Page 1 of 1

Temperature measuring

Posted: Tue May 24, 2016 9:40 pm
by Lenp
I have a test jig that measures the temperature of 4 individually heated assemblies. When the jig was built I used 4 PID controllers, with PT100 sensors, one for each assembly to be measured.
All was well but lately the system has become erratic. On start up the displays can show as much as a 4 degree C span and all the sensors are in the same ambient environment. A offset can be entered to get them all back in line but eventually the error repeats. Interchanging the sensors was inconclusive. I plan on substituting precision resistors for the sensors to determine if the controllers or sensors are drifting. If it is determind that it is the sensors, what is the opinion as to changing them to type K thermocouples. The controller can accoodate RTD devices or thermocouples, so it would be is a low cost fix.

We have used this same combination of components to replace bi-metal thermostats in several laboratory incubators, and have had great success. So far this problem has been a bit of an enigma!

I am waiting for delivery of a 4 channel data logging thermometer with NIST tracable calibration to help determine the problem, and to use it as a secondary calibration reference. It uses type K termoocouples so I wonder if these are more stable than the PT100's?

If it turns out to be unstable controllers, does anyone have suggestions for a replacement?
They are used for temperature measurement only, so no control function needed. They operate from 120VAC, and fit a 1/32 DIN panel cutout.

The components used are from
Miniature TD sensor Model Pt100MN
PID controller Model SYL-1512A

The floor is open for comments!


Re: Temperature measuring

Posted: Tue May 24, 2016 10:22 pm
by gunter
All 4 temperature probes are reporting a 4 degree shift? you mentioned incubators, is the 4 degree shift coming from whatever is being incubated?

Re: Temperature measuring

Posted: Wed May 25, 2016 5:33 am
by Lenp
No. The incubator was mentioned only because the same controller was used without the errors.
The problem with the jig setup is before any tests are starrted. Sometimes each controller is off and sometimes only one or two are in error.
I just realized that one difference between the incubator and test jig is that the incubator is powered 24-7 while the jig is onl only as needed.
This may point to controller instability so I'll let it run overnight to see if that affects the readings.
It's a bit unusual that digital equipment needs a 'warmup' time!

Re: Temperature measuring

Posted: Mon Jun 06, 2016 4:13 pm
by haklesup
Sounds like a calibration error. I'm not very familiar with the accuracies of these thermocouples. link implies they are the OEM so you could call them for support questions about calibration and accuracy of the specific components you used. ... vices.html

Re: Temperature measuring

Posted: Tue Jun 07, 2016 12:54 pm
by jwax
I'll second the vote for Omega- they are the temperature wizards of the world.

Re: Temperature measuring

Posted: Tue Jun 07, 2016 3:08 pm
by Lenp
I agree about Omega, they are top drawer...if they have what you need at a price you can afford.

I just received a 4 channel type K thermometer with data logging for under $200, complete with a bogus NIST certification.
The instrument works well, each channel is within 0.3C and each channel can be calibrated. It will log a large number of readings in memory, then with a A USB cable and proprietary software it dumps the data to a Excel compatible file, graphs and shows averages for the measuring period.

The NIST certificate is bogus because the cal date was 18 months old! They gave a hundred reasons, like it is good for 1 year from date of receipt....I guess, if it was on their shelf 10 years, it would still b valid! The real reason is that someone swapped a freshly calibrated unit for an out of calibration unit, to freshen their stock, at my expense.
In all, he instrument is still a right choice for me.

If there are any USB hackers out there I would like to monitor the communication between the software and the instrument so that maybe the data could be harvested in real time. Absolutely no protocol or data information is available from the suppliers or producer, so any success would be totally luck propelled!

Re: Temperature measuring

Posted: Tue Jun 07, 2016 4:28 pm
by jwax

Re: Temperature measuring

Posted: Thu Jun 09, 2016 4:07 pm
by haklesup
I should have looked up the PT100 and I would have realized it was an RTD. The resistive thermal device can be offset by series resistance in the interconnecting wires. Since a thermocouple relies on a more reliable voltage drop across a bimetal junction, this is done with low current source so any additional voltage drop across resistive connections is negligible. RTIs also can have larger tolerance in the coefficient than TC types.

A USB protocol emulator (USB Sniffer) you can look at signals with can be expensive, its essentially a serial logic analyzer with high speed. Once that work with USB 3 are very expensive. Check out the Beagle USB line of products bottoming out around $400. A Bluetooth sniffer is much cheaper, I got one free at Design Con as a dev kit sample once. Custom USB drivers in windows can be a pain in the neck too, stick with drivers that come with off the shelf uControllers unless you want to be forced into a 32bit OS where you can load unsigned drivers.

You probably don't need to do that though, Control codes for thermal controllers is standardized. for RS232 its called Modbus, I forget the name for USB but if you look at other PID controllers that have this documented, it may also work for the one you have.

Check this all software tool out too