Illegal?

[Brandon Holley]

didnt team 40 use LEDs, not cathodes

[Al Skierkiewicz]

Daniel,
Besides the fact that the inverter puts out a very high voltage and requires better insulation on wiring than that provided in the kit, there could be any number of reasons to prevent it's use. The inverter should be considered a custom circuit and therefore can only feed the RC under the robot rules.

<R61> All outputs from the custom circuits must connect to the Robot Controller or to other custom circuits.

You might claim that is a non-functional decoration but decorations can only draw power from the 12 volt main battery, the inverter is the power source to the lamp, not the main battery.

<R110> Decorations may draw power from the 12v electrical system as long as they are powered via a dedicated 20A or 30A circuit breaker and do not affect the operation of other control system components.

You might even argue that custom circuits can and frequently do have local power supply regulators and/or negative power supply generators, but again the voltage is not higher than the main battery and those devices still follow the rules for custom circuits.

The lamps themselves are not roboust as anyone who built lights in 2006 can attest. I would venture to guess that several hundred lamps gave their lives to an errant ball during last year's game. Broken lamps produce dust and slivered glass. I wouldn't want to do a belly check on a robot and lay down in a pile of phosphor and glass out on the field.

Small, intense, light sources play havoc with CCD cameras especially those without an iris mounted in front of the pickup. The particular device linked in the first post is a "black light" or ultra violet output device. Although it has some visible light output, there is considerable radiation outside the visible spectrum. CCD pickups are capable of detecting a rather wide spectrum including IR and UV and this device would likely cause some interference.

Sorry, but I have to stand by my original response. Really cool but not for use on an FRC robot.

[dtengineering]

Hmm... well... good points (as usual), Al. I would have assumed that properly protected blue or red cold cathode tubes would be okay (assuming green is the only target), but you've got a point about the transformer pumping the voltage up. Although I would suggest that some reasonable level of LED lighting in a colour (or spectrum range) different from any vision targets should be okay, I would suggest teams make a point of clearing up any potential misunderstandings in the Q&A forum next year before decorating a robot with shiny flashy things.

I would suggest to anyone terribly disappointed by the concept of having cold cathode tubes potentially nixed at tech inspection that they consider lightsticks, such as cyalume http://www.cyalume.com/ They are plastic, chemoluminescent tubes... crack them and shake them at the start of the day and they should last for several hours of cool glowing robotness on the field, and in the pits... without draining batteries, or exposing anyone to high voltages, or the potential for cracked glass on the playing carpet.

Jason

[Daniel_LaFleur]

Originally Posted by Al Skierkiewicz

Daniel,
Besides the fact that the inverter puts out a very high voltage and requires better insulation on wiring than that provided in the kit, there could be any number of reasons to prevent it's use. The inverter should be considered a custom circuit and therefore can only feed the RC under the robot rules.

<R61> All outputs from the custom circuits must connect to the Robot Controller or to other custom circuits.

You might claim that is a non-functional decoration but decorations can only draw power from the 12 volt main battery, the inverter is the power source to the lamp, not the main battery.

Again, I'd disagree. The entire lamp (including it's cold cathode bulb) is a COTS item and that single COTS item is powered soley (sp?) from the 12v battery (through a breaker)

Originally Posted by Al Skierkiewicz

<R110> Decorations may draw power from the 12v electrical system as long as they are powered via a dedicated 20A or 30A circuit breaker and do not affect the operation of other control system components.

You might even argue that custom circuits can and frequently do have local power supply regulators and/or negative power supply generators, but again the voltage is not higher than the main battery and those devices still follow the rules for custom circuits.

Wattage is wattage. In order to up the voltage you must reduce the current (through a constant resistance). There is no rule against having higher than 12 volts on the robot (or lower, which by your definition would not be allowed either), just that the electrical supply must come from the 12 volt battery.

As long as it is breakered properly, then electrically it should pass inspection, according to the letter of the rules.

Originally Posted by Al Skierkiewicz

The lamps themselves are not roboust as anyone who built lights in 2006 can attest. I would venture to guess that several hundred lamps gave their lives to an errant ball during last year's game. Broken lamps produce dust and slivered glass. I wouldn't want to do a belly check on a robot and lay down in a pile of phosphor and glass out on the field.

And here we agree. Which is why I said that it would have to be contained within the robot completely if it did break.

Originally Posted by Al Skierkiewicz

Small, intense, light sources play havoc with CCD cameras especially those without an iris mounted in front of the pickup. The particular device linked in the first post is a "black light" or ultra violet output device. Although it has some visible light output, there is considerable radiation outside the visible spectrum. CCD pickups are capable of detecting a rather wide spectrum including IR and UV and this device would likely cause some interference.

Superbright LEDs put out ~2000mcd (give or take some due to manufacturer and wavelength). 2000mcd, unless put directly in front of the camera (less than 6") does not have enough power to saturate the CCD elements within the camera. In addition, the camera is searching for a particular wavelength, which happends to be very close to one of it's primaries (green). Thus as long as the LED has very little green in it's composite the camera will easily filter it out. In fact, at a distance of more than 6" it will be able to filter it out better than the stadium spotlights.

Originally Posted by Al Skierkiewicz

Sorry, but I have to stand by my original response. Really cool but not for use on an FRC robot.

For cold cathode lamps, I agree. There is too much engineering to make tham safe and not enough return (other than the 'cool' factor) for use on a FRC robot.

For LEDs I disagree. I can see plenty of feedback information that could be useful here. But I would caution those who would use them to be prepared to prove that your setup does not affect the CMU camera in any way, and to be prepared to disable the LEDs should you be required to do so.

[Al Skierkiewicz]

Originally Posted by Daniel_LaFleur

Wattage is wattage...

As long as it is breakered properly, then electrically it should pass inspection, according to the letter of the rules.

Superbright LEDs put out ~2000mcd (give or take some due to manufacturer and wavelength). 2000mcd, unless put directly in front of the camera (less than 6") does not have enough power to saturate the CCD elements within the camera. In addition, the camera is searching for a particular wavelength, which happends to be very close to one of it's primaries (green). Thus as long as the LED has very little green in it's composite the camera will easily filter it out. In fact, at a distance of more than 6" it will be able to filter it out better than the stadium spotlights.

Daniel,
I have to caution you, you are mixing apples and oranges here. Wattage is not the issue. 680 volts to exposed skin on students, volunteers and mentors is the issue as well as 680 volts to robot frame when the lamp breaks and the cathode touches the frame. I will grant you it is low current but I bet you wouldn't like it if you touched the exposed voltage. There is a difference between safe primary wiring practices and exposed high voltage.

On the camera, I am speaking from my everyday television experience. Small bright light sources are especially nasty to CCD pickups because the the AGC circuits and other electronics tricks used to get a nice looking picture are confused by bright spots in a normally lit frame of video. There are no optical filters before the pickup so all light reaches the face of the pickup where there are individual pixels for the three colors. Most light sources are fairly broad in their spectrum but LEDs can have some rather narrow bandwidths. A white super bright LED for instance has a rather high peak in the blue spectrum and a rather broad peak in the green. That being said, a small bright light produces a sharp image on the face of the pickup which for a variety of reasons, leaks light through the surface of the sensor and also reflects light on the back of the lens which then in turn sends it back to adjacent areas of the pick up. Besides the effect it has on the pickup, the circuitry used for making video outputs and the adjacent leakage paths, the signal then has a significant impact on white balance and local heating on the face of the sensor. All of these things aggravate the ability of the sensor to determine the presence of the correct color light. Add to that the 60 degree field of view of the lens and you are now producing problems that most teams will not be able to overcome.

[Cuog]

From someone who has been zapped by a transformer from these things I must say it hurts like a .... well it hurts a lot is what I'm trying to say. Its not likely a lethal amount but it certainly cause issues when trying to remove a robot or get to the power switch on the bot. With using UV lighting, there is certainly a problem with the CMUcams, when I brought the CMU can home to work with it I had issues with it getting distracted by my computer which is lit with UV CCFLs. This occured at a distance of roughly 3 feet. It is also important to note that the CMUcams are quite sensitive to a number of wavelengths and things such as too much red or blue can affect its ability to see green. There is an up and down to using a color in the "center" of visible light in that it is strongest and easiest tosee but it is also the easiest to interfere with.

[Adam Y.]

Originally Posted by Cuog

Its not likely a lethal amount but it certainly cause issues when trying to remove a robot or get to the power switch on the bot.

What is the power rating of the ballast? According to Wikipedia 100 watt ballast is capable of killing someone at that voltage. Even at one watt the ballast is still dangerous because Wikipedia states that you can't let go. Unfortunately, the values appear to vary wildly between website to website on how much current can kill.

In order to up the voltage you must reduce the current (through a constant resistance). There is no rule against having higher than 12 volts on the robot (or lower, which by your definition would not be allowed either), just that the electrical supply must come from the 12 volt battery.

This doesn't sound right. You don't necessarily need a resistor to boost the voltage to higher than the supply. In fact you might be confusing the fact that a florescent tube actually has negative resistance at a certain point and a resistor is needed to ensure that it doesn't just keep on trying to draw more and more current.

[Cuog]

I say its not likely lethal as I'm still here posting this, although I pulled my hand back VERY quickly when I felt that shock.

[Al Skierkiewicz]

As a rule of thumb, GFI breakers are designed to trip at 5 ma. The minimum currents for death depend on what part of the body the current is flowing through, skin conduction, etc. Across the temples interfers with brain activity, across the upper chest obviously interferes with the millivolt heart impulses, across limbs and muscle damage will result. Anyone who has contacted the 48 volt battery line on a telephone will tell you it's no fun.

[Daniel_LaFleur]

First off, I'd like to thank you for your patience, and thoughtful discussion. On Most boards, I'd have been flamed down. This is why I love CD and FIRST so much. We can disagree, but we do so with facts and respect.

Now back to our regularly schedualed discussion

Originally Posted by Al Skierkiewicz

Daniel,
I have to caution you, you are mixing apples and oranges here. Wattage is not the issue. 680 volts to exposed skin on students, volunteers and mentors is the issue as well as 680 volts to robot frame when the lamp breaks and the cathode touches the frame. I will grant you it is low current but I bet you wouldn't like it if you touched the exposed voltage. There is a difference between safe primary wiring practices and exposed high voltage.

I do not believe I am mixing apples and oranges here. The 12V directly from the battery has the power to weld it's wires together. Power is power, wattage is wattage. When you setp up the voltage, you step down the current. In the case above the step up transformer is a 50:1.

So lets look at the circuit current at the 2 different voltage levels. Lets assume it's a 12 watt lamp (for ease of figures)

When looking for current Ohms law is I=P/E
Thus I = 12watts/12volts ----- or 1 Amp
Now at the 680 volt level
I=12Watts/680Volts ----- or 0.0176 Amps

In a short circuit The wattage and the current both climb until the current passes the reset point of the breaker. If it shorted to the chassis could it hurt ... yes ... and it could hurt you just as much as the 12 volt battery shorted directly to the chassis. Since your breaker is on the 12V input to the device the 680V would never pass 1 amp because the 12 volt supply would have already passed 20 amps.

Originally Posted by Al Skierkiewicz

On the camera, I am speaking from my everyday television experience. Small bright light sources are especially nasty to CCD pickups because the the AGC circuits and other electronics tricks used to get a nice looking picture are confused by bright spots in a normally lit frame of video. There are no optical filters before the pickup so all light reaches the face of the pickup where there are individual pixels for the three colors. Most light sources are fairly broad in their spectrum but LEDs can have some rather narrow bandwidths. A white super bright LED for instance has a rather high peak in the blue spectrum and a rather broad peak in the green. That being said, a small bright light produces a sharp image on the face of the pickup which for a variety of reasons, leaks light through the surface of the sensor and also reflects light on the back of the lens which then in turn sends it back to adjacent areas of the pick up. Besides the effect it has on the pickup, the circuitry used for making video outputs and the adjacent leakage paths, the signal then has a significant impact on white balance and local heating on the face of the sensor. All of these things aggravate the ability of the sensor to determine the presence of the correct color light. Add to that the 60 degree field of view of the lens and you are now producing problems that most teams will not be able to overcome.

OK. I'll propose a test. Take your digital camera (which has a CCD array just like the CMU camera) and find something in your house that has a bright LED (Your PC hard drive is a good example, but many household electronics have LEDs). Take a picture of it from beyond 1'. Did it wash out your digital camera? I'll bet you some donuts that it won't. Thats because the AGC in cameras works for each set of RGB CCD sets. So while 1 spot may be 'washed out' the rest of the picture is still clear. This would show up as a very small target with a very low confidence level.

[Al Skierkiewicz]

Daniel,
OK, we disagree. What I was alluding to as apples and oranges was your discussion on "wattage". Your discussion on conservation of energy is true if you remember the efficiency of the conversion is not 100%. A 100kV power line at 1 amp is 100kW. At the output of a 100kV to 100 V transformer, the current would be therefore 1000 amps less conversion losses.
The sensitivity of the human body and electronics is more concerned with voltage and available current. Although the main battery is capable of welding wire it is the current (400-600 amps in the robot battery) which is performing that action. At 12 volts, the human body does not produce enough conductance to make damaging current flow. What I am concerned with is the 680 volt output of the inverter (not transformer although one is used in the inverter circuit). According to specifications, the inverter will put out 680 volts loaded with two tubes at 5 ma. (The output current is limited by the series resistance of the secondary winding on the transformer) In reality when the tube(s) is broken, the open circuit voltage is likely much higher, say 1000 to 1200 volts. The internal design of the inverter circuit will not reflect a high current to the input 12 volt source so it would likely never trip the breaker. The internal impedance of the inverter might cause it to burn up before a 20 amp breaker would even get close to tripping on the primary. (remember that the nature of the breakers will reguire sustained currents greater than 166% of rating for several minutes before they would trip.) Either way, 680 volts (or higher) is a voltage that is very uncomfortable to say the least. It's effect on humans, pacemakers, or robot controllers remains to be seen but I think I can predict that the humans will survive but be very vocal while the the effect on pacemakers is not something I wish to test at a First event. The effect on the RC is predictably certain death as capacitors, integrated circuits, and board traces would be subjected to voltages much higher than their design.

Now on to the bright source and CCD pickups. Remember that the CCD produces an output voltage dependent on the light falling on the surface of the device. There is a point at which the light is so bright that the individual pickups are overloaded. As with any device, there is not an infinite dynamic range. There is a point at which too much light causes there to be an excess of photon to electron conversion. As the light continues (and the excess of electrons) a variety of problems occur. The light is transmitted through the layers to other pickup areas of the device and the temperature rises. (Please keep in mind that the energy density of an object is magnified by keeping the smae energy but presenting it to a much smaller area. Ant vs. magnifying lens) The transmitted light cuses errors in the other "pixels" and the localized heating causes additional electrons to be formed. Sort of a dominoe effect. Eventually the heating will cause adjacent "pixels" to saturate and if sustained, this can cause permanent damage to the device. Even after the light source is removed, there will be excess electrons for a period of time.
Small highlights do not cause the AGC to react so small, super bright objects, are not corrected by the AGC and the effect is that the output video will also be saturated or "clipped" at the maximum allowable output voltage determined by the designer. It is also important to point out that AGC does not act on pixels independently but rather the entire output signal. Should the light source cause sufficient disturbance that the AGC circuits react, the effect is to change the gain of the entire output signal, possibly causing the intended green object to be pushed into the black. What you would see looking at a monitor, is a bright white spot surrounded by black.

[Jessica Boucher]

Originally Posted by Brandon Holley

didnt team 40 use LEDs, not cathodes

Yes, we used red and white LEDs, covered with plastic shielding to dampen the brightness of the LEDs. They lit up when the robot was close to a tube.

[Elgin Clock]

Originally Posted by Jessica Boucher

Yes, we used red and white LEDs, covered with plastic shielding to dampen the brightness of the LEDs. They lit up when the robot was close to a tube.

I'm constantly impressed with how bright them suckers are even with a damper built over them. You need some shades to look at 40's bot most of the time. lol

Send me some specs and a link to where you picked them up Jess!!!
I want to brighten some things up as a side project...
AND... The Cavalier wants to be nice and bright for some CT summer night cruising.

[Brian Mocci]

195 used almost those exact fans on our 2006 bot. We had 2x 120mm fans with 4 blue LEDs cooling our transmissions CIMS. The fans passed inspection at a regional and at nationals.

[Daniel_LaFleur]

Originally Posted by Al Skierkiewicz

Daniel,
OK, we disagree. What I was alluding to as apples and oranges was your discussion on "wattage". Your discussion on conservation of energy is true if you remember the efficiency of the conversion is not 100%. A 100kV power line at 1 amp is 100kW. At the output of a 100kV to 100 V transformer, the current would be therefore 1000 amps less conversion losses.
The sensitivity of the human body and electronics is more concerned with voltage and available current. Although the main battery is capable of welding wire it is the current (400-600 amps in the robot battery) which is performing that action. At 12 volts, the human body does not produce enough conductance to make damaging current flow. What I am concerned with is the 680 volt output of the inverter (not transformer although one is used in the inverter circuit). According to specifications, the inverter will put out 680 volts loaded with two tubes at 5 ma. (The output current is limited by the series resistance of the secondary winding on the transformer) In reality when the tube(s) is broken, the open circuit voltage is likely much higher, say 1000 to 1200 volts. The internal design of the inverter circuit will not reflect a high current to the input 12 volt source so it would likely never trip the breaker. The internal impedance of the inverter might cause it to burn up before a 20 amp breaker would even get close to tripping on the primary. (remember that the nature of the breakers will reguire sustained currents greater than 166% of rating for several minutes before they would trip.) Either way, 680 volts (or higher) is a voltage that is very uncomfortable to say the least. It's effect on humans, pacemakers, or robot controllers remains to be seen but I think I can predict that the humans will survive but be very vocal while the the effect on pacemakers is not something I wish to test at a First event. The effect on the RC is predictably certain death as capacitors, integrated circuits, and board traces would be subjected to voltages much higher than their design.

I guess we are just going to have to agree to disagree

I would not disallow a cold cathode lamp. I would inspect it very closely to ensure that it cannot get damaged during a match, and if it was damaged it would be contained within the robot and safely shielded from any conductive surface. I would also do this for the battery, which I consider more of a danger considering that it is not (and cannot be) fuse or breaker protected.

TBH, First does a rather poor job of electronics safety on the robot. A cold cathode lamp should be fused (or non-resetable breaker,not a resetable breaker) with a 2 amp 12 volt (24 watt) or under fuse. The fact that any custom electronics must use at least a 20 amp resetable breaker (even if it's just driving an LED) leaves too much chance for injury. They should open up the requirements for electrical safety and require the teams to specifically state why they used each specific breaker.

Again, teams should design in safety and be able to explain the 'safety' features of their robot, not be dictated a 'one size fits all' answer.

Originally Posted by Al Skierkiewicz

Now on to the bright source and CCD pickups. Remember that the CCD produces an output voltage dependent on the light falling on the surface of the device. There is a point at which the light is so bright that the individual pickups are overloaded. As with any device, there is not an infinite dynamic range. There is a point at which too much light causes there to be an excess of photon to electron conversion. As the light continues (and the excess of electrons) a variety of problems occur. The light is transmitted through the layers to other pickup areas of the device and the temperature rises. (Please keep in mind that the energy density of an object is magnified by keeping the smae energy but presenting it to a much smaller area. Ant vs. magnifying lens) The transmitted light cuses errors in the other "pixels" and the localized heating causes additional electrons to be formed. Sort of a dominoe effect. Eventually the heating will cause adjacent "pixels" to saturate and if sustained, this can cause permanent damage to the device. Even after the light source is removed, there will be excess electrons for a period of time.
Small highlights do not cause the AGC to react so small, super bright objects, are not corrected by the AGC and the effect is that the output video will also be saturated or "clipped" at the maximum allowable output voltage determined by the designer. It is also important to point out that AGC does not act on pixels independently but rather the entire output signal. Should the light source cause sufficient disturbance that the AGC circuits react, the effect is to change the gain of the entire output signal, possibly causing the intended green object to be pushed into the black. What you would see looking at a monitor, is a bright white spot surrounded by black.

What you are saying is absolutely true, but there are 2 things that you are missing.

1> The 'super bright' LED averages around 2000-3000mcd in it's wavelength. At a distance beyond 6" (due to exponential energy loss for didtance) this does not have the luminecence to saturate most CCD arrays.
2> CCD color cameras (like the CMU camera) have 3 CCD elements per pixel. Each element is for 1 of the primary (Red, Green, Blue) color bands of light. Since the target lamp is fairly close to (#00FF00) process green, as long as you avoid LEDs that emit a large portion of their energy in that wavelength you should not affect the Camera. Red (628 nm) and Blue (472 nm) LEDs are good, while White (broad spectrum) and Green (525 nm) would be bad.

I know that Checkmate used LEDs on their robot last year with no interference with other robots.

Note to all who would put an LED on their robot: Be prepared to prove that the vision system will not be affected by the LED.