Two wheels running at different speeds (Bee-bot) -loose connection?

[lucky67]

This is a Bee-bot -a programmable robot for young children. Goes backwards/forwards and turns 90 deg right or left. Has two wheels. Has started pulling to one side. It has been well used (in a playgroup) -so possibly something has worn out.
Took it to pieces to see if I could find something obvious - nothing on the circuit board looks fried but there are a lot of components. No major damage to the wheels/axles. There was a tiny bit of sand inside (usually occurrence, gets everywhere) but none really in the moving parts -mainly just in the base.
There are two dc motors one for each wheel, uses a series of gears for each wheel. gears have a light grease film - look to be running smoothly, no missing teeth etc. Axles are lightly greased. All seems in order.
Marking the wheels and watching them turn one wheel is moving faster than the other. Which is why it pulls to one side. This happens whether the wheels are going forwards or backwards. sometimes one wheel seems to almost stop. To turn 90 deg the wheels go in opposite directions - difficult to tell whether one wheel is slower or not.
However whilst I was looking inside one of the wires to one of the motors came loose - I wan't rough with it, didn't pull on it etc so I think it must have not been the best connection. I've soldered it back but probably not the best job I've ever done. Ideally I think I would need to replace the wire. Now I'm wondering if that could have been (still is) the problem. Could that cause the motor to run more slowly? How likely is that?
Or could one of the motors be getting worn out?
Or any other suggestion?
It might just be time to replace the whole thing -but if it could just be that connection I would try replacing the wire. (They aren't cheap to buy)
Thanks in advance for any help/advice.

 

[danadak]

A "loose" wire certainly could cause reduced motor current, hence torque/speed.

Also a visual inspection of motor drive transistors around that area of PCB to not show
significant heat discoloration around each motor driver transistor, which could indicate
further problems. And use a magnifying glass look at solder job on the board.

Poor joints -

Regards, Dana.

 

[lucky67]

Thanks- really helpful advice. I need to clear my desk and then I can have a proper look. I'll update here if I manage to fix it - or I find a 'concrete' reason for the problem.
Thanks again

 

[ronsimpson]

Sometime a carpet fiber (string) gets stuck around the shaft.

 

[lucky67]

Thanks -but both axles were fine. I did think it might be something like that - 'an easy fix'. But nothing obivously wrong - except that the connecting wire more or less fell off by itself. (Why it wasn't my best solder job ever - wasn't expecting to have to do any soldering, didn't really have time -thought it would be clear what was wrong and if it would be fixable or a write off...)

 

[rjenkinsgb]

According to this video, the motors have encoders of some sort to monitor their movement.
Possibly just slotted discs?

And, dirt in the sensors can cause the robot to run in a curve.

Can you see the encoder sensors??