Posted by: @benson

↑

The biggest mystery for me is why cathoderays modbus alterations don't trigger a shut down, and my modbus alterations do.

I confirm system shutting down after exiting the SERVICEMAN MODE but not shutting down when parameters changed through direct modbus communication.

@benson you are using Home Assistant though, no ? Maybe this is the reason for the different behavior ?

HAPPY NEW YEAR everyone !!

@benson is using an esphome controller which integrates with Home Assistant. The esphome controller uses regular modbus, which is not different then for example using a python script that talks modbus to the controller. 

I had a few people, who use the same esphome configuration, that mentioned this behavior for certain modbus registers. In a specific case, I was not able to reproduce the things that the messenger had, although we where using exactly the same configuration and controller hardware. So I still am under the impression that specific hardware and firmware versions of the wired controller can behave differently regarding this.

Posted by: @mosibi

↑

So I still am under the impression that specific hardware and firmware versions of the wired controller can behave differently regarding this.

Seems to be the only logical explanation. Our ASHP was purchased/installed late 2024, and perhaps Midea decided to 'upgrade' the firmware to make sure that the serviceman changes via modbus behaved in the same way as changes via the control panel at some point.

Posted by: @benson

↑

Our ASHP was purchased/installed late 2024

Mine was installed in early Feb 2022 and may well have left the factory in 2021, meaning quite a long interval between manufacturing dates. The varying emphasis in the wired controller manuals on the turned off behaviour may also indirectly confirm changes were made to the firmware.

If you are stuck with a unit which turns itself off when accessed via modbus, then I think it should be possible to add some code to turn it back on, thereby removing the need to do it manually. There are many many versions of the wired controller manual, this is mine, a scan of the physical copy that came with the unit, and it shows what looks like the main on/off switches in bits 1 and 2. These are single bits in a read/write 16 bit register:

The exact code to turn those registers on (write a value of 1) may need a bit of research, possibly by looking at the code behind the DHW on/off button in HA. If the same code was added to the end of the WCC change code, it should turn the DHW back on after the WCC change has been made.

Posted by: @cathoderay

↑

The exact code to turn those registers on (write a value of 1) may need a bit of research, possibly by looking at the code behind the DHW on/off button in HA. If the same code was added to the end of the WCC change code, it should turn the DHW back on after the WCC change has been made.

My Airwell is installed in October 2022 and I do not know the exact manufacturing date.

Your screenshot of register 0 is valid for all Midea (clone) variants I have seen until today. The differences between the variants in the same refrigeration type (R32 and R290) are not that big, but between R32 and R290 there are quite some differences where mainly the R290 variants have more registers then the R32. If you are curious, you will find the ESPhome configuration I've created for register 0 bit 1 and 2 here and the base configuration and specific model configuration here

Regarding "Power DHW T5S": I have created an automation for myself in Home Assistant, which automatically powers on "Power DHW T5S" when it is turned off for 1 minute. And I only do this when my custom DHW method is not enabled. This way I do not have to add enabling "Power DHW T5S" to on at multiple pieces of logic, but I can purely let this one event based automation handle it.

Posted by: @mosibi

↑

If you are curious, you will find the ESPhome configuration I've created for register 0 bit 1 and 2 here and the base configuration and specific model configuration here

I had a look and it looks complicated to me! There may even be some bitwise operators in there. 

Posted by: @mosibi

↑

Regarding "Power DHW T5S": I have created an automation for myself in Home Assistant, which automatically powers on "Power DHW T5S" when it is turned off for 1 minute. And I only do this when my custom DHW method is not enabled. This way I do not have to add enabling "Power DHW T5S" to on at multiple pieces of logic, but I can purely let this one event based automation handle it.

That's a neat 'catch all/except' solution. I expect @benson may want to look into it a bit more.

Posted by: @cathoderay

↑

That's a neat 'catch all/except' solution. I expect @benson may want to look into it a bit more.

Yes that would be a good solution and the one that Sven (creator of my esphome device) suggested. I don’t want to trigger a tank reheat though whenever the automation runs.

I am however testing the principle as to whether just turning back on something (so central heating back on) is actually sufficient for the ashp to come out of deactivated mode after serviceman and for DHW schedules to kick in. This would be far simpler to implement as an automation. I will know whether this has worked at precisely 1500hrs today which is when my next tank reheat schedule is set for….

Posted by: @cathoderay

↑

I had a look and it looks complicated to me! There may even be some bitwise operators in there. 

Sorry for the offtopic:

The bitwise operators are required where one register contains multiple values. As you probably know the registers are all 16 bits and Midea places 1, 2, 4 and 16 values in one register and in some cases, 2 registers are combined into 1 value. Register 0 is an example of a register that holds 16 values and thus if you for example need to change bit 2 from 0  to 1, you normally would have to store the current state of all the bits and flip the value for bit 2 and then write the complete register at once. With the bitwise operator, that logic is done automatically by only updating value 0x04 (bit zero is 0x00, bit one is 0x02, bit two is 0x04, bit 3 is 0x08, ...) of register 0.

Btw... That's also where my nickname Mosibi comes from: "Most Significant Bit"

Posted by: @mosibi

↑

Btw... That's also where my nickname Mosibi comes from: "Most Significant Bit"

Then you clearly know what you are talking about!

I do find bitwise operations are not exactly intuitive (at least for me). I managed to find a way of reading (but not writing to) these bit registers that treats the registers contents as a binary string, and then extracts the nth 1 or 0 from the right hand end from that string (eg reading the position of the three port two way valve) that works for me, and just importantly I can still understand what it does when I come back to the code months or years later.  

Posted by: @cathoderay

↑

and just importantly I can still understand what it does when I come back to the code months or years later.  

As somebody with 30 years professional experience in computing, being able to understand the code is way more important that squeezing out the last bits of efficiency. (and then you are sitting together with a co-worker, debugging a strange printer error and he, while looking at the code where large parts where in HEX, says: Ah, I already see it!)

Posted by: @cathoderay

↑

Maybe the point is you have to manually turn something on. It doesn't matter whether it is space heating or DHW heating, as long as one is turned on the system remains on. If neither are turned on, the system remains dormant, ie both are off, and will remain so until one or the other is manually turned on.

As it turns out this is correct! All I needed to do was turn the heating back on, which makes things much easier.

My plan now is to set up an automation to run my adjustment script if the IAT runs too low, and to run the 'revert to default WCC' script when it rises sufficiently. I will use the "and if" conditions to avoid my DHW reheat times for any WCC adjustments just so these reheat periods are not interfered with in terms of shut downs. Then, all I need to do is set up a "turn CH back on" automation if it detects that the CH has been off for longer than a short period of time (perhaps a minute, or less).

Posted by: @benson

↑

My plan now is to set up an automation to run my adjustment script if the IAT runs too low, and to run the 'revert to default WCC' script when it rises sufficiently. I will use the "and if" conditions to avoid my DHW reheat times for any WCC adjustments just so these reheat periods are not interfered with

This is what my auto adapt script does. It also covers the situation when the IAT overshoots, when it takes the WCC down a bit. It deals with 'is the DHW heating on' question at the start with a check on the position of the three port two way valve:

# check DHW is NOT running (use three port two way valve position)
on_off = instrument.read_register(129, 0)
valve_pos = "{:08b}".format(on_off)[-5]      # 1 = DHW is on, 0 = DHW is off

# note valve_pos is a string, not an integer, hence need '0' not 0
if valve_pos == '0':

   rest of the code...

This code first reads the register value which comes back as a hex number. Then we have my way of reading a single bit without using bitwise operators. The valve_pos = line converts the hex value to its binary equivalent, right pads with zeroes to eight digits, and then extracts the fifth digit from the right, which is bit 4 (the bit numbers use zero based counting, python slicing uses one based counting). If that value is 1, the DHW heating is on, if it is zero, the DHW heating is off. If it is zero (DHW heating is off), then the rest of the code runs.