PDA

View Full Version : Consolidating PWM's to one serial cable


GMAdan
08-04-2007, 11:07 PM
Sorry if this question has been answered in a previous thread, but i have searched and i can't find my answer.

I heard and from what I read on ifi robotics' website one wire of the three wires on the (the red one) is not used because it provides power which is not needed in the system we use. So if this is true could I in the interest to keep the number of pins needed to a minimum just not connect the red wire? So in turn I would only need two wires per victor for the signal.

bear24rw
08-05-2007, 07:29 PM
Where did you see this information, i'v never heard that you only need sig and gnd.. but i agree it could be nicer to just have to run two wires

GMAdan
08-05-2007, 09:19 PM
from Ifi robotics
General Notes:

1. The red wire (center pin) on the PWM cable connected to the Victor is an open circuit. There is no reason to disconnect this wire in the cable. The Victor does not source or draw power on this line.

ZZII 527
08-05-2007, 09:20 PM
As far as I know this is true. Somewhere in the Victor documentation it says that it neither sinks nor sources current on the 5V line and does not need to be connected. I don't know if it violates any rules to just use two wires.

Edit: I guess it says it does not need to be disconnected. So not explicitly stated in the documentation, but still true that you technically don't need it.

Also makes me wonder if the signal and power ground are internally connected or isolated (I suspect the latter). Bad practice not to use a separate signal ground line, anyway, I would think.

Al Skierkiewicz
08-06-2007, 07:08 AM
The Victor does not use the power pin as the Victor derives it's power from the separate power input(s). Servos, however, still need the power input from that third pin. Don't forget that the center pin is tied to the 7.2 volt back up battery when the backup is connected.

GMAdan
08-06-2007, 11:46 AM
Thank you all for your replies. Just to make sure, this means I can exclude the red wire in my serial connections. I just to have to make sure to include them for the servos.

Qbranch
08-06-2007, 12:13 PM
I would suggest that you still put your ground wire for the PWM signal in your DB9 cable.... using common point grounding for signals like that when you have noisy compressors and motors usually introduces their noise into your small 5VDC signals that can cause wierd and hard to troubleshoot control problems.

Still, this would get you 8PWM sigals and one ground per cable for the Victor drives. Very interesting....

Did you also consider using CAT5 (Ethernet) cable? The RJ series connectors are quicker to disconnect/connect and are just as secure as the DB9 cables... but its pretty much a draw for which is better.

It'll be interesting to see how this turns out....

-q

GMAdan
08-06-2007, 12:26 PM
I would suggest that you still put your ground wire for the PWM signal in your DB9 cable.... using common point grounding for signals like that when you have noisy compressors and motors usually introduces their noise into your small 5VDC signals that can cause wierd and hard to troubleshoot control problems.

Still, this would get you 8PWM sigals and one ground per cable for the Victor drives. Very interesting....

Did you also consider using CAT5 (Ethernet) cable? The RJ series connectors are quicker to disconnect/connect and are just as secure as the DB9 cables... but its pretty much a draw for which is better.

It'll be interesting to see how this turns out....

-q

The CAT5 does sound like an interesting solution. Only problem i see with that is that the CAT5 wire to my memory ( i could be wrong) only has 8 pins. So that would make for a weird distribution of wiring.
Unless I put 4 victors per CAT5 wire (if i remove the ground) and then have a CAT5 wire just for grounding purposes, which may workout that I can still give each victor its own ground.
Thank you for the idea. Sorry i was kinda thinking out loud there.

Qbranch
08-06-2007, 12:51 PM
The CAT5 does sound like an interesting solution. Only problem i see with that is that the CAT5 wire to my memory ( i could be wrong) only has 8 pins. So that would make for a weird distribution of wiring.
Unless I put 4 victors per CAT5 wire (if i remove the ground) and then have a CAT5 wire just for grounding purposes, which may workout that I can still give each victor its own ground.
Thank you for the idea. Sorry i was kinda thinking out loud there.

No, your absolutely right. The CAT5 does have just that many connectors and would definitely make for wierd distribution... also CAT5 normally is single core and because of that it doesnt have a long mechanical life when it comes to bending/vibration, and running a seperate wire for grounding does introduce, again, the possibility for wiring in more noise (every wire is afterall an antenna weather we want it to be or not :rolleyes:).

So, if I were you, I'd stick with the DB9. Its a great solution..... think you might see some nifty connection boxes on 1024's robot next year... :cool:

-q

Al Skierkiewicz
08-06-2007, 01:40 PM
also CAT5 normally is single core and because of that it doesnt have a long mechanical life when it comes to bending/vibration, and running a seperate wire for grounding does introduce, again, the possibility for wiring in more noise (every wire is afterall an antenna weather we want it to be or not :rolleyes:).

The correct term is solid wire and you are correct, solid is not a good choice for moving objects as it is prone to fail with repeated vibration. The problem with a single ground is not so much as an antenna as that the noise for each of the signals is added in the common return wire (ground...although not really ground). This is sometimes referred to as "ground loop". In the case of multiple Victors, motor noise will be added as well since there is not a perfect return for motor current to the battery for each branch circuit. Some motor current is bound to flow on the common PWM wire.