybernetic control kept me comfortable all last winter but I was annoyed this year when I plugged the electric blanket into it, since I could then hear the blanket clicking on and off as the room temp dithered around the target temp.

The combination of some hysteresis slows down the dithering around the threshold, and my use of linking the heat output to increments below target works like a charm. Now the heat output is relatively stable and I can even estimate how cold it is outside from how many increments below target it is stable at. Last night was stable around 2 leds, a 2 dog night, tonight the weather report is predicting low 20s so I expect to be pushing 4 to 5 dogs. If I ever get solid 5 dogs and temperature drops it means my 1000W heater is topped out. I can see temp cs target on the display in real time so easy to see exactly what it is doing, and how hard it is working.

I have returned to plugging the blanket into a time of day timer so it is now well behaved and not off too much when it isn't cold enough outside like last night could have been.

I have a (very) slight academic interest in refining this further, but it is mostly a box that I've already checked off and moved on. I suspect the low density air is not low enough thermal resistance to be very uniform temperature everywhere so control will always be a little uncertain.

JR
Cancel the "cancel culture", do not support mob hatred.
Post Reply