Quote:
Originally Posted by Daniel_LaFleur
I guess we are just going to have to agree to disagree 
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. |
Agreed and working on it.
Quote:
Originally Posted by Daniel_LaFleur
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.
|
There are some things that do need to fit under an umbrella. I have discussed certain items with a few people I know at UL and in order for this device to be considered safe it must be contained inside a housing, like the enclosure specified for the light source in the KOP and be electrically protected (fused as you have pointed out). My concern is that power supplied by the main battery and fuse panel is on until the main breaker is opened. Often this does not occur until after the robot is carried off the field.
Quote:
Originally Posted by Daniel_LaFleur
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.
|
Daniel, don't forget that the luminent intensity specification is measured at 1 meter in the center of the eye's normal range (about 550 nM). White LEDs do have a broad peak in this range that is less than half as intense as the sharp peak at 470 nM. So even accounting for the inverse square law at 2 meters, the source at 470nM is still very bright. Obviously the further away the object is from the camera, the less effect it will have on the camera system as a smaller region of the pickup is illuminated. Remember that the pixel size for this camera is 8x9 micro meters and the entire image area is only 3x2.5 mm.
Now for further discussion...Dave, I expressed an opinion on custom circuit voltage convertors earlier in this thread. +/- 15 volt power supplies are very different than 1200 volt unloaded power supplies. Although the GDC has stated that fluorescent light sources are allowable, the implementation, type and use of any light source is subject to inspection. I am stating a general opinion of CCFL light sources based on the device questioned at the beginning of this post. As to the use of any light source on a robot that might interfere with the camera vision system, the decision is generally a consensus of opinions from the head ref, Lead Inspector, FTA and any FIRST official present. The height of the light source in 2006 and 2007 was less likely to be influenced by robot color or lighting due to the height of the target. (robot vision design should have prevented the camera from seeking a light source below 6 feet.) In 2005, the vision system was looking at the floor and that made inspections for interference a much more difficult task. As to good electrical design in this case, a 20 amp breaker to protect the wiring between the breaker panel and the
insulated enclosure followed by a 1 amp, fast blow, non-resettable, fuse inside the enclosure would,
in my opinion, be an appropriate method of safe CCFL/Inverter implementation.