Sign in to follow this  
Followers 0
Jim Blackwood

PLC suggestions?

11 posts in this topic

Hi guys, I wonder if I could ask for some advice? I need a transmission controller and think something along the line of a micro/nano PLC might do the job but figured it'd be best to ask the people who would know. This is a sports car application and the gearbox in question is an 8 speed automatic from a Lexus IS-F. It contains 9 solenoids, better than half of which use PWM signals for shift firmness control and intermediate lock-up of the torque converter. Both analog and digital inputs are needed, probably at least 4 of each, maybe more of the digital. We can go very basic with it, full manual control and manual switching of the converter clutch and probably get by with 8 outputs and 8 inputs, or provide automatic functions with 9 or 10 outputs and 10 or 12 inputs. I think these small PLC's can probably handle the job as they seem to be fast enough, but I could really use some help in selecting the best candidates for the job. Obviously small size, low cost and sealed enclosures would be benefits and I've seen some that look pretty close such as the Divelbiss HEC-4000-E-R but not enough input/output. Could you fellas make any suggestions? Thanks, Jim

Share this post


Link to post
Share on other sites
This is really the niche that Omron does best in as far as a major name brand PLC manufacturer goes. "nano" PLC's usually means smart relays. What you are talking about is more of a shoebox type device. The "nano" stuff really isn't cut out for that.

Share this post


Link to post
Share on other sites
Thank you, I hadn't heard the term "shoebox" so I'm guessing that's more of a size reference? I'll look into Omron, I may have a catalog. I also think I'll fire off an email to Galil, I'm using one of their controllers on my mill conversion and they could have something. Jim

Share this post


Link to post
Share on other sites
Galil's niche is motion control. PLC vendors for years have sold PLC's in the "shoe box" or "brick" format. Basically it is a solid box between about 4 and 12 inches long by a few inches wide. There is a terminal strip along the bottom, and sometimes one across the top, too. They typically have about 5-20 digital inputs and outputs directly on board the PLC. There may be some analog (not as common), and there is often an expansion port on the left or right end. They are designed for markets such as small pump controllers such as for a relatively fancy pool controller or other relatively "simple" machine automation. I've seen them show up in diverse areas and often you can mistake it for something else entirely. They can be quite sophisticated. For instance, Allen Bradley's original SLC and many of the Micrologix PLC series are "brick/shoe box" form factors. They are VERY powerful. They lack the capabilities to run a large system with lots of distributed IO and hundreds of IO's, but they can do virtually anything less than that. I've used a single unit to run anything from indicator lights to an entire process step on a production line. Anything that doesn't begin to look like high performance or "distributed" is a candidate. Prices range from a few hundred dollars up to a couple thousand depending on what capabilities you are using. For example a mid range Allen Bradley "brick" is the Micrologix 1400. It has a small screen with some programmability, 16 digital inputs, 16 digital outputs, a serial port, an Ethernet port, 2 analog inputs, capability to expand up to 7 additional IO cards, and limited distributed IO (using Modbus). It has relatively speaking plenty of memory and can accomodate a memory card for storing the program or logging data. Price is around $800-$1000. The smaller Micrologix 1100 version (cut everything I just said in half) is about $500. These are "Cadillac" models and have a lot more capability for a "brick" than average, but it's not too far off from the maximum capability that you can expect to see. "Nano" PLC's tend to have less than 8 inputs and less than 4 or 5 outputs. They almost never have any expansion whatsoever. They are intended as slightly more fancy replacements for discrete relays. The target market/price is usually to make them cheap enough to replace a couple relays in a cabinet somewhere. They simply don't have the IO counts and/or horsepower to grow much beyond this. Prices do not exceed a few hundred dollars.

Share this post


Link to post
Share on other sites
That helps a lot. Gets me in the ball park at least. May be no home run in the cards but I'll gladly take a walk. Here's exactly what I'm facing: Inputs: 1) Throttle position sensor (analog) 2) Manifold absolute pressure sensor (analog) 3) line pressure (analog) 4) fluid temperature (analog) 5) engine speed (a Hall effect sensor picking up teeth on the input shaft) 6) driveshaft speed (same) 7) intermediats shaft speed (same, optional but available and installed) switch inputs 8) park 9) reverse 10) neutral 11) drive 12) manual mode 13) shift up 14) shift down optional but desired (switch inputs) 15) lock up torque converter clutch (TCC) 16) clutch function Outputs: pulse width modulated 1) clutch #1 2) clutch #2 3) clutch #3 4) brake #1 5) brake #2 6) TCC lock up 7) Line pressure reduce switching solenoids (3 way) 8) solenoid SL 9) solenoid SR As you can see I need a PLC with 4 analog inputs and 7 PWM outputs. Plus additional standard I/O. Further expansion is not even a remote possibility. Given the choice, a sealed enclosure is preferred. Current requirements on the outputs should not be excessive, I can run some tests to determine actual current draw but I don't expect problems there. Speed requirements: The OEM controls allow shift times of 100ms (actual) with 200ms lag time from input to shift command. The more of the lag time we can eliminate the better. I see no reason why the PLC shouldn't be able to beat that time by an order of magnitude on a manual shift, provided it isn't all eaten up by processor time. Jim

Share this post


Link to post
Share on other sites
Jim are you wanting to replace jsut the existing controller with a PLC or is a change of sensors a possibility. I mention this because you can get some nice IP67 sensors which do DeviceNet Protocol and then use a SHoebox PLC with DeviceNet scanner to communicate with them. DeviceNet after all is based on the CAN protocol which was an automotive network for sensor first and foremost.

Share this post


Link to post
Share on other sites
Interesting question. The hall effect sensors are fixed, as are the temp and pressure, all internal to the transmission and using the existing internal harness and bulkhead connector. So I don't want to change those. That really only leaves the TPS and MAP which are input to the Megasquirt (MS) engine controller. I could probably use CAN bus out of the MS for those inputs and also engine speed but that only eliminates 2 or 3 inputs. Still, that's half of the analog inputs so it certainly could matter. Jim (Edit): The application is a custom build. The donor is a Lexus IS-F. Use of the OEM ECM and TCM might be possible but would demand a large number of "phantom" I/O to run as the original engine is not part of the build. JB Edited by Jim Blackwood

Share this post


Link to post
Share on other sites
Jim you may have thought of this but I did not see it in your detail. You say you are using a hall effect sensor for engine speed. THat is going to be a fast switch and is going to require a high speed input. That said I think AB and Siemens should have soemthing that will get the job done. Take a look at the Micorlogix 1400 from AB and the S7-200 from Siemens. Both will require some expansion cards. The only other one I can think of that could do something like this is the Unitronics.

Share this post


Link to post
Share on other sites
This is not a PLC but check out the boards on www.diy-efi.org. Those include firing which is where you can really improve things. When you attempt to control the engine itself, speed becomes a big problem (microsecond timing requirements).

Share this post


Link to post
Share on other sites
7 PWM outputs is unusual for a small PLC but this one has that many: http://www.unitronics.com/Series.aspx?page=Vision130 I had one older Unitronics PLC that was being used as an HMI with a Micrologix1500. It is an extremely poweful system with both HMI and PLC functionality included. I was a tad bit overwhelmed by the software at first, just trying to decipher someone else's programming but once I learned the layout of the tools it was really quite nice. Can you tell us the type of analog inputs you need? Some vehicles I have worked on use 1-5vdc, but you'll need to know whether you are monitoring voltage or current in order to choose the right controller in some cases. Anyway, I think you can get down to the exact models that would work for you if you add those details to your I/O list. Paul

Share this post


Link to post
Share on other sites
Thanks, those suggestions really help. I'd be just as happy using a laptop for the HMI but obviously there are advantages to doing it either way. There isn't much space in the car to locate the HMI but where there's a will there's a way. My analog inputs that wouldn't be available through the CANbus from the Megasquirt are the temperature and pressure signals from the transmission. Those are almost certainly resistance elements which could be configured as either voltage or current devices. I'm still waiting on a wiring harness to connect to the transmission bullkhead connector so I can't test them yet, but I did have the pan off and judging from their small size and simple wiring I'd be very surprised to find anything more complicated than that. I could possibly pull the pan again and do some resistance checks. Typically on these type systems I've been seeing 5 vdc used for sensor power. Probably 12 vdc could be used but the 5v is typically more stable. I'll also try to run some resistance checks on the solenoids, that should give me at least a ballpark idea of current draw there. I wouldn't be surprised to find that Toyota made use of the "peak-and-hold" method for extra fast shifts under hard acceleration. I don't know that they did, but it's certainly something I'd have put on the table at design time. Regardless, if the controller is built to handle it that would be something I'd consider. I've also looked into using the stock TCM/ECM, and using the Megasquirt(MS) and GPIO board (general purpose I/O). The stock system will require a large number of phantom I/O to work and prevent "limp home" mode operation, and I expect the MS will be more difficult in terms of programming than the PLC (C++, etc) although some code will be needed anyway to get it to communicate with the PLC. In either case I'll have to learn it, so simpler is better. I've had some exposure to programming but it's not my strong suit and doesn't come easy. For that reason I'm allowing myself the winter months to sort that part out. Obviously the advantage over the stock system is the ability to modify the settings and program. I've been reasonably happy with the MS engine controllers. They have a few glitches but overall work pretty well. Nowhere near as stable as industrial controls of course, but again one advantage is that much of the programming work has been done already. Jim

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now
Sign in to follow this  
Followers 0