This article is a review of the EARU Programmable Thermocouple K J PT100 to 4-20mA Converter TC RTD Input 4-20mA Output Head-mounted Temperature Transducer.
So the concept is that one can program the transducer module for a range of input sensors, and lower and upper limits to the conversion range. Another significant benefit is that it does cold junction compensation (and it does not seem possible to disable it if you wanted). The stated cold junction compensation range is narrow at 20-60°, it seems unlikely but bears testing.
Yep, the software (which you can only get after making a purchase… gotta be a bad indicator), is purely Chinese, but this handy dandy graphic goes some way to helping an English reader to use the software. The software is glitchy, so that exacerbates the problem.
Apparently the Chinese seller isn’t too hot on the software, having ordered a pre-configured transducer for a K type thermocouple for 0-100°, they sent a Pt100 RTD for 0-100°. Above, the software indicates 1878° for the transducer with a K type thermocouple at around 20°. Chinese Quality.
It is fortunate that I also purchased the relatively expensive programming cable as I could re-program the thing to my requirement.
So, here are the specs from the Aliexpress listing.
Programmable thermocouple K J S T N R B RTD Cu50 PT100 input 4-20mA output head-mounted temperature transducer
There is a special chip inside our cable. only this cable can work with software. software can not work with other cable.The software and operate manual will be sent to you by email, please leave your email address.PACKAGE: Per order only one transducer, without config. cable.NOTE:
- *This is programmable transducer , user can set the range by config.cable. *The input of transducer is set as user requirement by seller, if you want set the range by yourself, you can place the order of cable.
- * Its input sensor type and temperature range can be changed by our software. The default input options are factory settings for reference. The software supports PC systems such as Win10/8/7/XP.
- *This upgraded version of temperature transducer has better accuracy and experience, and is easier to use.
- Main Features
● Internal embedded cold-junction compensation.- ● Have USB port to do configuration.
- ● Signal input type, temperature range are programmable through software.
- ● Universal RTD TC thermocouple signal input, 2-wire 4-20mA output.
- ● 12-35V wide range power supply.
- ● Support 2-wire, 3-wire, 4-wire RTD PT100/Cu100/CU50 input.
- ● Support all the thermocouple signal input, type K, J, T, S, N, R, etc.
- ● Standard head-mounted package.
- ● Refined frosted external shell, flame-retardant material.
- ● High performance in withstanding shock, vibration, external interference, etc.
- ● In compatible with all the PLC, temperature controller, meter, etc.
- ● Industrial grade operating temperature range: -40 to +85℃
- Technical Parameters
- – Signal input: Thermocouple or 2-wire RTD, 3-wire RTD, 4-wire RTD.
- -Output: 2-wire 4-20mA current.
- -Output current limit: 20.3mA
- -Power supply: 12-40V, 0.2mA excitation current.
- -Measuring Accuracy: 0.5% TC input, 0.2% RTD input.
- -Temperature Drift: 0.02% FS/℃
- -Excitation current: 0.2±0.01mA
- -Load capacity: 250Ωor 500Ω
- -Protection circuit: Reverse polarity connection protection.
- -Software programmable: Yes, have programming usb cable, configuration software (English).
- -Support HART: Not supported
- -Operating temperature: -40-80℃
- -Storage temperature:-40-100℃
- -Material: PA66
- -Installation:M4x2 screw, standard head-mounted.
Above is the DUT. It is not the same colours as the advertised product, there are no brand or model marks. This is not unusual for Chinese sellers and products.
The programming cable does appear to have electronics in it, it instantiates a CP210x USB-serial driver on the port. The transducer does have a mini-USB jack on it, but it is NOT USB, the Chinese often appropriate standard USB connectors for non-USB use.
I could not get the thing to work properly in test mode using the software, so I proceeded to just try the reprogrammed transducer and measure its response. Initially I connected it to an existing project in place of a conventional ‘fixed’ transducer, and it did not work properly.
It seems that when the programming cable is plugged into the transducer connector, transducer current will not go below about 11.5mA, so the test facility works to a limited extent, for currents from 12-20mA.
I interpreted the first graphic above statement of response time of 1s as meaning that it provided valid output within 1s of power up (which is relatively slow compared to ‘fixed’ transducers… but this no doubt has a microcomputer doing smart things.
Above is a scope capture of applied 12V (blue trace) and the voltage across a 100Ω load resistor (red trace).
In fact, the 4-20mA output is updated about every second, but the response time for the cold junction sensor is much longer than 1s.
The red trace is extremely noisy, and the noise seems to depend on the start up process inside the transducer… demonstrating that the noise comes from inside the thing. It may be possible to filter this noise externally.
The other thing that the capture shows is that is takes almost 5s before the output stabilises at the correct value. It is very very slow.
Many of my applications are battery powered, and they wake occasionally, power the 4-20mA loop, make a measurement, shut the loop down, report the measurement and go to sleep. For the other projects with ‘fixed’ transducers, I allow 400ms for the loop current to stabilise, this transducer needs 6s or more, 15 times the power consumption.
So, it is possible to program the sensor type and the upper and lower limits for the 4-20mA output, and these can be read back from the transducer at a later time. It is also possible to program a correction to be added to the measurement, unfortunately this cannot be read back exactly at a later time… though you might get a close estimate by observing a reading, and whilst the temperature is held constant, set the correction to zero and calculate the change in observed temperature reading.
A test without thermocouple
The data logger was run with a short wire link in place of the thermocouple. This effectively then is measuring and logging the cold junction sensor.
The transducer was heated with a hot air blower and allowed to cool, giving an idea of the thermal time constant of the cold junction (assuming the sensor is thermally bonded to the cold junction terminals, as it should be). The thermal time constant is about 900s.
Above, a longer test run of the air temperature inside the house. It has a bit more jitter than I expected, but that may be valid.
A test with a Pt100 RTD
Above, a PID temperature controlled stirred beaker of water to provide a stable measurement target to explore jitter on a Pt100 probe.
Above is a plot over a few hours of the temperature controlled water bath. Note that the jitter is due to both the temperature regulation system and the measurement system. For this dataset, mean is 39.86° and standard deviation is 0.162°, that sort of total jitter would be quite acceptable for many purposes.
Conclusions
The following conclusions are made:
- the configuration / diagnostic software client is Chinese language, with more emphasis on visual effects than function and reliability;
- the programming cable is proprietary, and relatively expensive for what it is;
- the transducer has a minimum current of just under 12mA if the programming cable is connected, making the basic diagnostic facility of the software useless;
- the cold junction compensation in a basic test of a type K thermocouple seemed to work, it is slow, but it is more important that it tracks the cold junction accurately than that the system responds to ambient changes quickly;
- in concept, a transducer that is configurable for a wide range of sensors / inputs, and programmable scaling to 4-20mA which seems to work for the type K thermocouple case tested;
- the transducer is relatively inexpensive, and the flexibility is a plus… if it were not for the half baked programming / diagnostic software. and the high minimum current when connected;
- the stated cold junction compensation range is narrow at 20-60°, it seems unlikely and would be very limiting, but bears further testing.
- the accuracy of conversion from any of the input sensors to 4-20mA was not tested.
I referred the clearest of the problems, the failure on test currents below 12mA, to EARU, but they could not understand.