Digital Thermocouple Input for Micrologix 1000?

Disclaimer: I am more of a layman and not an automation expert. The business I own has some machinery controlled by a PLC and we're having a problem so I wanted to seek some advice here for direction.

We have a problem interfacing thermcouples to the digital inputs of a PLC. The machines we are running are controlled by an Allen-Bradley (Rockwell) Micrologix 1000 1761-L16-BWA. This PLC does not have analog input capability, only digital. We have two J-Type thermocouple signals we need to feed into the PLC. This machine had been accomplishing this by way of Sensorpulse MSP-TC-IN modules (photo). These modules handle signal conditioning and A/D conversion all in one compact module. Over a single-wire to the PLC, they provide a 16-bit input, the interpretation of which is enabled by customized logic code from Sensorpulse coded into the PLC program. This appears to have been an elegant and relatively inexpensive solution. However, these modules are long obsolete and the ones we have are beginning to fail. Sensorpulse also made a multichannel card which we have on another machine and it working perfectly (until it dies...).

So the question is this:

What is the most efficient and/or economical way to get two thermocouple inputs isolated, linearized, converted to digital and intelligbly inputted into this PLC? I assume that some revision of the logic program would be required as well and would this code come off the shelf with the module? In this case, the PLC and I/O cards are housed in a small explosion proof enclosure on board the machine, so form factor is an issue as well. It may come to getting a local automation tech to come in and customize a new solution, but I'm not sure how many arms and legs that might cost us. I at least want to be aware of the options out there.

Thanks!

replacing the micrologix1000 unit with a Micrologix1100 and a thermocouple card would probably be the easiest way. Also, a micrologix1100 can be programmed using the free version of RSLogix500, so you wouldn't have to buy the software.

A functionally similar device to SensorPulse is the Calex 8500 and 8600 series.

http://www.calex.com/SignalCond.html

Calex, and that product line, seem to be healthier and more modern than the SensorPulse product.

Omega Engineering also has a unit similar to the Calex. Both the Calex and the Omega output is frequency proportional to temperature - I'm not sure about the SensorPulse.

I know nothing at all about this device but for those interested here is what a quick Google search yielded.
links:
http://www.sensorpulse.com/Downloads/ProductSheets/msptcin_bull_rev2003.pdf
http://www.sensorpulse.com/products/Analog_IO/msp_spec.htm
http://www.sensorpulse.com/Support/Drivers/PLC%20Driver%20Menu.htm

Interesting device
Do you think the units circled below are in milliseconds?

I was thinking that if the data rate was timed around a 10ms bit width, you could decode this with a STI routine.

I had a customer with the same issue earlier this year that was using them to code a load cell signal. Only in their case they were using a PLC Direct plc and since it was originaly developed a mV input module had been developed. They also had an open slot so the conversion was relatively painless.

These work by encoding the signal almost like a slow form of serial communication. It normally only updates a portion of the decoded signal, and thus if you have a rapid moving value it can give erroneous results. It only periodically refreshes the entire value. They are not like a V/F or I/F converter but instead is literally sending a data encoded number.

Unless the precision is required Kens and Tom's suggestions are probably the least expensive solutions. Otherwise going to the 1100 is probably the next best bet.

This one makes me a little sad

In 1992 a young enginer and I were riding the bus back from Automation World in Chicago and sketched on the back of a napkin (really!) a device to get analog signals into a micro-PLC. At that time none of the little micro-PLCs had analog inputs, and big PLCs with analog inputs were very expensive - we only needed one per PLC anyway for the machine control system we were building. We built it ourselves - got a circuit board making kit from Newark, printed the circuit board mask on overhead transparency paper, etched the board by holding it up to the window on a sunny day ......

The thing worked great, we started selling a few despite lack of marketing and distribution, expanded the line, and I was feeling pretty good. Then Omega and Calex came out with essentially identical units (a case of parrallel invention I'm sure) and I knew I couldn't compete with that kind of distribution. We sold out our stock and moved on ....

At any rate, cleansc should be aware that he will have to make program changes to his PLC. The SensorPulse unit sends out essentially a binary number, with 16 bits and a pulse pattern to identify the beginning and end of the data. GE used to sell the same thing. The Omega, Calex and my old unit send out a square wave for a high speed counter input, with the frequency proportional to the signal. For a 4-20 mA unit at 4 mA the pulse frequency would be 200 Hz, at 12 mA 1200 Hz, at 20 mA 2000 Hz for example.

Thanks for all of your feedback.

Tom, it looks like you designed exactly what we need. Too bad you scrapped it!

I'd rather avoid the expense of having to buy a new PLC (1100) unless that's the only route. Are you all saying it would be possible to use the Calex and/or Omega units with only a change to the logic code? In the case of the Calex and/or Omega units, if they are generating a frequency output, I would have assumed this would require an analog input on the PLC to accept that signal (but I'm not the expert!). So, would the Micrologix 1000 have the ability to accept a frequency input as a digital input? You mentioned the high-speed counter function....does this have the ability to count square waves in discrete time intervals and translate those back into the process value?

I understand this would require some changes to the logic code, which I'll probably have to get an expert to do. Although I'm a chemical engineer by training, I successfully reprogrammed a PLC in a chemical plant years ago, but have forgotten more than I remember about it by now!

I know that the idea of replacing the PLC seems formidable at first but the Micrologix 1100 is not very expensive - and you can probably ask your AB dealer for a discount, esp if you are going to be buying a few. Moving away from the ML1000 to the ML1100 will future proof your system for at least a couple of decades and it will save you future programming costs.

cleansc said:

... Are you all saying it would be possible to use the Calex and/or Omega units with only a change to the logic code? In the case of the Calex and/or Omega units, if they are generating a frequency output, I would have assumed this would require an analog input on the PLC to accept that signal (but I'm not the expert!). So, would the Micrologix 1000 have the ability to accept a frequency input as a digital input? ....

Click to expand...

As is often the case, the answer is "It depends".

You can take the output of the Omega and Calex units directly into a digital input, but it will probably have to be a high speed counter input. The ML-1000 with 24 VDC I/O has, I believe, high speed counter inputs built in. If that is what you have then you may have to wire the Calex to it (although the SensorPulse was probably on this kind of input anyway if you have it). You will then have to have the program modified to eliminate the SensorPulse code and convert the Omega converter frequency to equivalent data. You don't need an analog input for this option.

If you don't have a high speed counter input - for example if your I/O is all 120 VAC - then you will have to replace your PLC. At that point you may as well go with a MicroLogix with analog input. This can be an ML-1100 as recommended above, but there are also ML-1000 units with analog inputs.
http://ab.rockwellautomation.com/Programmable-Controllers/MicroLogix-1000

I really appreciate the help.

OK, it seems that either of these options might be one way of accomplishing what we're after:
http://www.calex.com/pdf/8511_12_22_23.pdf
http://www.calex.com/pdf/8611_12.pdf
http://www.omega.com/manuals/manualpdf/M3209.pdf

According to the specs, it will accept input directly from the thermocouple, compensate/linearize, and provide an 24/7mA/50ms output suitable for input to PLC high speed counter. According to the Micrologix manual, it accomodates HSC and I believe the Sensorpulse may have used mode.

If I were naive enough to think I could try to modify the program myself, what software would be required to do this on the Micrologix 1000? I downloaded a RsLinx Classic Lite software from Rockwell, because I wanted to peek into the program. Would this software allow changes via a USB-Serial-8 Pin DIN DF1 connection from a laptop running that program?

Thanks

My only other concern is that the Micrologix 1000 1761 specs indicate that it has 1 (one) high-speed-counter at 6.6kHz. Should I interpret that to mean that it can only work with one input for the HSC?

I have two such inputs I need to read.

cleansc said:

My only other concern is that the Micrologix 1000 1761 specs indicate that it has 1 (one) high-speed-counter at 6.6kHz. Should I interpret that to mean that it can only work with one input for the HSC?

I have two such inputs I need to read.

Click to expand...

On input per channel is normally the way to go. You can play games to get around this, but generally it isn't worthwhile. It's starting to look like replacing the PLC is the way to go. Your programmer should be able to use most of your existing program, and the hardware cost for hte PLC is usually less than the other cost components (re-wiring, programming, etc).

Should I interpret that to mean that it can only work with one input for the HSC? I have two such inputs I need to read.

Click to expand...

Only one HSC instruction is allowed in each MicroLogix 1000 program. However, one HSC can use up to FOUR inputs, depending on how the ONE HSC is configured. Three of the 8 possible types of HSC use all four of the possible inputs I:0.0/0 to I:0.0/3.

EDIT: I finally RTFM and it says that the SensorPulse uses a single regular old digital input to read one thermocouple. So you can use as many as you have MicroLogix 1000 inputs (20 I think you said). As Tom said, the SensorPulse "driver" program that you load into the PLC probably counts input pulses over some short time period and converts that count into a 16-bit number that is stored in PLC memory as a temperature.

The SensorPuls MSP family of single channel analog I/O modules permits most any analog sensor or actuator (depending upon MSP model type) to be used with any model of PLC (even a PLC without analog capability).

REQUIREMENTS Mandatory:
• PLC with one (1) +24VDC discrete input.
• THHN, MTW or other stranded interface wire to interface MSP unit with PLC I/O.
• 15 - 32VDC, 50mA external supply voltage.
• PLC Driver Software (supplied with unit).
• Personal or Handheld computer with Windows/PC or Windows®/CE - to input PLC
driver to PLC.
• PLC to Personal/Handheld computer communication interface cable

Select appropriate PLC driver from within 'sp' folder on PC/HPC and copy it to your PLC's program. Then within the PLCs program, change the newly installed driver's default addresses to desired real-time addresses.

Click to expand...