News:

;) This forum is the property of Proton software developers

Main Menu

Live conductor temperature measurement

Started by shantanu@india, Mar 29, 2025, 06:10 AM

Previous topic - Next topic

shantanu@india

Hi,
Working on a test setup where upto 5kA at 50Hz(maximum voltage 10V) is being injected into a test conductor for a temperature rise test. I need to have contact type sensors to measure temperature at different points .
Can I use thermocouple/RTD?
I have a feeling that there is a possibility of 50Hz interference resulting in unstable temperature readings.
Can I use DS18B20?
I dont want to use IR sensors for the obvious reason of potential inaccuracy.
Regards
Shantanu

keytapper

Quote from: shantanu@india on Mar 29, 2025, 06:10 AMCan I use DS18B20?
Depending on the range of the measures. I found some time it responded wrong data. So I designed a filter.

Symbol TOLERANCE 30               '  anything as necessary
Dim LastRead As Word
Dim CurrRead As Word
Dim timeOut As Byte
Dim bCounts As Byte

  Repeat
    DelayMS 25                    ' Wait until conversion is complete
    ORead DQ_Pin, 4, [bCounts]    ' Keep reading low pulses until
    Inc timeOut                   ' keep counting for time process
    If timeOut > 100 Then         ' spending too long for a reply?
      CurrRead = $7fffffff        ' assert a non valid value
      Break                       ' giving up
    End If
  Until bCounts <> 0              ' the DS1820 is finished.
  OWrite DQ_Pin, 1, [$CC, $BE]    ' Send a Read ScratchPad command
  ORead DQ_Pin, 2,[CurrRead.LowByte, CurrRead.HighByte]

If Abs(LastRead - CurrRead) < TOLERANCE Then
  LastRead = CurrRead
End If
Ignorance comes with a cost

shantanu@india

Thanks for the idea. Yes filtering is definitely required. I've use shielded cables.
Regards
Shantanu

Fanie

#3
At 5000A you definitely have a short circuit !

You don't mention the expected temperature range ?

18B20 should work fine, and you can track temperature quite accurate with it too, up to 120°C max if I remember right.
I assume the conductor is quite heavy, so temperature will not cool down too quickly, although it may heat up faster than cool down.
A small thermistor will get you faster readings. 
With the 3D printing higher thermistor temperatures has improved, up to around 250°C.

To get a larger output from a thermistor, supply it with a higher voltage than the pic.  The measuring voltage should remain within the pic's A/D range and that of the thermistor of course.

Fanie

You can of course go the thermopile way if you connect the thermocouples in series, but usually only used with higher temperatures.  You will need less in series for a 3V3 pic than a 5V pic  ;)
 
Termokoppels in serie.jpg

At 500°C, a type K thermocouple will output about 20.644 mV.  500 deg is still below melting temperature of most metals, and you can get a nice full scale output of 3V3 if you series 165 type K thermocouples.  No amplifier required, just think of the saving  ;D

Don't connect them in series as it looks, because every 2nd one will cancel the value back to 0V.
Pos must connect to neg for each one.  Pos to neg and neg to pos will give no output.

shantanu@india

Quintessential Fanie with his ever practical suggestions for the past 20 years!! Thanks a lot.
Actually this is a type test setup to inject current into a 5 meter long aluminium conductor piece(used in O/H transmission lines) to raise it's temperature by 40 deg C. The amount of current depends on the conductor. It's a standard 3-phase variac connected to a high current transformer.
Regards
Shantanu

Fanie

Quote5kA at 50Hz(maximum voltage 10V)
Quote5 meter long aluminium conductor piece(used in O/H transmission lines)
You guys really do things differently in India !
Why do you want the temperature to go up 40°C ?
Temperature increase is usually associated with losses, you may end up with less than 10V.
Or do you have a 9V app drawing 5000A and you want to lose the 1V ?  :o

shantanu@india

No maybe you need to inject only 600A/700A to get the requisite temperature rise. The transformer is rated for high current. I also feel that the transformer is overdesigned.The transformer looks like this....
Regards
Shantanu

Fanie

Perhaps you should indicate why you need the higher temperature ?

Here is AI's response - For once I agree with it.
QuoteIn general, higher transformer efficiency leads to lower temperature rise, and vice versa, as more efficient transformers generate less heat. Lower temperature rise also translates to increased overload capability and longer transformer lifespan.
Here's a more detailed explanation:
Efficiency and Temperature Rise:
High Efficiency = Lower Temperature Rise: Efficient transformers are designed to minimize energy losses, resulting in less heat generation.
Low Efficiency = Higher Temperature Rise: Less efficient transformers generate more waste heat, leading to higher operating temperatures.
Temperature Rise and Transformer Performance:
Overload Capability: Transformers with lower temperature rise ratings (e.g., 80°C or 115°C) can handle higher overloads without compromising their lifespan.
Reliability and Lifespan: Cooler-running transformers are more reliable and have a longer lifespan because excessive heat can damage internal components and accelerate aging.
Cooling Requirements: Efficient transformers require less cooling and ventilation because they generate less heat.
Factors Affecting Temperature Rise:
Load: Higher loads generally lead to higher temperatures.
Ambient Temperature: The surrounding temperature also plays a role in the overall operating temperature of the transformer.
Transformer Design: The design of the transformer, including core material, winding materials, and cooling methods, significantly impacts temperature rise.
Temperature Rise Ratings:
C-Rise: Temperature rise rating, often referred to as "C-rise," is the average temperature increase (in Celsius) of the transformer above the ambient temperature when fully loaded.
Common Ratings: Common temperature rise ratings include 80°C, 115°C, and 150°C.
Consequences of Excessive Temperature Rise:
Reduced Efficiency: As the transformer temperature rises, its internal resistance also increases, leading to more power dissipation as heat and a decrease in efficiency.
Insulation Degradation: High temperatures can damage the insulation, potentially leading to insulation breakdown and transformer failure.
Premature Aging: High temperatures accelerate the aging process of transformers, reducing their lifespan.

So if your transformer is overdesigned, it will have a better expected lifespan than one which is just sufficient.  Here in the illegal colony of Sick Africa we frequently have transformers failing because of overload. 
Another factor is of course the cost.  If your design is not extravagantly more expensive it should be a preference.

ken_k

Quote from: Fanie on Mar 29, 2025, 11:57 AMYou can of course go the thermopile way if you connect the thermocouples in series, but usually only used with higher temperatures.  You will need less in series for a 3V3 pic than a 5V pic  ;)
 
Termokoppels in serie.jpg

Thermocouples are wonderful devices.
I have successfully used thermocouples in series. One easy to fall into trap is that the output voltage is relative to the cold junction temperature, this may or may not be a problem depending on the application. One solution is to connect all the cold junctions to a thermal mass at near ambient temperature and measure the temperature of the cold junction using a cheap semiconductor temperature sensor.  The "cold junction" temperature can be used when calculating the hot junction temperature.


ken_k

#10
Quote from: shantanu@india on Mar 29, 2025, 06:10 AMHi,
Working on a test setup where upto 5kA at 50Hz(maximum voltage 10V) is being injected into a test conductor for a temperature rise test. I need to have contact type sensors to measure temperature at different points .
Can I use thermocouple/RTD?
I have a feeling that there is a possibility of 50Hz interference resulting in unstable temperature readings.
Hi
5kA will produce quite a significant magnetic field and induced voltages may indeed be a problem. I believe some thermocouple and RTD amplifiers have notch filters and or low pass filters built in.