IOT heat pump controller - updated

Acuario · 12-03-17, 03:50 AM

I'm going to start a new thread for this rather than continuing the old thread so they don't get mixed up. The old thread is still relevant as this is really an evolution of the project and I suggest reading that first. IOT heat pump controller Here I'll detail the updates from the previous version.

The hardware and software are both installed and running on my heat pump. This is a live on-going project.

Hardware
I have now designed and built my own 'peripheral' board for the controller rather than using an assortment of modules glued together with veroboard. The full schematic is shown below (and in the attached PDF). There were only some very minor changes in the hardware from the original design.

The peripheral board is connected to the main controller with 4 wires, +5V, 0V and I2C control lines SCL and SDA. The board has connections for the 4x temperature sensors, 2 x relay outputs for the fan and reversing valve and an output for the solid state relay used to control the compressor. The extra outputs from the I/O expander are available for use (I'm using one to control a circulating pump). There are also I2C outputs available to connect extra peripherals such as (in my case) a BME280 pressure/temperature/humidity detector.

Software
There are significant changes in the software and this is the area that is most relevant for continuing development and experimentation.

The current version continues with the traditional defrost algorithm, i.e. if the evaporator temperature is below X for Y time then do a defrost. I have (and still am) experimented with different defrost processes. I have been trying differential pressure monitoring but despite some promising initial results (restricting the airflow through the evaporator using plastic bags) the few real life scenarios have not given similar results. I'm now investigating relationships between temperature and humidity. As we don't get many occasions when the evaporator freezes up this is a lengthy process...

Additionally in the latest version is the inclusion of weather data to predict heating requirements. This reads a data feed from openweathermap.org

The current version of the software is included in the zip file in this post. I use Visual Studio with the VisualMicro addon but the software should be able to be used in the Arduino IDE.

Acuario

Acuario · 01-11-21, 02:17 AM

It's been a while since I updated this post but development certainly hasn't stopped...

I've made significant changes to the heatpump controller by adding in predictive weather heating/cooling (there's another thread on this) and now an Android app.

I can now reliably control the system remotely via the app. Previously I used Telegram to send messages to the thermostat that then controlled the heatpump but this wasn't very reliable so I wrote my own app (that I can use to control the 3 heatpumps I have).

If anyone is interested, I used Basic for Android (www.b4x.com) to write the app. It uses MQTT via CloudMQTT to talk directly to the heatpump controller.

Previously I controlled the heatpump via the remote thermostat but I decided to move the control back to the heatpump itself. The problem I had previously using Telegram is that it only works using SSL and this uses too much memory on the esp8266. Moving to MQTT has solved this memory problem so I could move the communications back to the heatpump.

Acuario

jeff5may · 01-16-21, 12:05 PM

I've settled on the Trane American standard defrost model for control. Works like magic. The real one starts out with model number cnt5001 and has been revised over the years. Current model number is cnt05875. Has 3 relay outputs: outdoor fan disable, reversing valve cool mode, and indoor backup heat call.

The control works off of two thermometer sensors: outdoor air temperature (OAT) and evaporator temperature. Defrost initiation and termination are dependent on the following conditions:

Initiation happens after the evap temp reaches (OAT - 25 degF) and a timer expires if you wish.

Termination depends on OAT:

OAT*****ET (degF)
>22******47
10-22****OAT+25
6-10*****35
<6*******35 or 12 minutes

There's an option 2 in the board, for rigs that are touchy around 30 degF OAT:

>30*****47
6-30****70
<6******35 or 12 minutes

I gave up on the universal heat pump control idea a few years ago. Too many people had too many different expectations, and were very centered on their own specific rig. Air to air, air to water, and water to water were the 3 main groups of people, and nobody wanted to collaborate with anyone. I couldn't find anyone to take charge of their type of system who had coding or platform integration skills, so what happened in the end was the active members all went their own separate ways.

Since then, I've found that there are a lot of preassembled controls on the market, preconfigured for all kinds of different rigs. Dirt cheap to super expensive, pick your poison. Just go on Amazon or aliexpress and search for "universal (whatever you have) control". Example: universal ac replacement control

Also, there's a super affordable, robust second hand market for r22 equipment in America. The HVAC service guy swaps out homeowner with a r410a system, and the homeowner stashes it in storage until cleanup day. I could go on Facebook marketplace today and find a 2 to 3 ton outdoor unit for 100 bucks.

Acuario · 01-16-21, 12:57 PM

Interesting information about the defrost, I'll study it a bit more - I'm still playing with different ideas rather than just temperatures - the latest I'm investigating is using a small fan with an infra red optical link to detect the flow rate of air through the evaporator on the basis that less flow = more ice. Just waiting for the IR diodes to arrive.. The drop in pressure I tried before wasn't really reliable enough.

It's a pith the universal controller was abandoned. Yes, there are commercial pre-assembled controls out there but I doubt they give you the control and connectivity a home-made unit does (for me = no fun), for example adding a new method of controlling/monitoring remotely as above would probably be almost impossible with a commercial unit as it can't be modified.

All of my units have come from replacements, I'm currently running 4 home-made systems, 2 on my house, 1 on my greenhouse (in the process of being upgraded to run on solar as a combined solar water/heatpump system) and one in a flat where my mother-in-law lives.

jeff5may · 01-16-21, 01:43 PM

I know. It was difficult for me to let go of the idea. Keeping it simple was the hard part: everyone wanted to upgrade the basic control to tailor it to their personal site. I stuck with an Arduino uno and a little LCD screen with up, down, right, left and enter buttons. Had it working with either 10k ntc thermistors or one wire sensors. That was when I settled on the Trane defrost algorithm for air source outdoor unit control.

01-11-21, 02:17 AM	#2
Acuario Apprentice EcoRenovator Join Date: May 2011 Location: Tortosa, Spain Posts: 221 Thanks: 2 Thanked 81 Times in 46 Posts	It's been a while since I updated this post but development certainly hasn't stopped... I've made significant changes to the heatpump controller by adding in predictive weather heating/cooling (there's another thread on this) and now an Android app. I can now reliably control the system remotely via the app. Previously I used Telegram to send messages to the thermostat that then controlled the heatpump but this wasn't very reliable so I wrote my own app (that I can use to control the 3 heatpumps I have). If anyone is interested, I used Basic for Android (www.b4x.com) to write the app. It uses MQTT via CloudMQTT to talk directly to the heatpump controller. Previously I controlled the heatpump via the remote thermostat but I decided to move the control back to the heatpump itself. The problem I had previously using Telegram is that it only works using SSL and this uses too much memory on the esp8266. Moving to MQTT has solved this memory problem so I could move the communications back to the heatpump. Acuario Attached Thumbnails

01-16-21, 12:05 PM	#3
jeff5may Supreme EcoRenovator Join Date: Jan 2010 Location: elizabethtown, ky, USA Posts: 2,431 Thanks: 431 Thanked 619 Times in 517 Posts	I've settled on the Trane American standard defrost model for control. Works like magic. The real one starts out with model number cnt5001 and has been revised over the years. Current model number is cnt05875. Has 3 relay outputs: outdoor fan disable, reversing valve cool mode, and indoor backup heat call. The control works off of two thermometer sensors: outdoor air temperature (OAT) and evaporator temperature. Defrost initiation and termination are dependent on the following conditions: Initiation happens after the evap temp reaches (OAT - 25 degF) and a timer expires if you wish. Termination depends on OAT: OAT***ET (degF) >22**47 10-22OAT+25 6-10*35 <6***35 or 12 minutes There's an option 2 in the board, for rigs that are touchy around 30 degF OAT: >30*47 6-3070 <6***35 or 12 minutes I gave up on the universal heat pump control idea a few years ago. Too many people had too many different expectations, and were very centered on their own specific rig. Air to air, air to water, and water to water were the 3 main groups of people, and nobody wanted to collaborate with anyone. I couldn't find anyone to take charge of their type of system who had coding or platform integration skills, so what happened in the end was the active members all went their own separate ways. Since then, I've found that there are a lot of preassembled controls on the market, preconfigured for all kinds of different rigs. Dirt cheap to super expensive, pick your poison. Just go on Amazon or aliexpress and search for "universal (whatever you have) control". Example: universal ac replacement control Also, there's a super affordable, robust second hand market for r22 equipment in America. The HVAC service guy swaps out homeowner with a r410a system, and the homeowner stashes it in storage until cleanup day. I could go on Facebook marketplace today and find a 2 to 3 ton outdoor unit for 100 bucks. Last edited by jeff5may; 01-16-21 at 12:44 PM.. Reason: Information*

01-16-21, 12:57 PM	#4
Acuario Apprentice EcoRenovator Join Date: May 2011 Location: Tortosa, Spain Posts: 221 Thanks: 2 Thanked 81 Times in 46 Posts	Interesting information about the defrost, I'll study it a bit more - I'm still playing with different ideas rather than just temperatures - the latest I'm investigating is using a small fan with an infra red optical link to detect the flow rate of air through the evaporator on the basis that less flow = more ice. Just waiting for the IR diodes to arrive.. The drop in pressure I tried before wasn't really reliable enough. It's a pith the universal controller was abandoned. Yes, there are commercial pre-assembled controls out there but I doubt they give you the control and connectivity a home-made unit does (for me = no fun), for example adding a new method of controlling/monitoring remotely as above would probably be almost impossible with a commercial unit as it can't be modified. All of my units have come from replacements, I'm currently running 4 home-made systems, 2 on my house, 1 on my greenhouse (in the process of being upgraded to run on solar as a combined solar water/heatpump system) and one in a flat where my mother-in-law lives.

01-16-21, 01:43 PM	#5
jeff5may Supreme EcoRenovator Join Date: Jan 2010 Location: elizabethtown, ky, USA Posts: 2,431 Thanks: 431 Thanked 619 Times in 517 Posts	I know. It was difficult for me to let go of the idea. Keeping it simple was the hard part: everyone wanted to upgrade the basic control to tailor it to their personal site. I stuck with an Arduino uno and a little LCD screen with up, down, right, left and enter buttons. Had it working with either 10k ntc thermistors or one wire sensors. That was when I settled on the Trane defrost algorithm for air source outdoor unit control.