Discrete Inputs and Outputs for HVAC system

Or another window comparator I have used in previous designs was the LTC1042.

https://www.analog.com/media/en/technical-documentation/data-sheets/1042fa.pdf

. . . but this one needs an external reference whereas the LTC1843 has an internal reference. You can use a single reference and use a string of six resistors to generate your different reference voltages for your window comparators.

How about an Arduino?

How about a small Programmable Logic Controller?

(Not an endorsement - other brands are available).

Here's a possible solution. You have a Read-Only Memory (ROM). The address lines are logic levels controlled by the damper positions, 0 = closed and 1 = open. The data lines are connected to drivers for the motor relays 1 - 6.

The ROM is programmed so that addresses with one '1' output = 0x01, with two 1's as 0x02, ... with six 1's as 0x06. It only requires 64 bytes of memory.

Contents of ROM:

Address - Data

000001 00000001

000010 00000001

000011 00000010

000100 00000001

000101 00000010

000110 00000010

000111 00000100

001000 00000001

001001 00000010

001010 00000010

001011 00000100

001100 00000010

001101 00000100

001110 00000100

001111 00001000

010000 00000001

010001 00000010

010010 00000010

010011 00000100

010100 00000010

010101 00000100

010110 00000100

010111 00001000

011000 00000010

011001 00000100

011010 00000100

011011 00001000

011100 00000100

011101 00001000

011110 00001000

011111 00010000

100000 00000001

100001 00000010

100010 00000010

100011 00000100

100100 00000010

100101 00000100

100110 00000100

100111 00001000

101000 00000010

101001 00000100

101010 00000100

101011 00001000

101100 00000100

101101 00001000

101110 00001000

101111 00010000

110000 00000010

110001 00000100

110010 00000100

110011 00001000

110100 00000100

110101 00001000

110110 00001000

110111 00010000

111000 00000100

111001 00001000

111010 00001000

111011 00010000

111100 00001000

111101 00010000

111110 00010000

111111 00100000

Correction:

"The ROM is programmed so that addresses with one '1' output = 0x01, with two 1's as 0x02, ... with six 1's as 0x06. It only requires 64 bytes of memory."

should be:

"The ROM is programmed so that addresses with one '1' output = 0x01, with two 1's as 0x02, three 1's as 0x04, four 1's as 0x08, five 1's as 0x10, six 1's as 0x20. No dampers open would have 0x00, blower motor turned off. It only requires 64 bytes of memory".

Welcome aboard. There are a lot of good people on this site who are willing to share what they know.

Cheers!

Crossbar Diode Logic

One can trivially implement or's and and's with crossbar logic. You can implement non-inverting outputs(buffers) and inverted outputs with your optocouplers driven by the output of crossbar lines. Diodes might be tiny, extremely cheap, signal (so low power) diodes such as the surface mount 1N4148 or similar. Crossbar lines driving optocouplers can be used to drive derivative crossbar lines within the same crossbar array. Thus, you can build up any arbitrary logic you wish with a little care to deal with race conditions. With good physical design the printed circuit board could be universal, requiring a single step and repeat pattern for a crossbar line. Optocoupler output could be implemented with discreet jumpers making the levels of logic explicit and field modifiable. Spare logic function crossbar lines could be populated in anticipation of requirements. There are lots of diodes in this design but their resource demands are small and they allow you to reduce the optocoupler count profoundly over an all optocoupler design.

The generic approach to this hardwired random logic engine can be understood with minimal technology specific detail such that documentation could actually be etched right on the printed circuit board. Clever (perhaps truth table) symbology could make it adequate for an extra-terrestrial to understand with brief study.

If you have a pick and place device, the PCB diode placement can be less labor intensive than a by-hand process. Crossbar encode/decode logic is well established in legacy keyboard designs.

____________

thewildotter

Thanks for sharing this with us. It's fun to play around with these design problems.

I'm curious. If you have 9 dampers now, does your motor have contacts for 9 speeds or still 6?

One thing you might want to think about is flexibility, not having a major redesign if the number of dampers changes, for example.

Explain to the dumb ole Mech Engr why a passive voltage divider won't work with your LM3914. In the sketch below, S1 through Sn are dry contacts from each zone. Then the active 1/R_1-n's sum and divide with R_b to provide 0 - 5v to pin 5 SIG. You might want the R's to be trimmable so some zones could be biased relative to the others.

Can the unused LM3914 outputs be tied together so your fan doesn't turn off on overrange? If this would work, you might want jumpers across S_{1 -} S_n so the thing could be tuned from the control panel.

Ho are you going to control the compressor for the broad demands?

ANDY, the compressor is actually an air to water heat pump that feeds either hot (during winter) or cold water (during summer) into a tank. The airhandler takes either cold or hot water, when it wants and how much it wants.

The heat pump is only controlled by the tank temperature.

Normal A.C. units have a evaporator in de airhandler which would freeze up if not enough air passes over that coil.

That is a great way to do it I installed Ground source until last year, I think 78 is long enough to work out there, I'm now teaching HRA, in a local college. I like all the answers you received on the controls. Now you are going to have trouble deciding which one to use.

Use a small PLC and that way you future proof the design because the PLC manufacturer looks after the upgrade path and backwards compatibility.

I have seen the path where people use Arduino or Raspberry Pie but get someone called out who has no idea of them and you are asking for trouble, no stick with a small PLC and if you like you can have a small operator interface display for trouble shooting.

Does the fan need to be controlled directly by the dampers or can it be controlled by a static pressure sensor in the duct? If you can just maintain constant static pressure by varying fan speed it should work for any combination of open or closed dampers.

Paulnederland,

Your reply and thinking behind it deserve accolades. I am fully in favor of anything we can do to reduce energy demand while keeping the overall cost reasonable. I wonder, however, if you have access to the air handler's fan curves. I believe these are not a single graph, but a family of them to relate current, air flow, and pressure drop. Thus your reliance on the ECM control's ability to determine air flow is questionable.

It is a relatively trivial problem for HVAC contractors to determine the proper duct size for a given air flow, length, and pressure drop. The system can be designed with 9 individual ducts from a single plenum, which gives identical air flow to each open register at the designed pressure in the plenum. Or you can design with a somewhat oversized supply duct and branches off it so the pressure drop down the length of the supply duct is relatively small compared to the pressure drop down the branches.

With either of these approaches, if you are measuring the differential pressure between the supply air into the air handler and the discharge air at a nearby point on the plenum or duct, you are getting the desired air flow through the system to each open register.

You raise the valid concerns regarding dirty filters, closed doors, open windows, and so on. Windows have to be ignored because when they are open the system has lost efficiency. Doors can be ignored if you have adequately sized return air ducts from each room. Dirty filters can be ignored if you are taking your supply air pressure from a location after the filter and you have planned on a filter that is adequate for the system (not cheap 1" ones), and you use a dirty filter alarm.

Another post mentioned the accuracy and reliability of a Magnehelic type control, and I can endorse that.

I suggest that the above approaches are more than adequate for your system, and follow the KISS principle.

--JMM

Thank ALL for the inputs, suggestions and good guidance. Some of the sub-threads went into HVAC design issues and that was also helpful.

I eventually decided to go with the approach that BRAVE SIR ROBERT suggested - after I found a cheap and simple 8 channel voltage comparator with the adjustment capabilities that will enable stable controls and flexibility in case I expand the system. I needed only 6 channels. The spare 2 channels will be used for temperature control. I just need to build a simple voltage divider. My initial solution, using LM3914 would have worked but the standard 8 channel voltage comparator allows me to exercise my laziness; no soldering needed. The LM3914 would have had some undesirable limitations too.

Today, I get delivery of the air handler and I promise that after installation, I will report back on the findings. Might be a while.

Thanks again for all the help. Great people, great website, great help.

https://cr4.globalspec.com/member/6648/Brave-Sir-Robin