@cathoderay I indeed use what you say. I have £5 ESP wifi boards hiding in corners in various rooms (this is needed to keep the wiring to the sensors local and short  - to avoid spiders-web cabling messes) , with sensors taped to various pipes. That gives me the option to monitor literally every radiator if I want to. I haven't actually done that, but monitoring about 5 or 6 key ones  has allowed me to see what is going on much better and make balancing changes. will be coming back to this once it goes cold again!

If you only wanted to monitor pipe temps at a central distribution point, say an airing cupboard, then you could dispense with the remote ESP's and wire to a common collecting device. which could probably be a PC, but I've never looked into this so google is your friend.

I do have a HA PC, which serves as the collecting, aggregating and displaying platform. I'm comfortable with both the ESP and HA platforms (they are well integrated), and I have enough ropey DIY electronics capability to wire up  the ESP's to the DS18B20's. Early days I wrecked one with my awful soldering but at that price trashing one whilst learning is no big deal.

I also have an ESP board that is dedicated to modbus monitoring of a) SDM120M heat-pump specific electricity meter and b) the heat pump itself. Again that data goes into the same HA. so I have everything I need locally. I can also push to cloud for public consumption without exposing my devices, here is my stats : https://emoncms.org/app/view?name=WoburnSandsSamsung16kw&readkey=536f3634aa60c0be3d6e00f1c6128e7a

my experience of the reliability, responsiveness and accuracy of the DS18B20's that I've bought from a variety of random suppliers has been excellent. IMO we're not after revenue grade metering here just "good enough to see what's going on". there are other people who are experts on metrology I do not claim to be one.

Given that you dislike HA , you could do worse than take a look at openenergy monitor emonHP. https://shop.openenergymonitor.com/level-3-heat-pump-monitoring-bundle-emonhp/ will give you what you need as a paid for supported solution , ping the guys there for more explanation what bits you need, glyn hudson is very helpful (that is same guy who authored the "how to install your own heat pump" youtube, which I used as a reference when planning my own installation) . There is masses of discussion on their forums. They They also open source their code so if you felt keen, you might be able to build your own using their platform.

HTH

@iancalderbank - thanks as ever for your helpful detailed response. I might yet cave in and end up using either ESP (electrical) or the emonHP systems, if I really have hit a modbus dead end. I was rather surprised to find the internet isn't awash with modbus based methods, given modbus itself is so ubiquitous. I have also, as mentioned elsewhere, considered swapping my non-modbus heat pump dedicated SDM120 for the modbus enabled version, but as I know the relationship between the Midea reported kWh use and the SDM120 reported use, I can for now just add a correction factor.

A thought on water/pipe temp monitoring: if what we are after is delta t then two DS18B20 sensors with the same or very close error will in fact give the correct delta t. I'm assuming the errors as such are fixed not random, ie they are reliable (always give same reading for same state), even if not valid (always give the right reading for the state), and can be measured using @derek-m's method, put them in hot water at a known temp. Take a pack of ten DS18B20 sensors, pair them off as closely as possible, and you have the making of a good enough temperature monitoring set up. 

@iancalderbank 

Hi Ian,

Copper Heat Conduction?

My System was built with a large mass of unnecessary copper pipes , heat pumps buffers etc.

Copper is an excellent heat conductor.

I want to measure a specific Water temperature , in a specific pipe, not the Temperature of , the  common

 copper structure..

The more copper , the more interference is transmitted through the copper.

The interference is manifested in an electrical signal with  AC and DC components.

Clamps

The Clamped K type Thermocouples can be instantly moved to another location but are inaccurate with a better high frequency response.

The 18B20 are accurate , but, cannot be easily moved, and have a poor high frequency response .

ian

@iancalderbank 

hi ian,

Could I better present my measurement problem with a specific example.

I have a Heat Pump with a Heat Exchanger dividing the Water circuit into two water circuits.

I need to measure the Water Temperatures at the Two input ports and the two output ports.

However, all four ports are physically connected to the Steel copper chassis of the heat exchanger.

How can I measure the Water Temperatures, and not the chassis temperature,  at all four ports?

ian

ps. This has nothing to do with K type sensors or !8B20 Sensors.

I bought the K type sensors because of the ease of moving , clamping.

I will take your advice and try my 18B20 's by cementing the sensors to the pipes.

Many thanks.

@derek-m 

What is Cycle -Time ?

My definition of Cycle -Time is the time between Power pulses to the Compressor.

On my system, with Weather compensation the Power pulse from the compressor is , for example

While the resulting Temperature at the Input to the Heat exchanger is, for the same example

I have measured the Cycle-time as the time between The Power Pulses , in the example 15 minutes .

However, Looking at the following Heat Pump Temperature Graph , for example

The Cycle time would be in excess of 4 hours!

Cycle Time at the Heat Pump output

The Cycle time at the heat pump, as displayed on the front panel Water Temperature display ,  has a period identical to the period between Compressor Power pulses.

Cycle Time at the Heat Exchanger input

The Temperature at the input to the Heat Exchanger is approximately the integral of the Temperature at the Heat Pump.

My System , with a 50 l buffer tank has a far larger  time constant than the average Heat Pump.

The greatly increased Time constant is shown on the following example

  The Buffer is working !

How would you define the Cycle-Time?

ian

@derek-m 

Hi Derek,

Is it meaningful to ask for the Temperature of any point in a Heat Pump ?

I have been pondering the Value and meaning of the Temperatures on my Heat Pump.

The Temperature at the input to my Heat Exchanger is continuously changing.............

What is the "Temperature" at this point ?

The "Temperature" can only be expressed as a dynamic variable, not a static variable.

The "Temperature" can, also, only be measured in the Water Column with a sensor that has, itself , a negligible affect on the Temperature measured.

Sorry, an electrician speaking again !

By how much does the effect of the instrument , and it's location affect the measurement itself?

Model ?

The "Temperature" can only be expressed as a dynamic variable, surely only a dynamic model would be appropriate. not a static model..

ian

@iantelescope 

To me the cycle time is the period from the compressor starting, running, stopping and starting again.

The first graph would appear to show this to be at a frequency of approximately 10 minutes.

I assume the first graph and the second graph were produced on the 22nd August this year. What was the outside temperature on that day?

As far as I can interpret from the graphs, the compressor is started and runs for approximately 5 mins then stops. After a further 5 minutes the compressor is restarted.

Graph 2, I assume from the label, shows the LWT from the heat pump. Do you have a similar graph for the RWT?

As can be seen from the graph, after starting the compressor the heat pump commences raising the temperature of the primary circuit water from 20C to approximately 33C, but does more in a sawtooth fashion rather than a smooth curve. From what I can ascertain from the limited information supplied, there could be a number of reasons why the heat pump would operate in this manner.

1) The heat pump is obviously producing more thermal energy than can be dissipated into the heating system at a sufficient rate. A temperature sensor within the heat pump therefore senses that the LWT is being heated too quickly and stops the compressor to allow the produced thermal energy to be absorbed by the system. Once the temperature within the heat pump has fallen, the compressor is restarted.

2) The flow rate in the primary circuit is inadequate to transfer the thermal energy rapidly enough.

3) The flow rate around the secondary system is insufficient to transfer the thermal energy from the PHE, and hence the primary flow warms up too quickly.

4) The primary flow is mainly going through the buffer tank rather than the PHE, so once the buffer tank is up to temperature the thermal loading is reduced to whatever flow is going through the PHE.

5) The secondary flow is inadequate due to pump speed setting or closed TRV's.

6) The primary and secondary water pumps are not always running when required to be.

It could actually be a combination of any of the above.

The third and fourth graphs appear to indicate that you have some form of on - off thermostat within your system, which is disabling the heat pump once it has reached the temperature setting. As the room temperature falls to the point where the thermostat once more enables the heat pump, the heating system commences operation again.

I would suggest that you closely monitor pump operation and temperatures around your system, and then post the results for analysis.

Also confirm or deny any assumptions I have made.

@iantelescope

From the point of operation and control of a heat pump, it is more important to obtain the average temperature rather any slight changes that may occur during normal operation.

Sensors are unlikely to affect the actual medium, but can give incorrect measurements if not correctly sited and installed. The closer a sensor is placed to the heat source, the more likely that fluctuations can be measured because full mixing has not occurred.

You can try to create a dynamic model if you wish, but how are you going to cause the required variations to temperature and flow etc.? You already have a dynamic system, it is called Central Heating.

@derek-m

Understood!

I've been thinking about the problem of assessing quantitatively the effect of on/off heating (eg when you switch it off during daytime in an effort to save energy, which it is generally held (and I have no reason to doubt) doesn't work, albeit that I have never seen it proven either experimentally or by modelling.

I think I have an idea of how it can be modelled.  I don't think you actually need to know very much about the heat pump itself, its more about the house characteristics.

If I get some 'clear thinking' time over the next days/couple of weeks I will develop this and post.

Posted by: @jamespa

↑

@derek-m

 

Understood!

I've been thinking about the problem of assessing quantitatively the effect of on/off heating (eg when you switch it off during daytime in an effort to save energy, which it is generally held (and I have no reason to doubt) doesn't work, albeit that I have never seen it proven either experimentally or by modelling.

I think I have an idea of how it can be modelled.  I don't think you actually need to know very much about the heat pump itself, its more about the house characteristics.

If I get some 'clear thinking' time over the next days/couple of weeks I will develop this and post.

Hi James.

I am in the process of tidying up the model that I created some time ago, which takes into account some of the property characteristics as well as those of the heat pump. Whilst most of the calculations are straightforward linear regression to extrapolate intermediate data from that provided by Mitsubishi, the problems arise in deciding how to deal with instances when heat demand is greater than the supply, or the heat supply is greater than the demand, which in turn causes cycling to occur.

My other problem is that 'she who must be obeyed' keeps deciding that I need to perform 'urgent' tasks, take the other day for example, she decided that we needed to go out gallivanting, having come up with the lame excuse that we should celebrate our 20th Wedding Anniversary. I am at the moment writing a computer program to analyse 'female logic', and I assess that in about 10 years time when it has been completed, they should have built a computer 'large' enough to run it. 😋 

@derek-m 

Many Many thanks Derek.

My Confidence in my Heat Pump Installation , measurement and design continues to fall......exponentially!

Automatic Flow control.

1) My Samsung Heat Pump is capable of controlling the Water Flow rate in the primary, Heat Pump water Circuit.

2) The primary Water flow rate is controlled by  Pulse Width Modulating the Primary Grundfos motor.  

3)  The PWM motor signal is set by a set of parameters  within "Field Bits" 4051  .

4) The bits within field bit 4051 control , Delta_T , Depth of modulation ( 70% or 100%), and a dynamic control factor.

5) No signal is output from the Control board, and , I have been unable to get a replacement board .

6) I have tested the Grundfoss PWM motor with a PWM signal generated by an Arduino.

7) The PWM signal from the Arduino and the Grundfos motors work perfectly.

8) I am trying to produce a , hopefully, temporary replacement , with the PWM signal coming from the Arduino.

9) MY Question is therfore

How do I measure the Delta_T across the ports of my Heat Exchanger?

All of the temperature measurements are influenced by the copper/steel mass of the Heat Exchanger itself.

I want to measure Delta_T across the Heat Pump ports  not the Temperature of the Heat Exchanger Copper/Steel chassis!

I will have to get a replacement Control Board !

ian

@derek-m 

Hi Derek,

Please have a look at the Water Temperatures captured across my Heat Exchanger input ports.

The Delta_t Temperatures HP(out) - HP(return) were measured by a pair of K type Thermocouples mounted on clamps.

The Delta_T Temperature across the incoming ports is less than 2 C.

The flow rate during this measurement was 9 lpm .

Given that the Water power = Specific_Heat x Delta_T x flow rate

the Delta_T x flow rate will be constant, K_dfl   for a given Water Power.

With a Minimum Flow rate of perhaps 8 lpm and a Maximum flow rate of 16 lpm the Maximum Delta_T would be

Maximum Delta_T = K_dfl/(minimum Flow rate)  = 2 X 9 / ( 8 ) =  2.15 C !

and the minimum Delta_T would be

Minimum Delta_T  = K_dfl( Maximum Flow rate )  = 2  x 9/16  = 1.125 C !

The  range  of Delta_ T , 2.15 - 1.125 = 1 C is , surely, far too small to allow any control.

The Temperature signals also have a poor signal to noise ratio.

A Delta_T of 5 C will be unobtainable.!

Can I therefore control Delta_T by controlling the Flow Rate over such a small range?

The French Samsung Engineer said that I would encounter this problem.

DS 18B20 verses the K type

I have also used two DS18B20's to measure  Delta_T, the results being shown in the fourth graph , named Delta_T Heat Exchanger,C.

With the Delta_T shown in the fourth graph named Delta_T Heat_Exchanger ,C.

Notice that the Temperatures are much smoother than the K type reflecting both the effect of the water buffer and the limited frequency response of the DS18B20.

The other problem was the, twice a day, intermittent crashes to 0 , -60 C and more frequently -127 C with an example caught shown on the graph.

Temperature dependent Resistors

I will try Clamp type Temperature Dependent resistors for comparison with the DS18B20 and the K type Thermocouples.

ian