Notifications
Clear all

MD10C Question  

  RSS

huckOhio
(@huckohio)
Trusted Member
Joined: 1 year ago
Posts: 63
2020-09-16 1:17 am  

I am wiring the navigation layer for DB1, and I noticed some strange behavior that I hope someone can explain.  

IMG 4393

I was downloading the motor controller test code to the Motor Controller card (I included a connector for an FTDI board when I created the PCB) when I noticed the motors would start turning very slowly.  Without the 12V connected there should be no voltage going to the motors. 

I started troubleshooting and found ~.5V on the 12V line.  Through the process of elimination I found that the .5V is coming from the MD10C card.  The input to the MD10C is 12V so I am not sure where the .5V is coming from.  The picture below shows only one of the MD10C cards connected to the Power Distribution board and the meter probes across the 12V & GND input lines and I am reading .5V. 

IMG 4424

I have a battery pack connected to a 5V power supply connected to the 5V/GND input terminal connectors on the Power Distribution board.  The 12V input is open.  If you look closely you will also see the 5V fuse disconnected.  I checked a blank PCB and there is an open between 5V and 12V.  I disconnected all wires to the Power Distribution board and check to make sure there was an open between the 5V and 12V lines and there was.

Any suggestions on possible causes or what I should check next would be appreciated.  Thank you!


Quote
triform
(@triform)
Reputable Member
Joined: 1 year ago
Posts: 331
2020-09-16 1:28 pm  

@huckohio

It could be caused by a lot of things. Inductance, leakage, etc. Having a lot of wires running around, you can transfer or pickup voltage from other wires, motors, or other things in the environment. A wire is just an antenna 🙁 You will typically measure some sort of voltage in the mv range in setups like this. On most commercial robots you would have isolated power for the motors and the computers/controllers.  This may not be practical for most hobbyist uses.

On the cables, data, power, and otherwise, you may want to get some ferrite chokes to place on the wires to help some with crosstalk, etc.  I would also add a flyback diode to the motors. The MD10C may be leaking some voltage back in as well. You could try a diode on the motor's power in (you will lose a @.6-.8 v though) This will keep anything from at rest, leaking and I have had issues with that before myself. 


ReplyQuote
huckOhio
(@huckohio)
Trusted Member
Joined: 1 year ago
Posts: 63
2020-09-16 6:16 pm  

@triform

Thank you for the suggestions.  The interesting point is the 5V is only going from the terminal strip to one end of the 5V fuse.  No other boards have power, the wires you see have no power/signals on them.  12V is not connected.  I only see the .5V when the MD10C board is connected.  The strange thing is the 5V is only going to one side of an open fuse.  I have no idea how the .5V is jumping to the 12V line.  But when I remove the wires from the MD10C board this signal goes away.  It appears to be coming from the MD10C board, but that board has no power.  I really don't see a possible feedback look and there is virtually no current flow so not sure how inductance would cause this.  

I will use your suggestions as I continue to build DB1.


ReplyQuote
triform
(@triform)
Reputable Member
Joined: 1 year ago
Posts: 331
2020-09-17 3:21 am  

@huckohio

The MD10C may not be grounded well. Is this seen on both controllers or just the one? A floating ground can cause induced noise and again, wires are just antenna and you aren't talking about a lot of voltage. You can also try unplugging the motor and see if that changes things on the meter readings. 


ReplyQuote
huckOhio
(@huckohio)
Trusted Member
Joined: 1 year ago
Posts: 63
2020-09-17 2:25 pm  

@triform

Originally yes, it happened on both controllers.  This saga is getting weirder as I troubleshoot the problem 😤 .  Last night i replaced the MD10C board with a load resister and the voltage drop was 0.  I connected the MD10C and the voltage across the terminal (second picture) was still 0.  Nothing really changed - I've connected and disconnected the MD10C boards several times trying to isolate the problem.  At that point I just laughed and walked away.  Will check again today...maybe the gremlins will give me something else to change 😏 


ReplyQuote