Is a PLC the right way to go in this case?

Fasthotrod

Member
Join Date
Oct 2007
Location
Oklahoma
Posts
26
Good afternoon everybody!

I was recently tasked with a project and I think that my solution will be to integrate a PLC into it, but I'm not very familiar with PLC programming so I don't know if it's possible to do it easily or if I'll be getting in over my head.

Here is the situation: We are replacing a standby power system at a facility. The facility has a "legacy" computer system (legacy meaning: old) that monitors the entire facility, including the power system. The legacy system provides a readout on both a local and a remote computer terminal.

The new power system that is being installed is not compatible with the legacy system, and the legacy system cannot be modified.

Therefore, I need to come up with a solution that can get the information from the new power system and sends the proper signals to the legacy system.

I am looking at the following parameters for the power system:

AC Voltage: 0-150 VAC (x3 phases)
AC Current: 0-150 Amperes (x3 phases)
AC Frequency: 60 HZ
Transfer Switch Position: Commercial or E/G
Generator Battery Charger Current: 0-10 Amperes DC
Generator Oil Pressure: 0-75 PSI
Generator Water Temperature: 0-220 Degrees C
Generator Oil Pressure Alarm
Generator Temperature Alarm
Generator Over-Speed Alarm
Generator Over-Crank Alarm (Starter Time Out)

The Legacy system expects to see the following input:

AC Voltage: 0-5 VDC (x3 phases)
AC Current: 0-5 VDC (x3 phases)
AC Frequency: 120 VAC @ 60 HZ
Transfer Switch Position: NC Contacts
Generator Battery Charger Current: 0-5 VDC
Generator Oil Pressure: 0-100mV DC
Generator Water Temperature: Unknown at this time...
Generator Oil Pressure Alarm: NC Contacts
Generator Temperature Alarm: NC Contacts
Generator Over-Speed Alarm: NC Contacts
Generator Over-Crank Alarm (Starter Time Out): NC Contacts

I realize that I can just install voltage, current, temperature and pressure transducers with a 0-5 VDC output to feed the legacy system, but there are a couple of inputs that are a bit more complicated. (One input is 0-100 mV input for example.)

So it got me thinking... what if I were to use a PLC that acts as an "interrupter?" I could take the various inputs from the power system and use the PLC to scale them to whatever output I need.

For example, the legacy system expects to see a 0-5 VDC signal for an input voltage of 0-150 VAC. I could use a voltage transducer with a 0-5 VDC output, or I could use a transducer with a 4-20mA output, and use the PLC to give me an output of 0-5 VDC. I would think that there is a simple way to do this, correct?

The same holds true for the oil pressure reading. I could use an oil pressure transducer with a 4-20 mA output and use the PLC to convert it to 0-100 mV output. (The legacy system has a high impedance input, so I'm thinking I could use a 0-20 mA output from the PLC and use a simple resistor network to get the voltage I am looking for.)

The coolant temperature reading is a bit more complicated. The legacy system supplies -15 and +15 VDC to a solid state temperature sensor which provides a linear output based on the temperature. The sensor reads 1000 ohms at 25 C. The temperature sensor provides a voltage/current input to a variable gain amplifier, which then converts the signal to 0-5 VDC from 0-220 C. I'm currently working on the existing circuits to see if I can come up with something that will work with the PLC output... I'll add more to this part as I learn more about the circuit.

One more thing: The replacement power system has the ability to communicate via RS-232 and RS-434 using MODBUS. I believe that I can use the PLC to get the information I need from the power system, then have the PLC give me the respective output.

So what do you all think... am I off my rocker, or is this something that can be done (relatively) easily?

Thanks in advance. (y)

Mark
 
What you are suggesting would work.

Using a PLC to gather IO and then presenting that to a second system is quite common and sometimes called a 'data concentrator'.

A PLC can accept a large variety of physical inputs, scale them and store them in contiguous blocks to be polled by a second system via its serial port.

With detailed information on the 'replacement system' folks on here would probably be able to help you more.
 
The temperature sensor sounds like a 1000 ohm RTD. Look further into it. You can buy custom controllers (transducers). There's one company that has a large 'A' in their logo. Advanced or something. They make all kinds of controllers.

With RS232 and/or MODBUS as options, I think PLC is the best choice.
You'll be happy when you get down to the wierd parts, because the PLC will give you so much flexibility.
 
Thanks for the quick replies, guys! (y)

SLC_Integrator: The replacement power system is made by Kohler. They are using a Kohler DEC-340 or DEC-550 controller on the Engine Generator package, and a M-340 controller on the Automatic Transfer Switch. (ATS) I have the various kits that Kohler uses to communicate with their controllers via laptop or modem. I believe an associate of mine has used a PLC from AD and has pulled the data from the controller... but I haven't had the time to check with him to see exactly how he did it.

As for the remaining details regarding the power system, it's really up to me to come up with the associated measurement devices, so I am not "bound" to anything necessarily. If the system is equipped with 200:5 CT's then I would simply tap into that CT loop with a transducer with the same rating and either a 4-20 mA or 0-5 VDC output. NK Technologies has a nice little clip on device that is both the CT and x-ducer that works for 0-100, 0-150, and 0-200 amperes if that CT loop cannot be penetrated for some reason.

If you have some other ideas, I'm certainly glad to listen.

Keithkyll: The temperature sensor is a silicon PTC thermistor made by U.S. Sensor Corporation. (USP2700)

The value of the output based on temperature can be calculated: Rt = 1000 + 7(Tc - 25)

The device is becoming difficult, if not impossible, to find.

The circuit that the thermistor connects to is a variable-gain instrumentation amplifier, powered by -15 VDC on one rail, and +15 VDC on the other rail.

-15 VDC ----[47k ohm]--O--[Thermistor]--O--[47k ohm]---- +15 VDC

This allows for a small voltage to develop across the thermistor, which is the input to the amplifier. (Imagine that the amp is in parallel with the thermistor, attached to the terminal points --O-- )

For example: With +/- 15 VDC applied (30 VDC delta) and a temperature of 25 degrees C, the thermistor is at 1000 ohms. That means that the circuit has a total of 95k ohms between the +/- 15 VDC rails. Current though the thermistor is approximately 0.31579mA, and voltage drop across the thermistor is 0.3157 VDC. (+0.157 and -0.157 VDC into the circuit.) This voltage is 'read' by the variable gain instrumentation amplifier and converted to an output.

At 220 degrees C, the thermistor is 2365 ohms, so the voltage across the thermister (to the amplifier) is 0.736 VDC. (-0.368 and +0.368 VDC) This input provides an output of 5 VDC from the amplifier.

To make it simple, I am trying to see if I can pull that v-gain amp out of the legacy system and just provide a 0-5 VDC signal to their gear. That way, I could use either a temperature sensor/transducer package as an input to the PLC, or a thermocouple with a PLC thermocouple input card and send a 0-5 VDC signal to the gear.

What do you guys think? o_O

Mark
 
I wouldnt be surprised if the engine controller supports Modbus RTU. All ones I have done do just that.

We pull data from Murphy Millenium engine controllers, concentrate it in a PLC, bring in any extra or incidental IO and display it on an HMI and/or a SCADA system.
This sounds pretty similar to what you are trying to do.

The only thing is to make sure whatever PLC you choose is capable of being a Modbus Master.
AB products are not capable of Modbus Master unless you buy a Prosoft card which isnt cheap and they are a bit of a PITA to set up, especially if you havnt done one before.
 
Fasthotrod said:
...To make it simple, I am trying to see if I can pull that v-gain amp out of the legacy system and just provide a 0-5 VDC signal to their gear. That way, I could use either a temperature sensor/transducer package as an input to the PLC, or a thermocouple with a PLC thermocouple input card and send a 0-5 VDC signal to the gear...
A thermistor or RTD is typically chosen for fast response and/or accuracy. A thermocouple is chosen for ruggedness. Make sure your system can handle the error of a T/C.
You've given quite a bit of detail to how the thermistor is read. That much detail isn't needed if you are selecting a new amplifier. Just the Rt = 1000 + 7(Tc - 25) or 1000 - 2365 ohm spec (and related temps) is enough. The new controller doesn't necessarily need to use +/-15 as a drive voltage.
If you need accuracy, go with an RTD. These should remain easy to find in the future.
Retransmit of 0-5 analog output is standard stuff for any brand PLC.
 
It sounds exactly like what you are talking about, SLC_Integrator!

I went and grabbed some documentation regarding the Kohler equipment. It looks like each Kohler device has a RS-232/RS-485 port (not RS-434 as I said in my original post... sorry!) and has a provision for a MODBUS/Ethernet Converter. I looked in the box and bingo: I have the converter.

Everything I am reading in this manual says "MODBUS/KBUS" but there is a Remote Network (Ethernet) Connection diagram that shows various devices connected to an Ethernet Network.

[Network]--MODBUS TCP-IP--[Converter: MODBUS/Ethernet]--MODBUS RTU RS-485--[Device]

From what I can tell looking at the diagram, I can connect a device to a RS-232 port (one port per device) or I can connect them all to a single RS-485 connector, or use the MODBUS/Ethernet converter for each device and setup a hub or a switch.

Now all I have to figure out is how to do that... o_O

If I can poll the data from the E/G and ATS, I can skip the voltage, current, temperature and pressure transducers all together!

Is it weird that I am getting excited about this project? :oops:

Mark
 
keithkyll said:
A thermistor or RTD is typically chosen for fast response and/or accuracy. A thermocouple is chosen for ruggedness. Make sure your system can handle the error of a T/C.
If you need accuracy, go with an RTD. These should remain easy to find in the future.

I'm not sure I understand these statements.

Response of a temperature element is a function of how thick the thermowell is and its material of construction, not the type of element inside it.
Also I'm not sure what 'error' you are talking about with a thermocouple. A thermocouple is incapable of being 'out of calibration' in that it has a very simple temperature/mv correlation.
 
The temperature device is installed on an engine, so it needs to be able to handle the vibrations... but so far the thermisters have worked without any issues. I am ignorant to the differences in response times... can you give me a 'ballpark' figure for what the difference is? 1 second? 5 seconds? 30 seconds?

What really hurts this project is that thermister isn't being made any more. If they were, I could just buy a bunch of them and install them with the new system and keep the existing amplifier in place.

The +/- 15 VDC rails that power the thermister are part of the legacy system... changing that might cause a few headaches.

If I can change the amplifier, then all I need to do is pull the circuit card and put a 'dummy' card in place that allows me to connect to the legacy input points, and use the 0-5 VDC output from the PLC. I have recommended this, but I am getting some push back...

Unless I can come up with something that works, we may not have an option to pulling that circuit card... so I want to try and figure out if I can make it work before throwing in the towel.

I wonder if a motorized potentiometer would work? o_O

Mark


keithkyll said:
A thermistor or RTD is typically chosen for fast response and/or accuracy. A thermocouple is chosen for ruggedness. Make sure your system can handle the error of a T/C.
You've given quite a bit of detail to how the thermistor is read. That much detail isn't needed if you are selecting a new amplifier. Just the Rt = 1000 + 7(Tc - 25) or 1000 - 2365 ohm spec (and related temps) is enough. The new controller doesn't necessarily need to use +/-15 as a drive voltage.
If you need accuracy, go with an RTD. These should remain easy to find in the future.
Retransmit of 0-5 analog output is standard stuff for any brand PLC.
 
Thanks, rsdoran!

Truth be known, I have downloaded the DL-06 manual, printed it out, and been looking at it as the PLC for this project. (y)

I have very, very limited exposure to PLC's so I'm sure that I'll be asking all kinds to questions here.

My only real experience was playing with an AB SLC-500 that we used for monitoring contacts and turning on indicator lamps, and totalizing some mA inputs from some transducers and sending the signal to a display panel. I saw the ladder logic, played with the software a little, but that's it. I have also played with a Woodward SMART 3000, but not a whole lot. Every time it was more around 'does this thing work' vs. 'make this thing work.'

I downloaded the free program from AD but I haven't had a chance to play with it too much yet. Making a light turn on looks easy... taking a 4-10 mA input and creating a 0-5 VDC output, not so much. Grabbing data from MODBUS and creating a 0-5 VDC output? :unsure:

Thanks again, guys! Please keep the ideas and suggestions coming!

Mark
 
Last edited:
Ahem... Electro/Mechanical, thankyouverymuch! :p :D

I am an electronics technician and I have been in the electrical power industry for almost 18 years.

I got started in the USAF and I moved on into private industry and elsewhere. I won't bore you with the details, but I know my way around the electrical side... but I want/need to learn more. :oops:

That's why I am currently working on my BSEE here in Oklahoma... while working full time. :sick:

I am completely ignorant to PLC programming, but I'm sure I'll be able to pick it up over time. This web site looks like it's the place to get the help I need, and get pointed in the right direction.

Thanks for the links, Sensei. :geek:

Mark

rsdoran said:
Ahh grasshopper, you would not be the first mechanical person to have done this, take a look at these, it may explain many things before you even get started

http://www.plctalk.net/qanda/showthread.php?t=11015

http://www.plctalk.net/qanda/showthread.php?t=22335
 
SLC_Integrator said:
...Response of a temperature element is a function of how thick the thermowell is and its material of construction, not the type of element inside it.
Agreed. Thermocouples can be isolated, grounded, or exposed. Exposed has the fastest response. Grounded (bonded to the sheath) is next. The problem with exposed is obvious. Grounded can give you grief with noise. Isolated offers the best noise immunity, but has a slower response.
SLC_Integrator said:
Also I'm not sure what 'error' you are talking about with a thermocouple. A thermocouple is incapable of being 'out of calibration' in that it has a very simple temperature/mv correlation.
The charts are based on perfect purity alloys, etc. In actual practice, there is significant error. Add to that Cold Junction Compensation and lookup tables.

Quote from an FAQ at Omega Engineering:
[font=Verdana, Arial, Helvetica, sans-serif]THERMOCOUPLES generally can measure temperatures over wide temperature ranges, inexpensively, and are very rugged, but they are not as accurate or stable as RTD’s
and thermistors. RTD’s are stable and have a fairly wide temperature range, but are not as rugged and inexpensive as thermocouples.

Omega's website is a great resource for info regarding temperature.

Fasthotrod: T/C electronics will cost about $75.00 per point with AD - a good price. If you can't get Omega, Gordon or others to make you custom thermistors, then thermocouple is probably the best way to go. The error on T/C's in my experience has been up to 1 degree or so. It's always an offset, never linear. After everything's installed and at operating temp, take a reading and calculate the offset correction. I think there's a place to put this in the PLC. After that, you can expect .1C accuracy in your temp range.
If you need better than that, then stick with the thermistor.
I don't like RTD's in high vibration environments.




[/font]
 
keithkyll said:
[font=Verdana, Arial, Helvetica, sans-serif]

Fasthotrod: T/C electronics will cost about $75.00 per point with AD - a good price.


I'm sorry... I don't follow you. :oops:

If you can't get Omega, Gordon or others to make you custom thermistors, then thermocouple is probably the best way to go. The error on T/C's in my experience has been up to 1 degree or so. It's always an offset, never linear. After everything's installed and at operating temp, take a reading and calculate the offset correction. I think there's a place to put this in the PLC. After that, you can expect .1C accuracy in your temp range.
If you need better than that, then stick with the thermistor.
I don't like RTD's in high vibration environments.

I see no reason for it to be more accurate than that... 1 degree would be fine by me. 0.1 would be better obviously, but I don't know if the legacy system have that kind of resolution or not. It's simply for remote monitoring, so 1 degree (or less) is acceptable.

In that case, I could do one of a couple of things:

Buy the raw thermocouples and add a thermocouple expansion card to the PLC for the input, and scale the output from the PLC to the legacy unit.

OR

Buy a 'pre-manufactured' device that I can just attach to the engine. (I have used units from Wilkerson Instrument Company with JMS thermocouples in the past... TW304's with 1J1BKU's if I remember correctly.)

If I grab the temperature data from the E/G via MODBUS, I can eliminate the need for the sensing units all together. All I will need to do is grab the data via MODBUS and then output that data from the PLC (0-5 VDC) to the legacy unit. (Unless I get stuck making that existing amplifier card work.)

So do you guys have a lot of experience using the AD DL06? Is it a good unit?

Thanks guys! I really appreciate your help!

Mark
 

Similar Topics

Hi All, I have a programming background but have never used a PLC. I have what I think is a simple automation project however Im not sure of the...
Replies
19
Views
1,778
Good Day everyone, i am currently wiring up a panel and i am not sure if a L36ERMS compact guardlogix controller is just used for safety or you...
Replies
4
Views
554
Good day all. I work for a small company that would like to start automating the production of our goods. I have been tasked with doing some...
Replies
49
Views
13,481
Could someone point me to a DB9F right angle cable that would allow me to close the door on a L61 PLC? Certainly one exists but all I can locate...
Replies
4
Views
1,776
With Windows 7 End of Extended Support being just under 3 years away (Jan 2020), is the time right for someone like Inductive...
Replies
45
Views
13,062
Back
Top Bottom