Heatmiser heating system issue. Online/on-site(£) help needed. Oscilloscope?

[tolklu]

Hi everyone

We have a little problem with a heating system on one of the projects (quite big house in south Kensington) and we are looking for someone who might be able to help.
I'll try to make it as short as possible.

Heating system is based on heatmiser touch v2 thermostats sitting around the house. They are linked with data line/modbus which goes back directly into Crestron processor (whole house is controlled by Crestron). There are 2 loops. About ~20 thermostats on each loop.

The system has been designed by one company, then all the wiring and thermostats installation have been done by me, and then it was taken over by AV company which plugged it into their processor and did all the programming.

Below issue described by AV company:
“We are seeing heatmiser room stats randomly losing comms on the Modbus network. We have two independent loops from two Crestron processors.
Crestron Processor 1 is dealing with the 1st to 3rd floor
Crestron Processor 2 is dealing with Ground and basement.

We have seen the upper floors being mainly stable with minimal issues, which we have contributed to site installation conditions of power going on and off and commissioning.
However, for processor 2 we are seeing at least 2 – 3 times a week issues and the only way to resolve is to power down the heatmiser stats and repower, in a very few circumstances we had to reboot the Crestron processor as well.
Around a month ago we switched the upper floors and ground / basement to the other respective processors and then issue remain with ground and basement, this test I believe rules out the Crestron hardware. The Crestron module is a module that we have successfully used on other projects and has been working for 8-12 months without any known or reported issues.”

It's almost accurate.
For now the system is kind of stable, but is constantly reporting some communication issues and becomes unresponsive once every 2 weeks or so, but then goes back to normal operation (after few hours I guess).

During commissioning we have seen some thermostat not responding to Crestron commands. I have spent a lot of time on site with AV engineer trying to work it out, but it was clearly random, as one day it was one bunch o thermostats and another day it was the other bunch. No rule. We have tried replacing some thermostat for new ones, but that didn't make any difference.

At the moment, AV company have said that they don't care anymore and won't be participating it attempts to resolve this. In their opinion, it's electro magnetic interference between data line and mains cables. Yes, it's one of the possibilities, but very hard to confirm/rule out.

We have obviously kept the separation and used 2 core screened belden cable recommended by heatmiser. There is confirmed continuity of the screen on both loops. Screen is grounded.

We have reached out to heatmiser support, and they have asked to check voltage on modbus line when disconnected from processor. Readings below:

Mod 1 (downstairs)
DC
A to B 0.472
A/B to Ground/screen 0.008V
AC
A to B 0V
A/B to Ground/Screen 0V

Mod 2 (upstairs)
DC
A to B 0.241V
A/B to Ground/screen 0.015V
AC
A to B 0V
A/B to Ground/screen 0V

I have discovered that on Mod 2 0.472V turns into 3.9V as soon as the last stat is disconnected (front screen removed or cables psychically removed from the back part of it) from the comms line. Of course, it's the one with termination resistor enabled. Removing any other stat along the line doesn't affect the voltage at all.

I have tried that with Mod 1 and the voltage has increased from 0.241 to 0.441. Again, only when the last stat was removed.

According to heatmiser there's nothing interesting there, and they have asked for tests with oscilloscope and this is the reason why I'm writing to you.

Main contractor have been trying for weeks to get hold of the right engineer, but no luck. I have tried all my contacts and nothing as well. I know we can rent an oscilloscope, but it's not only about that.

Tensions are very high at this point and we want to do it the best possible way (whatever that means...)

So maybe there's someone out there with an oscilloscope and vast experience with BMS systems or with heating systems based on heatmiser or with ANYTHING from this list...

Any help will be highly appreciated.
Thank you.

 

[UNG]

Initially looking at this from an analytical perspective do you have any diary / log of these events while you suggest they are random over a period of time there may be some pattern to the loss of comms involving other installed equipment within the property or an external influence in the local area

While trying to find this fault timing could be an issue when you say

For now the system is kind of stable, but is constantly reporting some communication issues and becomes unresponsive once every 2 weeks or so, but then goes back to normal operation (after few hours I guess).

So what else is happening in the building when you get the loss of comms, are any other parts of the BMS affected

Thinking around the problem you have
Have you done any power quality analysis of the electricity supply
Does the BMS have an event log
Are all the cables continuous runs between devices with no joints

 

[plugsandsparks]

You need a Modbus protocol analyser rather than an oscilloscope. Its a software package which converts the modbus signals to meaningfull data so you can see what is being transmitted on the bus.

A couple of things you may wish to consider.

Modbus is pretty bullet proof, would be amazed if Em interference in a house would have an impact.
If this has not already been done and it may sound obvious but i would double or triple check both the AB wiring to ensure it is not crossed anywhere and check each thermostats Modbus address. Write them down as you go from room to room, each thermostat needs a unique address from 1-32 per modbus line.

After that get an analyser and leave it connected gathering data and wait for the "moment"

hope this helps.

 

[tolklu]

Thank you very much for your comments

Initially looking at this from an analytical perspective do you have any diary / log of these events while you suggest they are random over a period of time there may be some pattern to the loss of comms involving other installed equipment within the property or an external influence in the local area

Unfortunately, only the AV company has access to it, and they are not wiling to cooperate. We could probably "force" them to make it available, but I don't see it representing a significant value.
Owners are living in that house for the last 6 months, so use of equipment is quite repetitive and predictable. When ever I'm there, I can observe the situation on Crestron touch screens located around the house. When something goes south, there's a notification popping up on the screen informing that there is a comms error and this is happening all the time. Let's say if I watch it for 30 minutes, I would see it about 6 times. Random intervals. Sometimes it's displayed for 5s and sometimes for 20s.
When the whole system goes down (becomes unresponsive) screens are just displaying values of 0.00 as they are not getting any information. As I have written before, that happens quite rarely since clients moved in. They are frequently checking the screens and immediately notifying us if something goes wrong. Then I'm questioning them about activity in the house, but they are using the same equipment every day (and I'm looking really hard for anything to blame)

While trying to find this fault timing could be an issue when you say

So what else is happening in the building when you get the loss of comms, are any other parts of the BMS affected

As you can imagine there are loads of (sensitive) equipment inside that property and that's the only issue.

Thinking around the problem you have
Have you done any power quality analysis of the electricity supply

No. I don't see how that could affect the data side of the thermostats, but correct me if I'm wrong.

Are all the cables continuous runs between devices with no joints

There are only 2 joints on Mod 1 as 2 thermostats have been removed at some point. Connections are easily accessible as those thermostats have been located in the same cabinet as few other thermostats. Connections are sound, inside (now empty) backboxes, screen joints etc.

 

[tolklu]

You need a Modbus protocol analyser rather than an oscilloscope. Its a software package which converts the modbus signals to meaningfull data so you can see what is being transmitted on the bus.

Sounds interesting. Can you elaborate on this a little more, please

A couple of things you may wish to consider.

Modbus is pretty bullet proof, would be amazed if Em interference in a house would have an impact.

If this has not already been done and it may sound obvious but i would double or triple check both the AB wiring to ensure it is not crossed anywhere

checked multiple times by me and confirmed by other person

and check each thermostats Modbus address. Write them down as you go from room to room, each thermostat needs a unique address from 1-32 per modbus line.

Again, checked multiple times. What's more, during commissioning I could watch them on AV software in real time with ID numbers displayed and being in order.

 

[plugsandsparks]

Sounds interesting. Can you elaborate on this a little more, please



checked multiple times by me and confirmed by other person

Again, checked multiple times. What's more, during commissioning I could watch them on AV software in real time with ID numbers displayed and being in order.

Handheld Serial Data Analyzer with RS232 & RS485 inputs, Modbus & timestamping

A 'multimeter' for troubleshooting serial data - this handheld device uniquely accepts RS232, RS485/422 and CMOS/TTL level inputs

www.new-forest-electronics.co.uk

If you are holding the baby i would double check with both Heatmiser and Creston tech support that these products are actually compatible over Modbus together, check fireware of both Thermostats and Crestron box. I have used Heatmiser products over wireless and over the internet and i would make no assumptions. Most AV guys i have met are plug and play merchants, if it does not work , its usually "power surges", "interference" "black death" lol

 

[tolklu]

Thank you @plugsandsparks
This looks like a very useful device. I'm already in the process of getting hold of it.

Unfortunately some parties involved are getting obsessed with an oscilloscope test. I will have to tick that box.
Does anyone know someone with an oscilloscope in their van?
Or how to get hold of such a person?

 

[plugsandsparks]

Thank you @plugsandsparks
This looks like a very useful device. I'm already in the process of getting hold of it.

Unfortunately some parties involved are getting obsessed with an oscilloscope test. I will have to tick that box.
Does anyone know someone with an oscilloscope in their van?
Or how to get hold of such a person?

I use one of these
PicoScope 2204A 2 Channel 10 MHz USB Oscilloscope : Amazon.co.uk: Business, Industry & Science - https://www.amazon.co.uk/Technology-PicoScope-2204A-Oscilloscope-channels/dp/B00GZMRZ3M

I have only really used it on unknown signals what i am trying to figure out, i guess "they" are looking for a nice clean set of binaries with no noise, good luck, lol
You will always see noise, questions is how much is too much ?
Thats why an analyser will show data errors if in fact noise is an issue.
Dont fall into the trap of putting a scope on, seeing non perfect waveforms and concluding EM interference, you need to see it in data errors, hope this helps and NO i am nowhere near London.....
If you do get an analyser you need to look for transmits and timely responses, thats where i would go first.
Then, when you leave, cos it will never do it whilst you are watching, get client to note exact time fault happens, go back to data and look at the device that initiated a command and see what happened

 

[baldelectrician]

Have any thermostats been swapped from one zone to another to see if the fault shifts.

 

[tolklu]

i guess "they" are looking for a nice clean set of binaries with no noise, good luck, lol

yeeeeeeep

If you do get an analyser you need to look for transmits and timely responses, thats where i would go first.
Then, when you leave, cos it will never do it whilst you are watching, get client to note exact time fault happens, go back to data and look at the device that initiated a command and see what happened

that's very helpful, thank you

 

[tolklu]

Have any thermostats been swapped from one zone to another to see if the fault shifts.

Yes. It doesn't shift.
More likely to shift if you reboot the system.

 

[telectrix]

Teyjccnjddjgfjhdgnjjf gfdjc jggyr

 

[tolklu]

Teyjccnjddjgfjhdgnjjf gfdjc jggyr

I believe we have tried that at some point. Didn't help either.

 

[saifkjo]

Hi everyone

We have a little problem with a heating system on one of the projects (quite big house in south Kensington) and we are looking for someone who might be able to help.
I'll try to make it as short as possible.

Heating system is based on heatmiser touch v2 thermostats sitting around the house. They are linked with data line/modbus which goes back directly into Crestron processor (whole house is controlled by Crestron). There are 2 loops. About ~20 thermostats on each loop.

The system has been designed by one company, then all the wiring and thermostats installation have been done by me, and then it was taken over by AV company which plugged it into their processor and did all the programming.

Below issue described by AV company:
“We are seeing heatmiser room stats randomly losing comms on the Modbus network. We have two independent loops from two Crestron processors.
Crestron Processor 1 is dealing with the 1st to 3rd floor
Crestron Processor 2 is dealing with Ground and basement.

We have seen the upper floors being mainly stable with minimal issues, which we have contributed to site installation conditions of power going on and off and commissioning.
However, for processor 2 we are seeing at least 2 – 3 times a week issues and the only way to resolve is to power down the heatmiser stats and repower, in a very few circumstances we had to reboot the Crestron processor as well.
Around a month ago we switched the upper floors and ground / basement to the other respective processors and then issue remain with ground and basement, this test I believe rules out the Crestron hardware. The Crestron module is a module that we have successfully used on other projects and has been working for 8-12 months without any known or reported issues.”

It's almost accurate.
For now the system is kind of stable, but is constantly reporting some communication issues and becomes unresponsive once every 2 weeks or so, but then goes back to normal operation (after few hours I guess).

During commissioning we have seen some thermostat not responding to Crestron commands. I have spent a lot of time on site with AV engineer trying to work it out, but it was clearly random, as one day it was one bunch o thermostats and another day it was the other bunch. No rule. We have tried replacing some thermostat for new ones, but that didn't make any difference.

At the moment, AV company have said that they don't care anymore and won't be participating it attempts to resolve this. In their opinion, it's electro magnetic interference between data line and mains cables. Yes, it's one of the possibilities, but very hard to confirm/rule out.

We have obviously kept the separation and used 2 core screened belden cable recommended by heatmiser. There is confirmed continuity of the screen on both loops. Screen is grounded.

We have reached out to heatmiser support, and they have asked to check voltage on modbus line when disconnected from processor. Readings below:

Mod 1 (downstairs)
DC
A to B 0.472
A/B to Ground/screen 0.008V
AC
A to B 0V
A/B to Ground/Screen 0V

Mod 2 (upstairs)
DC
A to B 0.241V
A/B to Ground/screen 0.015V
AC
A to B 0V
A/B to Ground/screen 0V

I have discovered that on Mod 2 0.472V turns into 3.9V as soon as the last stat is disconnected (front screen removed or cables psychically removed from the back part of it) from the comms line. Of course, it's the one with termination resistor enabled. Removing any other stat along the line doesn't affect the voltage at all.

I have tried that with Mod 1 and the voltage has increased from 0.241 to 0.441. Again, only when the last stat was removed.

According to heatmiser there's nothing interesting there, and they have asked for tests with oscilloscope and this is the reason why I'm writing to you.

Main contractor have been trying for weeks to get hold of the right engineer, but no luck. I have tried all my contacts and nothing as well. I know we can rent an oscilloscope, but it's not only about that.

Tensions are very high at this point and we want to do it the best possible way (whatever that means...)

So maybe there's someone out there with an oscilloscope and vast experience with BMS systems or with heating systems based on heatmiser or with ANYTHING from this list...

Any help will be highly appreciated.
Thank you.

Hello, does the house have a PV system (photovoltaic electricity/energy generation system)?

 

[tolklu]

Hello, does the house have a PV system (photovoltaic electricity/energy generation system)?

Hi. No, it doesn't.