Post by thread:[EE] require Ultrasonic anticollision system

anti collision system is to be used in loco moving on same rail between points in a small area... Temperature is very high around 80-90 degree celcius.and enviornment is dusty.. movement of the loco would be around 3 km per hr. suggestions required...:) On Wed, 15 Nov 2006 Jinx wrote : > > > I am looking for ultrasonic or say any other sensor for > > anticollision system for loco > >000 or Orient Express ? >

> anti collision system is to be used in loco moving on same rail between > points in a small area... Ah, OK, that gives some idea of scale. If you went with ultrasonics, probably not a bad choice, fairly simple to implement I think I'd tend to over-engineer (more-than-adequately-engineer ?) this and use a piezo tweeter for the transmitter, which would need 40kHz power drive. Quite doable, the basis of ultrasonic cleaners. Something bigger than the Tx/Rx pair you can buy for robotics anyway. You might even code the 40kHz to ensure good detection and test both reflection time and signal identity It's possible that a car wrecker might have an ultrasonic reversing unit found in a lot of late model vehicles, could be a cheap way to experiment

Alan B. Pearce wrote: >>> Temperature is very high around 80-90 degree celcius >>> >> That's going to be quite punishing on electronics. Is this a >> situation where a Peltier could be used ? >> > > Sounds like he is setting out to make a positioning system for use in a > steel works or similar where a train needs to position crucible hoppers > under a furnace outlet. > > You will have trouble with reliability at 90C. I'd start off by trying to get the environment below 50C before I'd do anything else. Water cooling, etc whatever. This reminds me of downhole (oil drilling) instruments, THEY have to operate at 90C. But its a very costly set of instruments. --Bob

Ya we coukld have enclouser to bring down the temp...but want some stuff that can gimme distance ...and reliable,,,, On Wed, 15 Nov 2006 Bob Axtell wrote : {Quote hidden}>Alan B. Pearce wrote: > >>> Temperature is very high around 80-90 degree celcius > >>> > >> That's going to be quite punishing on electronics. Is this a > >> situation where a Peltier could be used ? > >> > > > > Sounds like he is setting out to make a positioning system for use in a > > steel works or similar where a train needs to position crucible hoppers > > under a furnace outlet. > > > > >You will have trouble with reliability at 90C. I'd start off by trying >to get the environment below 50C >before I'd do anything else. Water cooling, etc whatever. > >This reminds me of downhole (oil drilling) instruments, THEY have to >operate at 90C. But its a very >costly set of instruments. > >--Bob

Binu Pillai <binu_s_pillai <at> rediffmail.com> writes: > Hello All > I am looking for ultrasonic or say any other sensor for anticollision system > for loco moving at a very very > slow speed. Is there any suggestion for any good product for the same. What is 'very reliable' (what happens if it kills one in 200 workers) and how much time do you have to react. Since everything is made of metal in there magnetic sensing seems to be the best way to go, apart from wiring the rail side with a magnetic loop indexing system. Works even in darkness and under water ... why does it have to be ultrasound ? If the thing that moves at 3km/h has to make a decision based on the sensor reading how much time does it have for that ? Iow, if it's a 40ton load at 3km/h you will have to look ahead for about 10 meters at least if you want to stop in time. Ultrasound will not work well at 10 meters in a noisy environment with heavy vibration of the equipment (worsens the response and dithers the return echo out of the receiver's tuned band due to doppler). Dust in the air disperses ultrasound (steam does it too) by the way. Also how big is the something you are about to hit and what is it made of. Metal car ? Person ? Animal ? A wire ? A chain ? A metal rod ? So what do you need: a) a reliable distance measurement system, reading in what (10cm/10 meters is 1% - this begins to be seriously accurate and can be made reliable only by indexing the way somehow - probably inductively) b) a reliable relative distance measurement (harder than above) c) a reliable way to know your loco is about to hit something 10 meters away with +/- 50cm precision (5% - fairly easy but more likely with microwave than with ultrasound). d) other Electronics work great at 90deg C and above but usually using extended temperature parts, and no electrolytic caps. This is not a serious limitation for what you need imho. Vibration is a much worse limitation, and metal dust can be a big problem together with condensing steam and sulfur gases. hope this helps, Peter

Bob Axtell wrote: > Alan B. Pearce wrote: > >>>> Temperature is very high around 80-90 degree celcius >>>> I haven't been paying attention. Did this resolve? The reason I was concerned with temperature is that most folks don't pay attention to how most instrumentation handles heat normally. If you start OUT at 90C, then you simply have no reserve for normal device heating. You have to pay close attention to what your resistors are doing, your output drivers. Example: you design a normal 25C product. Inside the case, due to normal heating of the device, poor circulation, etc you get to 35C very easily. If you start out at 90C, where can the heat go? That's why I was concerned with temperature. I have designed sewer pipe robots that have to operate at 40C, and there was a decreased lifetime (enclosed DC gearmotor in a sealed case) due to trapped heat. --Bob

Very true Bob...the temperature parameter has to be considered seriously....the datasheet gives a fairly good discription of the temp-othere parameter response... On Thu, 16 Nov 2006 Bob Axtell wrote : {Quote hidden}>Bob Axtell wrote: > > Alan B. Pearce wrote: > > > >>>> Temperature is very high around 80-90 degree celcius > >>>> >I haven't been paying attention. Did this resolve? > >The reason I was concerned with temperature is that most folks don't pay >attention >to how most instrumentation handles heat normally. If you start OUT at >90C, then >you simply have no reserve for normal device heating. You have to pay >close attention >to what your resistors are doing, your output drivers. > >Example: you design a normal 25C product. Inside the case, due to normal >heating of the device, >poor circulation, etc you get to 35C very easily. If you start out at >90C, where can the heat go? > >That's why I was concerned with temperature. I have designed sewer pipe >robots that have to >operate at 40C, and there was a decreased lifetime (enclosed DC >gearmotor in a sealed case) >due to trapped heat. > >--Bob

yes momentum would be massive...But braking system can be connected to this system if i can give relay output.Right now we are controllig the whole system over Wireless...so just have to add one more interlock to the system... 1)distance of 10m would be critical and it has to stop at that distance. 2)distance of 15m can give an alarm. 3)breaking distance is 3.5mts On Thu, 16 Nov 2006 Alan B.Pearce wrote : > >I haven't been paying attention. Did this resolve? > >I don't think the OP has resolved anything at all about the project, and the >least worry at the moment seems to be temperature. Current discussion seems >to revolve around required accuracy, and how to stop a train with >considerable momentum with a tight position accuracy. >

Binu Pillai <binu_s_pillai <at> rediffmail.com> writes: > yes momentum would be massive...But braking system can be connected to this > system if i can give relay > output.Right now we are controllig the whole system over Wireless...so just have > to add one more > interlock to the system... > 1)distance of 10m would be critical and it has to stop at that distance. > 2)distance of 15m can give an alarm. > 3)breaking distance is 3.5mts I think that it is worth trying an industrial microwave doorway opener. These have a distance setpoint and some have vanes that help to direct the beam where it is needed. You can also add a dielectric lens to make the beam do what you need. Here is an example of device: http://www.bircheramerica.com/industrial-door/microwave-sensors.htm (I have not used this, but similar ones, and there are others) While this is 'just' a Doppler motion detector it can be adjusted to react at a desired distance for an object of given reflectance. Peter P.