Sensing Different Metals Using Prox's

[kyle 0]

I am trying to sense two different metals using a prox. With this i am hoping to not have to waste a AC input card on one or two prox's. The two metals would be stainless steel and aluminum. I was thinking that an analog prox could be used so as to send out two differnt voltages for the different metals, but was uncertain if there is something out there to convert the signal to a DC signal without having to buy an AB A/D card. Anyone have any suggestions or soulutions?

[beegee 7]

I would try to use a standard prox but adjust it mechanically (distance) so it only sees the stainless steel and detect the presence with another sensor Aluminum has about 2/3 of the sensing distance of stainless steel. So with bigger diameter proxes (30mm) you should have about 8mm wit Alunminum and 12 with stainless steel. beegee

[gravitar 3]

Well when you say "sense two different metals" do you mean that, whether the target is aluminum or steel, you just want to turn on an input to tell you that something is there? Or do you want to discriminate between the two? One way would be like you mentioned, using an analog prox and using a limit test to decide which it is. You'd have to have pretty consistent sensing distance though, because it seems to me that you could have an aluminum partup close that would read in the same range as a steel one farther away. You mention "Waste".. How are you defining that? Are you REALLY short on discrete inputs, but have unused analog inputs? In terms of cost and design simplicity (Which is how I usually define waste :), I would think that a ferrous-selective and nonferrous-selective prox looking at the same target (but spaced far enough apart that they do not crosstalk, of course) would be the simplest and cheapest solution

[BITS N BYTES 66]

Have used Contrinex sensors for this type of application. Their sensing technology differs from the run of the mill proximity switched and may work for you. Give them a try! http://www.contrinex.com/

[paulengr 44]

You can buy DC discrete input prox switches too. Most plants are converting slowly from AC to DC at least on the input side due to the improved safety factor (potential for electrocution goes way down, grounded DC tends to not take everything down and/or cause as much damage or bodily injury, and total lack of potential cross-talk problems). In fact, they tend to be significantly cheaper since the prox switch doesn't have to have any sort of on board power supply. Are you actually using voltage signals instead of current loops? If so, why? Current loops automatically adjust for different impedances in the connections and wiring, so once you set one up, you never have to calibrate it in most circumstances. The only two reasons to use voltage-based analog signalling is for speed reasons (current loops can't respond fast enough above about 20K samples per second), or because that's the only option available on some instruments. If you can reliably gaurantee the position of the prox switch relative to the metal, and set your single prox switch just right (just out of the range of the aluminum), then you can detect either "stainless present", or "no stainless present". If you want to detect a third possibility..."no metal present", then this means you need two inputs REGARDLESS of how you set up your prox switches. Personally with the "finnicky" nature of this, I'd want both inputs anyways. It's not a waste if it eliminates a troubleshooting session. With "distance sensing" analog prox switches, watch out. Some of them are of the multiple coil design. They use two coils that are a set distance apart inside the sensing head to measure both the distance and the magnetic response of the material. They use the response to scale the distance so that the distance remains correct regardless of the material. In your case, this may be an advantage with 2 analog sensors (one for distance, one for material), or it may be an issue (it doesn't work anymore). If you want to "save" your analog inputs but still use an analog sensor, Action I/Q (or was it I/O?) among others make analog "switches". These can be set to close an output at a discrete voltage or current level. If you are reading an analog sensor but just need a discrete on/off and don't mind having lots of pots for electricians to fiddle with, this is another way to go. Since pots can and will get fiddled with and anything you ever set up will end up completely screwed up, that is the #1 reason that I avoid voltage-based inputs like the plague. Those two little pots will get screwed up in pretty short order by somebody that doesn't know what they are doing, even if it's a production employee that just witnessed an electrician doing it. Finally, sticking with AB, remember: with AB, there might be better stuff out there, but you'll never pay more! There are other options than AB's overpriced A/D cards (which are also fairly noisy as analog I/O cards go). AB A/D cards for the older (PLC-5 especially) PLC's are very expensive (some 1771 cards are over $2K now). But did you notice the programmable mixed remote I/O in the catalog? It is significantly more useful in most cases than an NDC card or a full block of Flex I/O or Point I/O because it provides a good mix of I/O similar to the Point I/O stuff and can be networked. It isn't obvious because for some reason AB thinks it's a PLC, not an I/O card. It's called a Micrologix 1100. It comes with two analog inputs, 10 discrete inputs, and 6 discrete outputs, for $550US list. The analog inputs kind of suck (voltage only...easily fixed with a metal film resistor), and 10 bit input only, but it is still 25% of the cost of most of the analog input card options from AB. It even comes with a convenient programmable LCD screen and 3 push buttons that you can use to create menus for any on-board configuration or diagnostics for your "smart IO". I have several of these in the plant that are being used as nothing more than "smart I/O". In fact one in particular currently does nothing but drive 5 outputs for a bunch of lights. It was a lot less expensive than trying to run ANY AB output cards to a remote spot to drive some signal lights. Just set up MSG blocks instead of block transfers or discrete I/O mapping and you're on your way. If you want to make the leap in a different direction, I have currently installed a network of I/O cards from Acromag (there are several other vendors) that speak Modbus/TCP, which is a "universal protocol". They are running 24/7, monitoring vibration and temperatures on a couple large (800 HP) fans without any hiccups whatsoever. They are connected to an AB PLC-5. If I did it with AB I/O, I'd end up with at least 4 input cards, totalling almost $10K (thermocouple/RTD cards are close to $2.5K now by the time you add I/O adapters if needed). With the approach I took, it's 4 input cards, two Ethernet switches, and a single Digi One IAP, using Modbus/TCP. Total cost was about $3K. If possible, Modbus is the way to go in terms of protocols. It lacks bells and whistles (unlike any of AB's newer protocols) but there is almost no device out there that doesn't supprt it. Every PLC that I know of has translators or can directly speak Modbus. On many PLC's (including AB), the code is relatively simple so you can write your own serial-port based conversion if you want to. There is even a Micrologix that speaks only Modbus and none of AB's own protocols for Remote I/O. I used a Digi One IAP to do the conversion from Modbus and Modbus/TCP to AB's older PLC-5/SLC 5/05 Ethernet protocols.