??? 01/29/09 00:56 Modified: 01/29/09 01:00 Read: times |
#161803 - It's not as easy as one might think Responding to: ???'s previous message |
Chico Magalhães said:
Thanks for the answer Richard.
I'm not very experienced with electronics and microcontrollers, so, sometimes I cant understand, or take much time to understand... also my English is not very good. English isn't my first language either, so I certainly understand what you mean. I'm more articulate in English now, having spoken and written it for over 50 years, than in my native language, however. I think I proposed something, but I missed to give more information on what I need.
I want to use two HDD voice coils from an old drives to create a laser XY scanner. I think it don't need more than 10 ~ 16 bit resolution. Many beginners believe that all one needs to attain 10 - 16-bit resolution is a converter. It's never that simple. Consider for a moment how much power supply noise there is in a typical MCU-controlled system. 10 bits implies one part in 1000 or 5 mV per increment in a 5-volt system. 16 bits implies one part in 65536 or about 74 microvolts. A typical noise level in a microcontroller-based system is about 100 mV. If you have 100 mV of random noise on your ADC reference that's 50 millivolts of error. Now, there certainly are ways of doing that, but you should ask yourself whether you have the equipment with which to measure, accurately and precisely, how precise your reference is, and how much noise there is in YOUR system. Very experienced engineers have difficulty attaining 16-bit resolution in microcontroller-based systems. I believe that, with proper equipement and patience, you might attain that level of precision, too, but not easily. That is why I recommended a mostly digital approach. If you print a scale on bright white paper, that can be photographed on film, then you can photograph it with a macro-capable 35mm camera and, having reduced it by a factor, just guessing, of perhaps 10, you can use that as a digital signal source. You will place your optical sensor on the head assembly, behind yet another piece of film with a very fine slit in its dark surface, and allow it to move back an forth. You then apply a relatively uniform light source, compatible with your sensor, to the piece of exposed and developed film I mentioned, which film has to be affixed to whatever base you use, and whenever the sensor/slit pass one of the slits in the fixed film, light will fall on the sensor, generating a pulse, which, of course, you have to amplify and filter. That will tell you where you are, as you can count those pulses, and the magnitude of the voltage driving your voice coil will tell you the direction in which it is traveling. At first I don't want something professional, something simple is ok. How simple it can be depends very much on how precise it must be. I was thinking in some circuit that controls the HDD head by itself, also the feedback, the microcontroller just sends a analog value of the position. Is this ok? What the name for this type of circuit? I think the word you're looking for is "servo." That applies in general to any system that uses a feedback path. It can mean a number of things, however. But, if you think something driven by the microcontroller, also feedback, will be ok for me too.
If you have some idea, please tell me. Thanks. Well, Chico, you haven't given very much information about what you plan to achieve, and without that information it's hard to specify what you might attempt to do. BTW, How did that stepping motor work out for you? RE |
Topic | Author | Date |
Closed loop for HDD head coil help | 01/01/70 00:00 | |
Maybe it will work ... or maybe not. | 01/01/70 00:00 | |
Gradient transparency + led + photo-diode? | 01/01/70 00:00 | |
it will not be easy | 01/01/70 00:00 | |
more info | 01/01/70 00:00 | |
It's not as easy as one might think | 01/01/70 00:00 | |
harder than I thought | 01/01/70 00:00 | |
There are ways ... | 01/01/70 00:00 | |
speed is essential when drawing | 01/01/70 00:00 | |
Yes, faster is better in this case | 01/01/70 00:00 | |
Relativity | 01/01/70 00:00 | |
That would likely work | 01/01/70 00:00 | |
Flexible Thinking - Flexible Arm | 01/01/70 00:00 |