Sign in to follow this  
Followers 0
alpha roni

[Help] How to build Belt weighing System using SLC 500

12 posts in this topic

Hi All, I want to build a Belt Weighing system based on SLC 500 PLC. The belt will continuously run and transfer bag, and the load cell will continuously measure the bag weight, then the SLC 500 will calculate the bag weight (either from the max value of its value or with other statistics calculation) and PanelView will show the final bag weight. I have some problems here: 1. I want to build digital-full system, in which I can do tare, zero, net (tare) weight, gross weight from PanelView. What kind of component should I take? Can I use common transmitter? 2. What kind of signal conditioner or transmitter so the PLC may read 4-20 mA or -/+ 10 Vdc from this load cell? 3. What kind of calculation in SLC 500 so we may have the exact (best representative) value of this bag weight? in this case the system is continuously sampling the bag weight so there are a lot of value of each bag. 4. How is the callibration (zero, span) work in PLC? I still can't figure it out. Any suggestion, tutorial, and link are welcome. Thanks in advance for help.

Share this post


Link to post
Share on other sites
You may look at something like this... http://www.ab.com/db/encompass/bps_dir_upd...region=AMERICAS

Share this post


Link to post
Share on other sites
Another option would be this: http://www.hardyinst.com/process_weighing/...gh+scale+module

Share this post


Link to post
Share on other sites
Take a look at rice lake systems http://www.ricelake.com I have used them on several projects. Good pricing and tech support - AB interfaces

Share this post


Link to post
Share on other sites
The mechanical design of the conveyor is perhaps the most crucial thing to build your own check weigher. You must have little or no vibration on the load cells due to the conveyor drive mechanism. Also, you will likely have two or four load cells that must be summed (or averaged by the weigh module/input card). I would look at an existing check weigher design in action if possible for ideas. You will find that as the bag passes over the conveyor, that the weight reading forms something like a trapezoid when plotted over time. For that reason, you will need to calculate when the bag is centered on the conveyor and take readings only during that time, and average them. So, you'll need a photocell to detect the bag as it enters the conveyor, and accurate speed information of the belt so this can be calculated. I 2nd the recommendation for the Hardy Instruments weigh module. I think the high accuracy and weighing system specific features will pay off for you. We have several Metler Toledo check weighers that are pretty impressive machines, but they're expensive and (being obsolete models) difficult to maintain and integrate with our SLCs. I, too have been considering rolling my own since I have a lot of experience with load cells in web handling applications, and a running Checkmate scale I can look at for mechanical ideas. One thing I notice is that the conveyor frame is completely free of vibration even while running, and that stands out as critical for success in my opinion. Paul

Share this post


Link to post
Share on other sites
You've received a lot of good feedback and I assume you know something about what you're trying to do, but I'll state one obvious fact that is often misunderstood about "flying weight" systems. The single pass flying weight almost never equals the static weight of the same package. The best you can hope for is a very small mean deviation between flying weights and be able to approximate the statc from the flying.

Share this post


Link to post
Share on other sites
About the macahnical design, I'll use ex-ventomatic Belt Weigher that was used by my client. So I think the mechanical will be OK. Yes, I will use two load cells and there'll be two photocells to indicate bag eneter and bag out. About the calculation, measuring bag when it's on the center of the conveyor is an option. Another option is I'll have some of samplings values (being latched) during bag is on the conveyor, and I'll compare them and choose the greatest one. Is handy instrument an easy-to-programmed input module? what is the output of this module? a raw value of weight or the final value (the greates one)? Thank you very much for your help.

Share this post


Link to post
Share on other sites
Load cells use a Wheatstone bridge configuration. If you don't use a scale directly (build your own), then you need a millivolt input card, which is exactly what a thermocouple or RTD card does. You also need a DC power supply for the other side of the bridge. I suggest you read up on how strain gauges (and load cells) work to understand this. It is a fairly simple process to interface directly to a load cell. A weigh belt system on the other hand is NOT simple from a mechanical design point of view. I highly recommend you enlist someone experienced in this rather than build it yourself. It is well worth the money. You will get a raw millivolt reading. You will have to calibrate this reading (in raw counts), just like with a real scale. The goal here is to calculate a linear regression line using good old math. Look up the key word linear regression. Let's take for instance a 2-point calibration. In this case, we get two sets of readings. The first reading is weight 1 and raw reading 1, call them cal1 and raw1. The second will be cal2 and raw2. Whether cal1 is "0" or cal2 is "0" is immaterial because it works the same way. Now calculate m=(cal1-cal2)/(raw1-raw2). This gives you m in the formula Y=mX+b. Now insert the values from the first calibration point and solve to get b (b=cal1/(m*raw1)). b also represents "zero" and m represents "span". So now you can calculate the weight from a raw reading at any point by calculating weight = m*raw + b. These formulas can be calculated using the standard arithmetic instructions (ADD, SUB, MUL, DIV) using floating point algebra. It is possible to scale this using SCP (scale with parameters) but in this particular instance, SCP is much more difficult to use because the calibration weights will determine the absolute maximum and minimum readings and it is usually difficult to come up with calibration weights that are actually outside the operating range (especially when one calibration point has to be less than zero weight). Tare/zero/net in a real scale are simply ways that the scale stores a set point from the raw readings. The "scale motion" type stuff is done by monitoring the amount of noise (such as measuring the differences of period readings) and using an arbitrary cosntant to decide when the weight is "constant". If anything of this sounds at all intimidating, buy a good Rice Lake unit and use that scale to do all the work. It already has everything you need programmed in it and your work load is limited to interfacing to the scale head. In a pinch you can even use it as a manual backup system.

Share this post


Link to post
Share on other sites
The Hardy Instrument module manual is here (LINK). It uses a 20 bit A/D converter and has an excellent assortment of features designed for weighing applications including vibration rejection. I think it would be a wise investment for your system rather than using the RTD card. The 1746-WS is fully configurable and programmable through the backplane, and can return 32 bit weight values. So, there is some complexity to set it up, but the manual is very good. Paul

Share this post


Link to post
Share on other sites
Thank you for brief explanation. In your calculation about callibration, should I pick a module like Hardy Instrument or I may use common load cell transmitter (like m-system)? I want to be able to callibrate the Belt Weigher digitally, from the PLC or HMI, so I need not to drive the screw in m-system transmitter. May I use this kind of transmitter? Thank you

Share this post


Link to post
Share on other sites
First, you can skip buying ANY transmitter and use a load cell directly if you use a thermocouple/RTD input module to get the reading directly. Using a transmitter to convert the signal to 4-20mA has really no advantage other than saving on input cards if you already have an extra 4-20mA input available. However, there is a huge disadvantage to this. You are limited to the accuracy/precision of the input card. For most AB cards, it will be either 12 or 16 bit. With most weigh bridges such as Hardy, it will be 20+ bits. This can drastically increase the precision and/or accuracy of your system. You will need to do some math to figure out what the required accuracy is and what you can achieve with your intended system. In other words, the load cell will have a specification such as the number of millivolts output per unit of weight. The input card will be able to measure a certain number of millivolts per count, or will give a full scale range and the number of bits of accuracy (such as 12 bits) and you will need to calculate <full scale range in millivolts> / <2^bits>, and then divide to get the units of mass per count, which is the minimum precision you can measure (and do some similar calculations to determine accuracy), only considering the load cell and not the mechanical limitations of the overall system. When going from 16 to 20 bits for instance, the precision increases by a factor of 16 (2^4=16). It is usually possible to program all kinds of configurations into a scale head such as the ones from Hardy to give you raw counts if you desire or to do some of the math for you as well. I just always did it with a scale head. The biggest reason that I've ever even considered using an input card directly is that most of my applications involved multiple load cells (4-8), and I always thought it would drastically help out troubleshooting by being able to directly measure every load cell individually instead of getting an average or summed value.

Share this post


Link to post
Share on other sites
Pauleng is correct. To eliminate a lot of pain and excessive problems, a scale head is an excellent approach. We had a processing plant that had numerous weighing systems tied in to a SLC500 system as discussed. We had ongoing problems so during a controls upgrade we installed Rice lakes Weighing systems and used their AB interface for ethernet/IP. It has resolved our problems and provided an easier way for the techs to maintain calibration using the Rice Lakes Units.

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