Heating circulator pump control

[alternety]

Thank you for your comments. The system is not a lousy design. It is quite well done in accordance with industry standards. The only mistake we are aware of is the variable speed control. The contractor needed help and received bad information from a vendor. This approach to circulation is not a common practice, but is beleived by some to be a better approach. No ones grandmother had a system doing it this way. Nor are most current installations. Generally speaking heating systems contractors ignore things like electrical energy cost and focus only on the boiler. This approach addresses one aspect of that and eliminates a bypass valve that can become noisy and an approach that dumps more heat into the boiler room. I know how other systems are designed and built. The contractor (and he is a good one) knows how systems are designed and built and how to do it.

I feel that the attacks on the system design are both uncalled for and incorrect. I came here for help. I would still like that help. I do not need attacks and sniping.

I apologise if I have not made the problem clear.

Problem:

1 - I have a 0-10 V or 20 ma signal from a SENSOR that represents differential pressure across the heating manifold.

2 - The pump that generates the pressure IS a variable speed pump controlled by 0-10 or 20 ma interface. It reaches full speed with all zones calling for heat.

3 - I need a controller to

a - start and stop the pump based on an external "heat demand" signal.

b - Control the speed of the pump to maintain a constant pressure across the manifold as zone vlaves open and close.

​

Thank you for your help and courtesy,
Harold

​

 

[rdrast]

Then your only solution, is to invest a few hundred dollars, and buy something similar to:
http://www.partlow.com/content.aspx?id=26

Or google on other 'Process controllers'

Best of luck to you, and I'd advise stocking up on spares of all pump related items, including pressure sensor and process controller.

 

[danw]

Part of the confusion in responses is that I'm sure readers understand which differential pressure you are measuring and why.

It might be interesting to understand where the physical differential is developed; between which two points in the system; and why would maintaining a constant dp across those two points might achieve your desired control.

But even without that elaboration, if you need to maintain a constant dp with a pump with a variable speed input, a stand alone PID controller will attempt to do so, depending on how it is tuned, and to some extent, as to whether the dp signal is ranged suitably for the application.

The idea to use a digital input to a PID controller to "stop the processing and drop the analog output to zero" is a nice idea, but such an implementation of DI on a PID controller is VERY manufacturer/model specific. I've learned the hard way, believe me.

Having used 1/4 DIN controllers for years, I have seen the number of specific actions (like, "go to manual mode" or "go to 2nd PID set") accomplished by a DI input to a controller increase substantially, if you go this route, but be sure that whichever PID controller you select actually has the specific action you need, because the implementation of DIs is all over the map, even within a given manufacturer's line.

Most 1/32 DIN and 1/16 DIN controllers do not have a DI option, or if they do, it is a dedicated action, like "reset" action to unlatch a high limit controller lockout.

1/4 DIN controllers which have optional DI inputs are typically above the $250 price point. eBay is loaded with 1/4 DIN controllers, but I see a lot of outright junk there - 15-20 year old devices for which parts haven't been available for a decade. Most seem to be sold as-is. Even getting a table to decode the cryptic model numbers for older, obsolete units is a challenge. The life cycle seems to be about 6-8 years for a PID controller model, so a 15 year old one could be 3 generations out of date.

I currently use Honeywell UDC controllers. The present UDC 2500 model has an optional DI input (dry contact only). One action option is

MANUAL FAILSAFE OUTPUT—Controller goes to Manual mode, output goes to the Failsafe value. ATTENTION This will cause a bump in the output when switching from Automatic to Manual. The switch back from Manual to Automatic is bumpless. When the switch is closed, the output can be adjusted from the keyboard.

If manual failsafe = 0 (a separate setting in the unit), then your output goes to zero and the pump motor turns off.

The tuning range is also a consideration. The smaller 1/32 & 1/16 DIN controllers are typically for thermal processes and frequently do not have a range of tuning constants sufficient for tuning a flow loop, that is, the gain won't go low enough or the PB won't go high enough.

A UDDC 2500 basic model with 0-5Vdc input, 0-20mA output and a DI is
DC2500-C0-1000-100-00000-00-0.
Honeywell wants $40 for a pair of precision 100K resistors in series for 0-10Vdc voltage divider for the input (the native input is apparently 0-5Vdc) but you could jigger a voltage divider with a couple 100K ohm resistors yourself. You can develop a 0-10Vdc output by driving the 0-20mA current output through a 500 ohm resistor to get 0-10Vdc. The units come from the factory as 4-20mA, but there's an easy-to-do calibration routine (need a milliamp meter) to adjust to 0-20mA. Current pricing on the distributor's web site is $470.

The previous generation was the UDC 2300. The corresponding model number for a UDC 2300 is DC230B-C0-20-10-000000-x0-0. Same thing applies for a voltage divider on the input and a a dropping resistor for the output.

Avoid blue bezel UDC 2000 models which are antiques. Gray keypad UDC 2000s are also obsolete.

Dan

 

[Steve Bailey]

There are dozens of manufacturers of single loop process controllers out there. Omron, Eurotherm, and Omega come quickly to mind. Automation Direct offers a few models.

One problem presented by your system is that you are using a feedback signal that measures pressure to modulate a command that represents flow. You would have better (easier to tune) control if your sensor measured flow rather than pressure. If your pump is fixed displacement, once the system has been purged of air you may find that the variation in pressure due to changes in pump speed is very fast and your pump speed oscillates. If you don't already have some type of accumulator in the system, I'd suggest you consider installing one.

 

[alternety]

Where the pressure is generated

Let me answer the question about where the controlled pressure is developed.

In a radiant hydronic heating system, the energy source heats water which is made available to the individual zone loops that are the actual heating pipes under the floor. Simple (or old) systems may have only one or two loops/zones. My system has about 18 zones and the heat demand of each loop may be quite different. Various factors affect this. Room size, solar input, glass, use. A bathroom may be a very small load but the great room a very large one. The actual system is more complex than descibed below, but this model is useful for this discussion.

The zones may have a number of ways to get hot water. The most common ones are control valves and individual circulator pumps. With this many zones, valves are the solution I have chosen.

Typically, with either approach, the hot water is made available by supplying it to a a relatively large diameter supply manifold which distributes the water to smaller attached pipes that carry it to the zone. All of the loops return their cooled water to a return manifold. In a large system, this main manifold has taps that supply a second level of manifold with individual zone valves. If this second level wants water it is drawn from the main manifold. This is what I have implemented.

The way flow (which equates to BTUs/time interval) is controlled is to set up the individual zone loops to be of equal pressure drop at rated flow. Some of this is acomplished by keeping the loops about the same length. If a room needs more BTUs there is another loop added. The flow may be further controlled by flow balancing controls in each loop. This is what I use. The flow for each loop is determined by the heat loss calculations which are done on a per-room basis. The whole flow balancing process depends on having an essentially constant pressure across each loop.

It is necessary to use a pump that will supply the maximum flow with all loops demanding heat and at this flow rate maintaining the required pressure to allow all the control valves to do it correctly. The problem is when a zone representing maybe 1 or 2 percent of full system heat is the only zone active. You then have a serious flow mismatch, pressure soars, upsetting the flow control, and really annoying the pump. The most common solution to this is to place a bypass (i.e. pressure relief) valve across the main source and return manifolds and run the pump at full speed as long as any zone is calling for heat.

The constant speed pump with bypass has a number of issues. Excessive power consumption, noise from the bypass valve, and more heat radiated in the boiler room because the return manifold is hotter than required. One favored solution is to have all loops constantly circulating water and modulating the temperature. I am not going into that one but I do not favor it. Heating system designers tend to ignore electricity consumption and focus only on boiler efficiency. If it was good enough for grandmaw it is good enough for you. Historically, the boilers are also grossly over sized.

The solution to oversizing is a room-by-room heat loss analysis. The solution to circulation overhead is a variable speed pump supplying the main manifold with a constant pressure across source and return. This is accomplished by measuring the pressure across the manifolds (i.e. differential pressure) and adjusting the variable speed main pump to maintain this selected pressure. This pressure is calculated based on all system parameters during the system design. Pressure rather than flow must be controlled. Flow is properly highly variable.

The result is lower energy consumption and potentially somewhat lower noise levels. Noise may be a problem in pumps at low motor speeds; particularly with simple speed controllers.

I hope that provides a suitable basis to understand what I am trying to accomplish. I know there are a lot of controllers out there. I have spent a bunch of hours trying to find the one that works. Working beats price to a large extent, but I would like to spend the least amount of cash. This is my system. I am following up on all specific device recommendations. I am hoping your expertise can help me to find the right one without a whole bunch of additional hours of searching.

Thanks for your interest,

Harold

 

[CroCop]

D0-05AR $119
F0-2AD2DA-2 $149
PC-PGM-105 $99

Automationdirect.com

AC Input PLC, Relay output. 2 Voltage in/2 Voltage out Analog Card. Software. Supports up to 4 PID loops, Autotune, you can feed forward, drive the output to high or low, filter the input incase of noise, etc. Has a decent PID interface. Software is a reasonable price. You can jump to the DL06 with 4 slots for expansion cards, and up to 16 PID loops. With the software you can make the system much more flexible then a regular stand alone controller.

 

[nif]

Hello!

The process is quit simple, but I think that you would need
a pressure-accumulator in the system otherwise you will get pressure spikes in the system.
Only a pump is not fast enough to keep the pressure constant, you will break the pipes or the pump in my oppinion.
To add a pressure-accumulator to a existing system is easy you only add a pipe and the accumulator.

 

[steveb1475]

Check with Newark electronics they carry a lot of controllers. You should be able to get by with a temperature controller that has either 0-10V or 4-20ma inputs and outputs, you did say that your sensor and motor speed controller are configurable for either right?

 

[steveb1475]

Actually, this one from automation direct should be better than the above post.

 

[rsdoran]

I assume it was my comments that you were refering too earlier, the simple fact (as I see it) is you designed a system that does not work as expected for whatever reasons.

I am not an engineer but lets say I have dealt with a few boilers here and there.

Typically, with either approach, the hot water is made available by supplying it to a a relatively large diameter supply manifold which distributes the water to smaller attached pipes that carry it to the zone.

That large diameter supply is usually called a "header".

I am not sure about your assessment of boilers and their efficiency. In most cases boilers are rated 20-30% over demand for contingencies. The average efficiciency is 75-85% but newer systems offer as high as 90%, the inefficiencies are primarily caused by heat loss in ventilation and convection, depending on certain factors some of these losses can be recaptured.

Fluid dynamics shows a relationship between pressure and flow. A pump provides flow to a system whose resistance will create pressure...as it pressurizes you could possibly vary the pump speed which in turn could vary the flow but at some point the flow will need to stop or be bypassed...how this is accomplished can depend on varying factors.

Grundfos pumps with single phase motors may offer 1, 2 or 3 speeds but I am not familiar with one that offers variable speed control.

From your last post I read it that the pressure differential is read between the header (main manifold) and loop supply manifold to determine when water is needed, in other words if a loop opens a pressure differential allows flow which draws water from the loop supply manifold which in turns lowers pressure so water is drawn from "header" which in turn lowers pressure so pump needs to run to fill header and re-establish pressure at desired level.

I am not sure but I think you want to control the pump speed (flow) based on the "difference" in pressure...ie header pressure is 150ps1 and drops then start pump and set speed (flow) proportional to the difference.

If my above assumptions are correct then I can think of a couple of things that could possibly be done using simple controllers (in the $250 or less range) like the ones mentioned above from automationdirect.

There are several ways to do this, personally I would still use a bypass valve for that "just in case situation", this would allow the pump to run for X seconds or minutes after the pressure differential is gone, which in turn could eliminate excessive cycling. Technically I think bypass valves are required by code but not positive about residential, I know it is in most industrial/commercial applications.

One method (depending on the response of the motor speed controller) is when X amount of pressure differential is reached then run the motor at full speed then slow it as the differential decreases then run it at slow speed for X time period (this is where a bypass valve comes in) in case demand returns. Starting a motor (cycling) can cost more then it running longer once started.

If the pump motor was 3PH then a VFD with analog input could possibly be used as a controller and control the flow based on the pressure differential.

If I did not understand the application and none of this is applicable then I apologize for wasting our time.

 

[WildBill49]

If you decide on the Micrologix 1000, try going to E-Bay. I recently bought one there for $54.00. I got the 24vdc power supply for $19.00. I also got some din rail and panduit for almost nothing. I have a problem with the baseboard heating in my house. The zone for the bedrooms runs along the outside walls. I have had freeze problems that last two years. I have the Micrologix 1000 turning that zone on every 30 minutes, for 3 minutes. This sends some hot water out there to keeps from freezing.