I have two desk lamps. They can be seen here:
http://www.blakespot.com/list/images/g5_desktop.jpg
The right lamp, when I turn it on, usually causes one or both of the 20x4 LCDs I'm running (I am running two IO-Warrior 24s and two LCDs) to freeze up and the backlight goes dark. Sometimes just one goes down, other times neither. Happens when I turn on the lamp.
The lamp is plugged into a different outlet on a different fuse than the computer the IO-Warriors are plugged into. The lamp is not touching the IO-Warriors or their cables. Is it due to my using plastic shielding (not meta) for the IO-Warriors and LCDs?
See pic:
http://www.bytecellar.com/media/dual_status_lcd.jpg
Please...any advice appreciated! Will this damage the IO-Warriors?
blakespot
desk lamp blanks out IO-Warrior 24...
Moderator: Guido Körber
-
Guido Körber
- Site Admin
- Posts: 2856
- Joined: Tue Nov 25, 2003 10:25 pm
- Location: Germany/Berlin
- Contact:
Whoa, how long are those cables to the LCDs? The signalling on them is running at a couple 100kHz. Anything beyond a foot or 30cm is really critical.
Since the lamps have quite large transformers in their bases (at least they look like it) there is going to be quite some magentic pulse when you switch them on plus there is going to be a constant AC magnetic field around them, though that is going to be lower than what you get on switching on.
Since the lamps have quite large transformers in their bases (at least they look like it) there is going to be quite some magentic pulse when you switch them on plus there is going to be a constant AC magnetic field around them, though that is going to be lower than what you get on switching on.
The ribbon cable on one is about 2.5 feet. The other's is about 1.25 feet. Do you think keeping the cables under a foot would help the situation? It's the ribbon cable and not the unshielded LCD or IO-Warrior 24 units that are picking up the pulse, you'd say?Guido Körber wrote:Whoa, how long are those cables to the LCDs? The signalling on them is running at a couple 100kHz. Anything beyond a foot or 30cm is really critical.
Since the lamps have quite large transformers in their bases (at least they look like it) there is going to be quite some magentic pulse when you switch them on plus there is going to be a constant AC magnetic field around them, though that is going to be lower than what you get on switching on.
Do you feel that, as is, this is a damage risk to the IO-Warriors, the LCDs or even the Mac? Thanks.
(And indeed, the base of these lamps is a very heavy transformer.)
blakespot
-
Guido Körber
- Site Admin
- Posts: 2856
- Joined: Tue Nov 25, 2003 10:25 pm
- Location: Germany/Berlin
- Contact:
With EMC it is always hard to say what is the cause of the problem without elaborate measurements. But as a rule of thumb you should avoid any long cables and have (proerly done) shielding where ever possible. The IO-Warrior starterkits are designed for the test bench, when you use them in a typical home or office environment it is hard to say what kind of interference they will pick up.
The Mac is likely not endagered by your setup. One of the main causes for trouble is the cable elngth in itself. The interface signals of IO-Warrior are not made for long connections. Long connections add a couple parasitic effects, the most obvious is the resistance of the cable but ther is also capacitance and inductance. All of which together results in a degration of the signal.
Since IO-Warrior does not drive the high level of the outputs but rather uses internal pull up resistors it is in a bad position for driving any longer cable. On the other hand the more current you dump into a line to get past the parasitic effects of the cable the more RF you emit and the more likely you get a reflection of the signal which results in overshoots and undershoots (i.e. signal going above or below supply voltage lines, also known as ringing).
So the first rule is: If you can solder the LCD right onto the IO-Warrior board then do it.
The Mac is likely not endagered by your setup. One of the main causes for trouble is the cable elngth in itself. The interface signals of IO-Warrior are not made for long connections. Long connections add a couple parasitic effects, the most obvious is the resistance of the cable but ther is also capacitance and inductance. All of which together results in a degration of the signal.
Since IO-Warrior does not drive the high level of the outputs but rather uses internal pull up resistors it is in a bad position for driving any longer cable. On the other hand the more current you dump into a line to get past the parasitic effects of the cable the more RF you emit and the more likely you get a reflection of the signal which results in overshoots and undershoots (i.e. signal going above or below supply voltage lines, also known as ringing).
So the first rule is: If you can solder the LCD right onto the IO-Warrior board then do it.
-
Guido Körber
- Site Admin
- Posts: 2856
- Joined: Tue Nov 25, 2003 10:25 pm
- Location: Germany/Berlin
- Contact:
Thanks. Also ... and I know IO-Warrior is for the testbench as you've said ... but a friend of mine tells me one NEVER has electronics in an ungrounded case. My case is completely plastic. But...isn't the IO-Warrior 24 grounded through the USB connection to my Mac? Would the LCD not similarly be grounded through the IO-Warrior (then through the Mac)?
Thanks.
blakespot
Thanks.
blakespot
-
Guido Körber
- Site Admin
- Posts: 2856
- Joined: Tue Nov 25, 2003 10:25 pm
- Location: Germany/Berlin
- Contact:
The starter kit is meant for the test bench, the chips do fine even in industrial environemnts.
If electronics need to be shielded depends on many things. Two of the questions are how much RF energy the circuit does emit and how much can it take from the outside.
The ground connection is one thing, simply connecting ground to a shield of a case can even cause more harm than good. It all depends on how it is done and what the surrounding circumstances are. One really basic rule is that any shield may connect to ground at only one point. if it does connect at more than one point it is no longer a shield, it becomes part of the ground connection and since that is then redundant it is part of a winding of a coil.
What your friend was refering to was most likely a protective ground. This is something you do on mains powered devices that have a metal case. To make sure that there can be no mains voltage on the case if the cable is broken you connect the case to protective ground which ensures that the fuse will blow if mains connects with the case.
Protective ground can also be used for shielding to some degree. But as I said before that is not always necessary.
To cover all this in detail fills a book or two...
Just as a basic rule make sure to use short cables. For the USB make sure to use real USB cables. There are many cheap ones in the market that you caan maybe use to dry your laundry on but should not let close to any computer. A real USB cable is about 5mm thick and has a braided shield layer which you can usually see through the outer plastic.
The thin, highly flexible USB cables are good for mice where the cable length stays short (USB spec allows a simpler unshielded cable for low speed devices and maximum 3m), but if you go for the maximum 5m cable it absolutely has to be the properly shielded one.
If you want to put a shielded case around your electronics you should connect the case only to the shield of the USB cable. But keep in mind that any unshielded cable coming out of that box is also a break in your shield that can feed RF into the box.
If electronics need to be shielded depends on many things. Two of the questions are how much RF energy the circuit does emit and how much can it take from the outside.
The ground connection is one thing, simply connecting ground to a shield of a case can even cause more harm than good. It all depends on how it is done and what the surrounding circumstances are. One really basic rule is that any shield may connect to ground at only one point. if it does connect at more than one point it is no longer a shield, it becomes part of the ground connection and since that is then redundant it is part of a winding of a coil.
What your friend was refering to was most likely a protective ground. This is something you do on mains powered devices that have a metal case. To make sure that there can be no mains voltage on the case if the cable is broken you connect the case to protective ground which ensures that the fuse will blow if mains connects with the case.
Protective ground can also be used for shielding to some degree. But as I said before that is not always necessary.
To cover all this in detail fills a book or two...
Just as a basic rule make sure to use short cables. For the USB make sure to use real USB cables. There are many cheap ones in the market that you caan maybe use to dry your laundry on but should not let close to any computer. A real USB cable is about 5mm thick and has a braided shield layer which you can usually see through the outer plastic.
The thin, highly flexible USB cables are good for mice where the cable length stays short (USB spec allows a simpler unshielded cable for low speed devices and maximum 3m), but if you go for the maximum 5m cable it absolutely has to be the properly shielded one.
If you want to put a shielded case around your electronics you should connect the case only to the shield of the USB cable. But keep in mind that any unshielded cable coming out of that box is also a break in your shield that can feed RF into the box.