??'s about replacing a self-heating Tstat (for steam)

ChrisJ

> @Precaud said:
> Do you enjoy being a jerk?

No.

mattmia2

I think tekmar has some stuff that is very tewakable. You could buy 10 until you get the right one or just buy the expensive one with all the control you want in the first place.

SteamingatMohawk

You get what you pay for..." in most cases"!

mattmia2

Note that you aren't paying for the $10 in electronics, you are paying for the technical sales and support people who can help you get the right product for your application and help you configure it properly.

ChrisJ

mattmia2 said:

Note that you aren't paying for the $10 in electronics, you are paying for the technical sales and support people who can help you get the right product for your application and help you configure it properly.

To be fair,
The green screen VisionPros I've taken apart I couldn't build that PCB for $10. Of course, if you run a few hundred thousand maybe the price could get that low but I doubt it. Maybe 2 or 3 times that.

You're also paying for their hardware and software engineers time to make the product. The better touch display etc.

Precaud

Yeah, this has nothing to do with value, whether real or perceived. It's about a forum member taking cheap personal shots versus staying on topic.

I found two models last night that have very granular (starting at ±0.2º) differential control and settable minimum on/off cycle lengths.
I also ordered a replacement relay for the Honeywell I have.

Heating season is about over here. There won't be opportunities to check these things out in use. I'll probably set it aside until fall.

ChrisJ

Cheers @Precaud

SteamingatMohawk

@Precaud Earlier, you said, "Generally true, but having watched this one cycle three times, I can say with confidence it uses integer temperatures above and below the set temp. So a diff setting of 1º is actually a 2º differential. "

If the thermostat only reads out in whole numbers, you cannot determine the actual differential to a smaller increment. You would have to have a digital thermometer that reads out in 0.1 degree increments to be accurate.

For example, if the thermostat reads in whole numbers, the actual signal for say, 70 degrees could be anywhere from 69.5 to 70.49 degrees and you wouldn't see the difference on the display. I hope this is helpful.

Precaud

Perhaps, but from what I saw, it looks like they are truncating to integers internally, not rounding. But it's a moot point, the unit has already been returned. Supplyhouse agreed with me, the diff spec as given is misleading. Their experiment showed that it always cuts off 1º above the set temp, and the diff setting only determines the cutin point.

I ordered a Lux non-programmable DMH110A which has much finer swing increments (from ±0.25º in 0.25º steps), with miminum on and off cycle length of either 2 or 5 minutes. It would be nice to have more choices on the latter,and even have each one be independently programmed, but this should be sufficient.

I think having separate settings for cutin and cutoff increments, and separate on and off cycle lengths would be ideal for steam. Pretty close to what @PMJ does with his timer. I'd like to try tighter control on the cutoff, and looser on the cutin.

PMJ

Precaud said:

Perhaps, but from what I saw, it looks like they are truncating to integers internally, not rounding. But it's a moot point, the unit has already been returned. Supplyhouse agreed with me, the diff spec as given is misleading. Their experiment showed that it always cuts off 1º above the set temp, and the diff setting only determines the cutin point.

I ordered a Lux non-programmable DMH110A which has much finer swing increments (from ±0.25º in 0.25º steps), with miminum on and off cycle length of either 2 or 5 minutes. It would be nice to have more choices on the latter,and even have each one be independently programmed, but this should be sufficient.

I think having separate settings for cutin and cutoff increments, and separate on and off cycle lengths would be ideal for steam. Pretty close to what @PMJ does with his timer. I'd like to try tighter control on the cutoff, and looser on the cutin.

This. @Precaud , you are definitely on to it now. Good show.

Precaud

Thanks @PMJ . I think most of it could be accomplished by having asymmetrical hysteresis; i.e. not being restricted to having the temp set point in the middle of the swing. That would automatically give you a longer off cycle which automatically scales to seasonal heat demand. For example, a 0.25º cutoff and 0.75º cutin. The minimum cycle times can then continue to be more for eliminating short cycling. Make sense?

Jamie Hall

Mercury T87...

PMJ

@Precaud , practically speaking isn't the deadband still 1deg? The cutin and cutout are still 1deg apart.

Anyway, as you become successful at evening out the heat and lengthening the calls, the Tstat becomes less and less of a factor. The real control comes from managing what the boiler does in many cycles within a single call for heat. You want fewer slowly satisfied calls with mote cycles not quickly satisfied calls with few cycles.

Precaud

LOL @Jamie Hall . If I correctly understand how "anticipators" work, I think what I described is a digital implementation of it. Both approaches shorten the heat-on part of the cycle.

Precaud

PMJ said:

@Precaud , practically speaking isn't the deadband still 1deg? The cutin and cutout are still 1deg apart.

In the example I described, yes, But it doesn't have to be - it could be anything. But the key concept for me is accounting for the heat production lag caused by the mass of the system. Sorta like the planet's seasons. Longest heating day in the north is June 21, but warmest temps are 6-8 weeks later, even though less heat is being received then.

PMJ

Precaud said:

PMJ said:

@Precaud , practically speaking isn't the deadband still 1deg? The cutin and cutout are still 1deg apart.

In the example I described, yes, But it doesn't have to be - it could be anything. But the key concept for me is accounting for the heat production lag caused by the mass of the system. Sorta like the planet's seasons. Longest heating day in the north is June 21, but warmest temps are 6-8 weeks later, even though less heat is being received then.

RIght, there is a lag. That is what the anticipator did, shut the boiler off early anticipating that more heat is coming. It really did work pretty well.

The thing is, the anticipator is expecting the tsat to be satisfied in one burn. It simply stops that burn early but run enough to still get to temp which does make the overshoot less. It does a lot better than nothing. But single burns to satisfaction do not produce even heat, at least no where near as even as it can be.

Here is what I have trouble getting across. Perfectly even heat would mean that the tsat is NEVER satisfied. There would only be one cut in and never a cut out. The temperature would simply never fall below the cut in again but also never make it to the cut out. That would be perfection which of course is impossible. But the closer you get to it the longer the calls are with many burn cycles within each one.

So logically to be successful evening out the heat the tsat simply cannot be used to cycle the boiler because fundamentally you don't want it (the Tstat) doing any cycles. The burn cycles on an on/off boiler simply must be controlled another way to really make things even.

Of course single burns from cut in to cut out produces good enough heat for most folks. When I moved in there was a T87 and a vaporstat on my system. All I can say is if I told my wife we were going back to that she'd throw me out. These systems really can be way better. I've lived both. I truly believe they wouldn't be tearing these systems out all around me if they all performed like mine does.

Jamie Hall

Actually, in a sense (but only a limited sense) perfection is possible, @PMJ . The limited sense is that whatever your control system is it has to be set up with a proportional band -- which can be very narrow -- and it has to be running a proportional control. Think of it this way (I am thinking about airplanes here, but the principles apply -- car cruise controls, at least the very good ones, are similar). You have some sort of desired control parameter -- in the case of home heating, that would be the space temperature at the sensor. What needs to happen is that if that parameter changes, the control device needs to send a signal (not on/off !) to a driving system, which will then change slightly (again, not on/off) to restore the controlled parameter to where it should be -- only it will never quite get there if the damping is correct. Overdamped, and the response will be too slow. Underdamped and you will get overshoot. Seriously underdamped and you get divergent oscillations, which can ruin your whole day (off topic: the first DC-8 aircraft were inherently very seriously underdamped in a narrow, but important, range of speed and power, and took serious skill to fly through that range).

More concrete: consider a really good automotive cruise control. You set it to maintain a "constant" speed. What then happens is that the control senses variations from that speed and varies the engine output to return to that speed. There is a band of speeds (relatively narrow in the good ones) within which the control varies the engine output from idle to full power -- but it's not bang bang. Going up a steep hill heavily loaded, the actual speed will droop and the engine output will be high. You crest the hill, and the speed starts to increase and the engine output will drop. If the relationship between degree of droop or overspeed and engine output is well thought out, the result will be very close control of speed with no overshoot, but at the same time never quite holding the set speed (the difference will be unnoticeable to the driver).

The problem -- particularly with steam -- is that we are plagued with both a control device -- the thermostat -- and a controlled device -- the boiler -- both of which are either on or off. One could liken it to a cruise control where the control device was on (too slow) or off (too fast) and a controlled device which was either on (full power) or off (idle). So long as the outside parameters driving the system -- say road grade, for instance -- were held constant, one could come up with a satisfactory means of holding a remarkably constant speed (though not exactly) by timing or anticipating or some such. As soon as those parameters change, though...

Control theory is fun!

Precaud

@PMJ , I repect your approach and have learned a lot from reading your accounts. If I had a two-pipe system and it were my only heat source, I'd probably follow your lead. However, your objectives and conditions of use are different than mine.

On most days, my steam system is a backup, filling in the gaps of the passive solar heaters. Only on cold, cloudy days does the steam operate during the day. On normal, sunny days, the steam is off from say 8:30am to 4pm. It's heat curve has to splice in with the solar heaters' curve. So I need my system to be responsive to, and subservient to, the tstat. My main objective is not to hold house temps so consistent that I don't notice it. It's primarily to orchestrate a smooth crossover between the two systems in the late afternoon / early evening.

If my Honeywell RTH230B didn't self-heat, I could probably live with it. With CPH set to 3, it's reasonably good. But the self-heating is a deal-breaker.

These are my primary heaters. Even on the coldest sunny days, they heat the main floor to 75º by mid-afternoon, and cost almost nothing to operate. ROI was one year.

PMJ

@precaud,

I understand everyone's needs are different and don't wish to force anything.

It would appear that you need your system to come on line based on a call from the tsat when the other system is still providing heat, just not enough anymore. That would require an even smaller portion of the boiler's capacity making control even tougher. Still looks tough to me from the Tstat alone but best of luck with it.

PMJ

@Jamie Hall , I will say this journey has been interesting and fun. I am under no illusion that mine is the only way. I do wish others would jump in and try things. The control platforms are available and so inexpensive. The dead men would love this stuff.

Precaud

PMJ said:

It would appear that you need your system to come on line based on a call from the tsat when the other system is still providing heat, just not enough anymore. That would require an even smaller portion of the boiler's capacity making control even tougher. Still looks tough to me from the Tstat alone but best of luck with it.

Actually, this solar system is a low-thermal-mass system designed to produce heated air, not to store it like classic Trombe walls and such. It heats up and cools down quickly. The thermal mass is in the house itself.

Yes, temps peak around 3pm and tail off from there, but there's a hard cutoff when the internal temp of the heaters reaches around 95ºF. The fans cut off and the vents close. In the house, it's typically 74-75ºF at that point, and tails off to whatever the tstat is set to, at which point the steam starts up and takes over.

One could argue that steam is probably not the ideal system for such a setup, but we have to dance with the partner we have. Steam and passive solar do have one nice thing in common. Both are gentle air heaters; neither super-heat the air.

HVACNUT

Has anyone suggested the Honeywell Prestige IAQ. Excellent thermostat.

Precaud

Looks like total overkill for my needs.

ratio

Precaud said:

Looks like total overkill for my needs.

Perhaps, but they're the one (consumer grade) thermostat that I know of that, not only controls via PID, but exposes tuning variables directly. It should be able to control very accurately, once dialed in.

NB: I have not made a search for any others.

PMJ

ratio said:

Precaud said:

Looks like total overkill for my needs.

Perhaps, but they're the one (consumer grade) thermostat that I know of that, not only controls via PID, but exposes tuning variables directly. It should be able to control very accurately, once dialed in.

NB: I have not made a search for any others.

Anyone out there using a PID control for residential steam?

I have several applications of PID control at the plant. None of them with the cause/effect separated by large amounts of time though like would be a home heating system and hugely varying response times at that. I think it would have a really tough time - but then never tried it. Anyone?

ratio

PMJ said:

I have several applications of PID control at the plant. None of them with the cause/effect separated by large amounts of time though like would be a home heating system and hugely varying response times at that. I think it would have a really tough time - but then never tried it. Anyone?

Perhaps I'm wrong about it. My understanding of PID control is not exhaustive and my experience is limited.

It is quite interesting, however.

PMJ

ratio said:

PMJ said:

I have several applications of PID control at the plant. None of them with the cause/effect separated by large amounts of time though like would be a home heating system and hugely varying response times at that. I think it would have a really tough time - but then never tried it. Anyone?

Perhaps I'm wrong about it. My understanding of PID control is not exhaustive and my experience is limited.

It is quite interesting, however.

Interesting it is. Glad you brought it up. Perhaps someone has tried it and will have a comment.

mattmia2

Interesting to see steam in a southwest masonry/adobe style house.

Precaud

Yeah... high-thermal-mass house and heating system... a very nice combo.