Radio temperature

[VK3KYY]

I'm trying to add a system to monitor the internal temperature of the radio

The CPU has an internal temperature sensor, but its not very accurate. The data sheet quotes accuracy of +/- 4.5 deg, but I think this is worst case.

I've made a special version of the firmware, which shows the raw temperature data value from the CPU on the Battery Screen, and also on the VFO screen during Tx.

The version is here

https://drive.google.com/drive/folders/ ... sp=sharing

I'd be interested in any data points that people can give me. i.e Temperature of the chassis of the radio, and the raw temperature value

Ideally, I'd like more than one data point per radio. e.g. a Data point at 20 deg C or 15 deg C and another at 40 or 50 deg C

These are the data points I have captured from my radio,by putting a thermocouple probe between the battery and the chassis.
The batter is not fully in its normal position, and is held at the bottom by an elastic band.
The radio is also wrapped in a towel, so that the heat is contained inside the radio.

gd77-temperater-plot.png (9.86 KiB) Viewed 3888 times

This data is for a GD77.

I've noticed that the Baofeng DM-1801 values are a bit different, and the Baofeng RD-5R values are different from the DM1801 and from the GD-77


BTW.
Width the raw temperature data, initially the CPU seem to take 15 or perhaps 30 seconds to get to a stable temperature.


Anyway. Any data would be interesting, and useful.


Thanks

Roger

Edit. If anyone wants to read the technical data sheet (Application Note) about this.

See

https://www.nxp.com/docs/en/application ... &pageNum=1

But the values in the Application Note are much lower than values from the CPU in the GD77

[DG3GSP]

Hi Roger,

I did a bit measurements, but only at ambient temperature:

Value - Temperature (deg C)
875 - 22
867 - 25
864 - 26
863 - 27
862 - 28
860 - 29

I use the same method as you - thermocouple probe between the battery and chassis.

Please excuse, but i didn't have time for more measurements at higer or lower temperatures. Maybe at another day..

By the way - I see your post "Profiteering" in announcements: what a shame, no Ham-Spirit from this guy

Best regards

Peter

[VK3KYY]

Thanks Peter,

From my measurements I have established that the raw value changes by 2 for ever 1 deg C

This seems consistent across my radios.

However the absolute offset value for 0 deg C varies by +/- 10 between each radio.

To calculate the 0 offset use the calculation

Offset = RAW_VALUE + (ACTUAL_TEMPERATURE * 2)
So from you values, the 0 deg C offset was



Raw Actual Offset
875 22 919
867 25 917
864 26 916
863 27 917
862 28 918
860 29 918

So the average is 917.5

The calculation for actual temperature in deg C is

T = (Offset - Raw)/2


I will need to add a new setting in the options, so that people can calibrate their own radio temperature readings if they want greater accuracy.
This will set the offset value. But I will need to make the user interface easy, and just have a adjustment of +/- 5 deg in 0.5 deg steps

Thanks

Roger

[VK3KYY]

Can anyone tell me whether in the USA, the temperature is always displayed in deg F, or whether perhaps technical equipment displays the value in deg C ?

[KI5GZK]

I believe Fahrenheit is more common, but often there is an option to choose which to show.

[W0RMT]

Can anyone tell me whether in the USA, the temperature is always displayed in deg F, or whether perhaps technical equipment displays the value in deg C ?

In the USA deg F is ubiquitous in daily use, but scientific and technical equipment at least has the option to display in deg C. Most of us who work in science use deg C in our work, though we might also use deg F in our daily lives. FWIW I prefer deg C.

[VK3KYY]

Can anyone tell me whether in the USA, the temperature is always displayed in deg F, or whether perhaps technical equipment displays the value in deg C ?

In the USA deg F is ubiquitous in daily use, but scientific and technical equipment at least has the option to display in deg C. Most of us who work in science use deg C in our work, though we might also use deg F in our daily lives. FWIW I prefer deg C.

The decision was taken, that as this is a technical device, it was not worthwhile all the additional complication of having to handle 2 different temperature systems, and only Deg C will be supported.

I know in the conversion if easy, but to support both requires an additional display option, plus a load of extra texts, all of which need to potentially be translated, as well as additional voice prompts, and the voice prompt memory area is already 95% full for some verbose languages.
And the remaining program space is also now getting quite limited.

[W0RMT]

Oh and I'll be happy to collect some data if I can find a flat thermocouple. I've got some nice lab grade equipment from work which can collect triggered recordings which would make this easy. If I can just find that probe instead of the larger stainless probes I usually use.

[W0RMT]

I'm more than happy to *not* have measurements in deg F. It is a silly system...

[VK3KYY]

The best way to "calibrate"is to partially remove the battery, and put the probe between the battery and the chassic

Then wrap the radio in a towel, and turn it on, and leave it 5 or 10 mins, so the CPU temperature and chassic tempearature are roughly the same.

Then adjust the calibration setting to get rid of any offset.

There is no need to take multiple readings, just wait for the temperature to be relatively stable (probably at around 25 deg C, or possibly 30 deg C)

The sensor seems to be quite linear. Internally the data is 2 steps per deg, but the sensor is exceptionally noisy, and has to be put though a very large low pass filter (long average of 250 samples), so get rid of the noise.

Also bear in mind, that the sensor is in the CPU not the chassis.