T87F anticipator appears faulty

Jamie Hall

https://forum.heatinghelp.com/discussion/comment/1871637#Comment_1871637

and then a new one three years from now… and then another new one… and then…

One of my T87s was installed, to the best of my recollection, somewhere in the early 1950s. one of the others (which was installed as a backup to a digital) around 2010, but it was NOS and who knows how old it ones.

But then I'm odd. I like things which just plain work, and keep working.

pacoit

@coby , Fair statement. I think we are just curious about just exactly how the T87F 2055 works.

@109A_5 , Nice work! My tstat is also a T87F 2055. Yes, it looks just like the one in your pictures. I'm still not clear on the function of the resistor's flying connection to the axle of the bimetallic spring?

Regarding my tstat, since I was measuring the resistor from a fixed probe at right end, and moving the other probe leftward, I naturally read increasing resistance. But, if the fixed probe should have been at the left, then I got it all wrong… However, this means the tstat is working correctly in that respect!

More generally, I see now that the proper formula for adjusting the anticipator according to the system amps appears to be: P= I^2 * R (Power = current^2 * resistance). So, to maintain the same anticipator heat output, a higher amp circuit requires a lower resistance setting, which is in keeping with directions to move the wiper to the left.

Some T87F instructions do say that, for a steam system, set anticipator to 1.2. Does anyone here have a steam system with anticipator set to less than 1.2?

Jamie Hall

Mine are set to 0.3 more or less… which seems to work well enough. But Cedric's home is a little odd…

mattmia2

bimetallic spring?

Jamie Hall

https://forum.heatinghelp.com/discussion/comment/1871892#Comment_1871892

Does he mean the coil that holds the mercury switch?

mattmia2

maybe. I don't think there is an electrical connection to that but I could be wrong. I have a different series one laying around at home, not sure if it is a 2 stage or some older gas valve. The furnace it was on had a second stage valve sort of kludged in to it for some reason.

109A_5

" coiled bimetallic spring "

The confusion puzzles me, is that not what it is ? It is coiled and it is spring and it is bimetalic (not the variable resistor, the part the Mercury switch is mounted to). And in the case of the anticipator circuit it is a very subtle path but none the less it is a parallel current path to the left most (about 25%) of the variable resistor the small Left end. I suspect the current through coiled bimetallic spring warms it slightly causing it to expand.

@pacoit " Some T87F instructions do say that, for a steam system, set anticipator to 1.2. Does anyone here have a steam system with anticipator set to less than 1.2? "

Yes I do, it is about in the middle of the range.

BTW not my pictures, they were found in that previously mentioned article, that contains some erroneous annotations IMO.

Some time ago I read that in the early days of steam the thermostat directly controlled equipment that drew 1.2 Amps or more. So since the 1.2 anticipator setting effectively disables the anticipation resistance. It was more for best function and not burning up the anticipator resistor when used with that old equipment. I wish I could find that article now. For the most part I believe that 'Steam' setting is now obsolete information.

With modern equipment the anticipator should initially be set by the current flowing through the thermostat circuit. Then varied if needed.

109A_5

@pacoit Maybe this helps some. Item 3 points to the where 'flying connection' of the resistor is connected (apparently crimped) to the metal scale. Item 2 provides contact pressure for the connection to the axle of the coiled bimetallic spring (Orange on the wiring diagram) beneath the anticipator resistor.

pacoit

@109A_5 , Thanks for the detailed diagram. Unfortunately, I still don't understand it. One specific point is the connection from the number 6 up and then right to the dot; to my wiring diagram newbie eyes, it looks like it bypasses the mercury switch so that the circuit is always made/closed? Also, if the orange is the bimetallic spring/mercury switch, then what is the switch above?

109A_5

The dot is just a connection node of three things;

Anticipator's wire wound resistor via a wire to the Mercury switch.

The Mercury switch heat contact.

The coiled bimetallic spring.

Schematics and wiring diagrams some times (actually often) look different but represent the same information electrically, they often have different priorities, hence the two different styles of drawing. The picture is more like a wiring diagram and the other image is more like a schematic. Point 6 is actually connected to point 1 by a wire, all are actually connected at the Mercury switch (the Heat contact of the Mercury switch). The coiled bimetallic spring is also connected to Heat contact of the Mercury switch at the Mercury switch. The metal bracket is soldered to the switch contact pin along with the wire from point 6.

The Mercury switch is not bypassed. It only switches the connections of the resistors (both) to the R terminal of the thermostat.

The coiled bimetallic spring and part of the anticipator's wire wound resistor are wired in parallel.

I could redraw the bimetallic spring and the anticipator's wire wound resistor differently if you think that will help.

Jamie Hall

The mercury switch is next to the "R" in the diagram, which shows it as an open SPST switch (Note that some models of the T87 used a mercury switch which SPDT, with the other contact for cooling). Note that the big bimetal coil to which the mercury switch is attached — shown with a resistor symbol in yellow — isn't really a resistor, or certainly not much of one. It isn't much of a spring, either, though it looks like it might be. It is simply a coiled bimetal, and as it changes temperature it coils more or less — tipping the mercury switch one way or the other.

The true genius of the T87 is the combination of this coiled bimetal and the mercury switch. As the coil coils more — and the temperature drops that is what it does — eventually the mercury switch tips. At that point, the mercury in the bulb moves to the now lower end, completing the circuit — but also putts bit more coiling force on the bimetal. Now as the temperature of the bimetal rises, it tries to uncoil — but the temperature has to rise to a point slightly above the temperature at which it first tipped, since the weight of the mercury at the end of the switch is trying to keep it more coiled. But — eventually it overcomes that and the mercury switch gets tipped back the other way. This incredibly simple arrangement gives the T87 an inherent "dead band" (sometimes referred to as "swing") of about half a degree Fahrenheit, ensuring that small transient variations in temperature — a breeze, perhaps — won't cause the thermostat to turn on and off rapidly.

The anticipator has a completely different purpose, and exists to warm the thermostat (slightly) to cause it to shut off sooner than it would all by itself — to compensate for the fact that the radiation continues to give off heat for some time after the boiler is shut off. Except for very simple installations in small spaces, setting the anticipator is something of a black art, as the degree of anticipation which is exactly right for one space may be quite different from that wanted in some other space.

109A_5

The Orange resistor between 1 and 2 is the coiled bimetallic spring.

The Blue resistor between 5 and 6 is the wire wound anticipator resistor.

3 is the flying tap wire of the wire wound anticipator resistor about 25% physically from 6.

4 is the variable wiper of the anticipator resistor.

The path of 3 to 2 is the metal scale and the copper leaf spring to the axle of the coiled bimetallic spring.

EdTheHeaterMan

There is a lot of discussion here about the innerds of a T87F 1075 or 2873 or any of the many other versions of that model.

This is the basic diagram with no frills

The three wires that are connected to the mercury bulb are connected to the terminals so that R is the common and Y is the make on temperature rise. This is good for cooing applications or on heating applications where you need power to close a valve on a heating system. The proper sub base would need to be selected based on the needs of the system

R is also common the the W terminal thru the heat anticipator and will make on temperature fall. The more of the resister that is included in the circuit, the more heat is generated to cause the thermostat to get satisfied early. The reason you may want the satisfy early is to reduce the over shoot that some systems have as a result of the design. In this diagram there is a switch that is closed between R and the mercury bulb. This switch would be located on a sub-base and not on the actual thermostat. So when you are looking at the sub-base that comes with the heat only model (shown) you are only connecting wires to the R and the W. There is no switch in the heat only sub-base, so the switch in the wiring illustration is not actually there. The Y terminal is non existent, so this would be for a basic heating only system that does not require power to stop the heating cycle.

Any further discussion is pointless, so just get a replacement thermostat if it is not doing what you need it to do.

109A_5

@Jamie Hall I call it a coiled bimetallic spring since it has all these properties.

It is coiled.

It is bimetallic.

And the spring functionality gives a bit of hysteresis when the Mercury moves to the other end of the glass tube upon a change of state. Without the spring functionality there would be much less hysteresis.

I believe the small portion 25% physically of the wire wound anticipator resistor is so small placing the current through the coiled bimetallic spring also enhanced the anticipation operation at the higher current settings. So the two parallel heat generators are used to give the correct functionality curve.

Jamie Hall

Whatever … I think the common conclusion might be that the T87 basic thermostat design is incredibly simple, totally reliable, wil work essentially forever… what's not to like?

pacoit

What's not to like? Hmm… I'd like it to have an adjustable differential(?) between boiler-on and boiler-off temperature. My T87F differential seems to be approaching 2F. It doesn't seem like much, but it is noticeable, especially if some parts of the house tend to cool faster than other parts (e.g, weather-facing sides); and if one has vacuum steam heat, pulsing the boiler a bit more predictably can add efficiency (keep pipes warm and vacuum up). Me thinks 1 or 2 $5 temperature controller/fan switch that allow setting On and Off setpoints is all that is needed in a thermostat. Don't need an anticipator or on/off timers. Have one sensor for the room, and one on the last radiator to know when radiator is filled or needs filling; the radiator sensor mostly runs the show as long as the room sensor is calling for heat.

I think the T87F is pretty cool, too. But I think all old technology is pretty cool; it teaches so much about ideas and concepts and solutions. A computer chip conveys nothing, no matter how long you stare at it; it's all inside and invisible.

109A_5

The adjustable anticipator resistor provides an adjustable differential or dead band or hysteresis, kind of a side effect of being able to match to a variety of system thermostat currents.

Mine seems to easily hold under 0.5 degrees with my steam heat. The temperature here is so steady at the thermostat that whenever I look at it, it is the same with no noticeable deviation.

My T87F holds the heat so tightly where it is centrally positioned in the house that better steam balancing through the house or better improvements of the house's envelope would be needed. The thermostat is a single point in the house, different heat loss amounts in different parts of the house is not a defect with the thermostat.

" Have one sensor for the room, and one on the last radiator to know when radiator is filled or needs filling; the radiator sensor mostly runs the show as long as the room sensor is calling for heat. "

Sounds like a bit of @PMJ's technology and that is fine if you want or need to go that way.

As for the age (50 + years old) of the basically electrical / mechanical technology it works great for me, I can't complain, and have no desire to change it.

By your original descriptions I suspect your T87F either has a defect or you are adjusting it incorrectly for the connected system. Or poor location in the home or a draft through the hole in the wall where the wires enter.

" My primary control states 0.8A. "

Did you measure the actual thermostat current ? Do you have a damper ? The thermostat may only see the damper control current and not the primary control's current.

pacoit

Yes, I had stated the tstat current of 4.8A. Yes, I have a flue damper. And yes, of course, this is "a bit of @PMJ's technology". Very applicable to my new (or restored) vacuum system. The next few monthly gas bills will tell what savings, if any, the new system provides.

109A_5

I went back and (my last post) looked for the current measurement but I must have missed it.

Anyway I would think if you are using @PMJ technique the anticipator would be disabled for the narrowest thermostat hysteresis of just the movement of the Mercury on the bimetallic spring. Maybe it simply does not matter. I thought with @PMJ technique the actual thermostat call for heat was very long (many boiler cycles) since other influences limited the boiler's run time to minimize the temperature roller-coaster ride of the radiators.

" Don't need an anticipator or on/off timers. " I believe he does use at least one timer to limit the system's fill of steam.

pacoit

Yes, with @PMJ's technique the call for heat can last many boiler cycles. I currently have the anticipator max left (basically, off).

I found that even with @PMJ's technique, timers are not needed. Instead, for the delay-off timer, I place the last radiator temperature sensor farther down the radiator and set upper temp setpoint experimentally for desired fill; for the delay-on timer, I select a larger/smaller differential between upper/lower setpoint temperature experimentally. This provides full range of adjustment and flexibility.

109A_5

Ok sounds interesting. Are you using an electronic temperature controller to measure the radiator temperature ? Thermocouple, RTD or Thermistor ?

pacoit

I'm using a cheap "temperature controller". It's sold as a thermostatic fan switch. There are many to choose from. I happened upon this one, 2 for $10:
XH-W3002 DC 12V 120W Temperature Controller

My final setup may be different.

109A_5

OK, I was not sure if it was something like that or a simple snap switch. Maybe that's all you need since it looks like you can set the Start and Stop temperature independently. Looks like it has a delay too if needed.

Too bad they don't make a 24 VAC model. I think the much more expensive ones like Omron and others have 24 VAC power as an option. However a 12 VDC power cube is inexpensive.

pacoit

FYI, the XH-W3002 delay function is just a power-on delay, not a signal on/off delay.

I went with a $10 24VAC-to-12VDC transformer attached to the boiler 24VAC output.

109A_5

Well that is confusing they just state "Set delay start" not power up delay start. And the just the words "Start" and "Stop" instead of "High" and "Low" temperature. Must be the Chinglish.

pacoit

To be more precise, the delay timer delays the "on" or "out" signal. I just checked. Also, this review is quite complete:
https://www.omchsmps.com/xh-w3002-temperature-controller/

109A_5

Thanks for the link.

Kind of an awkward product. No 24 VAC version. No dry contacts available. No degrees Fahrenheit mode. With a few tweaks it could be much better IMO. Misinformation pictured, that plug is not used for 230VAC. But it is inexpensive.

pacoit

I agree. There are others. Try searching "Temperature controller" on Amazon or Ebay.

109A_5

I have temperature controllers, mostly thermocouple types. I'm just surprised there is not a specific or inexpensive more universal one that is closer to that application type. Maybe there is no demand for such a device except a few folks experimenting with steam control.