SOLO no longer detected by Windows

General discussion of using Roboclaw motor controllers
evangifford
Posts: 16
Joined: Tue Jun 09, 2020 11:32 am

SOLO no longer detected by Windows

Post by evangifford » Tue Jun 09, 2020 11:51 am

Hello,

I Have been working with a SOLO 30A for a few weeks now.

I'm using a bench power supply set to ~17V (to simulate a 4S battery pack) and a micro-usb connection to a Windows 10 Pro machine. The motor is being driven at <10A in closed loop with encoder and microswitch endstop (with 4k7 external pull-up on S3 and ENC A/B).

Today, while trying to troubleshoot a problem when HM1 is triggered, even though S3 is floating - I started having a new problem - the SOLO is no longer recognized by Windows!

Expected behavior (worked previously):
Plugging the SOLO into a USB port on a PC, it will be detected and display in Device manager as a COM port
and Motion Studio will detect it and allow configuration.

Actual behavior (new behavior):
Plugging the solo into a USB port on a PC, it is no longer detected by Windows in Device Manager (or Motion Studio)

The LED sequences are:

(powered up LED sequence, micro usb disconnected)
All LED on for (~500ms)
ER off, S1 green, S2 off (1 second)
After 1 second, ER Red, S1 off, S2 off (forever)

(Powered on, micro usb connected to PC)
All LED on for (~500ms)
ER off, S1 green, S2 off (forever)

To trouble-shoot, I've done the following:

Change USB port.
Connect directly to yet another port (not using hub).
Restart PC.
Change PC.
Take SOLO PCB out of metal housing and check micro USB connector pins under a microscope (look good, no cracking or loose USB connector)

I was working with it one minute, the next it was no longer detected. No short circuit, disconnect/reconnection of wires or any sign of overcurrent or damage, etc.

Is it possible to put it into DFU mode, or otherwise restore the firmware? What else can I do to troubleshoot this?

Thank you,
Evan

User avatar
Basicmicro Support
Posts: 1387
Joined: Thu Feb 26, 2015 9:45 pm

Re: SOLO no longer detected by Windows

Post by Basicmicro Support » Wed Jun 10, 2020 8:10 am

1. Test with a different USB cable.
2. Test with a different USB port on the PC.
3. Test with a different PC all together.

USB cables do fail. I've killed a dozen in the last 5 years easily so always check with a new cable.
USB ports can be damaged if not handling the motor controller correctly, eg disconnecting ground but leaving power will then potentially use the USB ground to complete the circuit and burn out the USB port ground.
Similar to above, The PCs entire USB system can be damaged if not handling the controller propler.

Assuming you have tested all of the above, the next likely cause would be a damaged USB connector on the Solo. You cant fix that but we can. In that case send the unit in to us with a note about the problem and we will confirm the problem and attempt to replace the USB connector.

The last possibility is the USB port on the Solo is damaged at the processor. This can also be repaired but there may be a small charge in this case.

Make sure you include your contact information in the note if you return the board for repair.

evangifford
Posts: 16
Joined: Tue Jun 09, 2020 11:32 am

Re: SOLO no longer detected by Windows

Post by evangifford » Wed Jun 10, 2020 8:20 am

Hello, thanks for your reply.

As I mentioned in my post, I believe I have already done all the troubleshooting steps you mentioned. I've even inspected the USB SMD solder joints with a microscope - they look ok (no cracks, the jack is firmly mounted to the PCB).

Is it possible to put a SOLO Into DFU mode and re-flash? It doesn't have a mode button like other models.

I am in France, so shipping will be significant, but possible.

Thanks,
Evan

evangifford
Posts: 16
Joined: Tue Jun 09, 2020 11:32 am

Re: SOLO no longer detected by Windows

Post by evangifford » Wed Jun 10, 2020 8:24 am

Regarding your comment about killing usb cables, I did try 3 different cables, no joy.
For the USB port, I plugged a generic FTDI into the same port as SOLO - The FTDI is detected by windows as VCP.

Even so, I did try using a different PC (which I have used successfully in the past with SOLO)

After this experience, I'd assume it would be best practice to always use a USB isolator when working with a RoboClaw?

evangifford
Posts: 16
Joined: Tue Jun 09, 2020 11:32 am

Re: SOLO no longer detected by Windows

Post by evangifford » Wed Jun 10, 2020 8:31 am

Also, does the fact I'm powering it with a bench power supply have any significance? This problem seemed to occur when I was performing forward and backward movement (using PWM slider in Motion Studio) with small ramps - so quick direction changes.

Also, I am not providing separate logic power, does that pose a problem?

Thanks!

User avatar
Basicmicro Support
Posts: 1387
Joined: Thu Feb 26, 2015 9:45 pm

Re: SOLO no longer detected by Windows

Post by Basicmicro Support » Thu Jun 11, 2020 8:04 am

When using a power supply you have to worry about regeneration from the motor causing an over voltage which can damage a lot of things including the USB port. Whether you caused an over voltage high enough to damage anything depends on what you were doing. It doesnt take much. Speeding up will never cause an over voltage. Its the slowing down part that does it(turns the motor into a generator). Even a motor running at 12v can easily pump the voltage up over 60v when slowing down using a power supply becasue power supplies do not sink current.

When using power supplies you either need to set the maximum main battery voltage to just above the power supply voltage(1 to 2v usually) so there isnt time for the voltage rise to speed up to the point it cant be handled or you need to use a voltage clamp.

Unfortunately, no. there is no way to force Solo into bootloader mode other than through a USB command.

A USB isolator would only protect the PC, not the Solo. About the only thing it would do to protect the Solo would be prevent a ground loop through the USB port, like when using a power supply. If you were running the motor and disconnected ground from the power supply the USB ground may have become the power ground instead.

evangifford
Posts: 16
Joined: Tue Jun 09, 2020 11:32 am

Re: SOLO no longer detected by Windows

Post by evangifford » Thu Jun 11, 2020 8:17 am

Ok, thanks for this useful information.

My bench power supply does shunt the inductive voltage spike - it even makes a little click when the protection circuit is triggered - but perhaps it's not fast enough.

Are there really no protection flyback diodes in the Motor A/B leads?

Would you agree this SOLO is dead? What should I do with it?

What can I do to prevent this in future?

Thanks,
Evan

User avatar
Basicmicro Support
Posts: 1387
Joined: Thu Feb 26, 2015 9:45 pm

Re: SOLO no longer detected by Windows

Post by Basicmicro Support » Fri Jun 12, 2020 8:28 am

That click is your power supply disconnecting itself from the load to protecting itself. It is NOT sinking the motor regeneration. A large voltage spike is still happening and is what probably killed your USB connection.

This is NOT an inductive spike. Inductive spikes last for nanoseconds to microseconds. Clamping diodes are designed for that much power. They are NOT designed for continous clamping over many seconds.

This is MOTOR REGENERATION. EG the motor is acting like a generator. This lasts for seconds at a time or longer(whenever the motor is slowing down it act as a generator. If that generation is higher than the internal resistance it will produce excess power which, if it isn't sunk by your power source(like to charge a battery back up) will instead cause a large voltage rise that will continue for the entire time your are slowing down. To sustain that much power dissipation the clamping didoes would have to be larger than all the mosfets on the board combine(and cost more than the entire controller as well). That is NOT how you deal with motor regen.

You need a voltage clamp which dissipates the energy in a large RESISTOR if you don't have a power source that can sink the power that is flowing back into it(eg a battery).

Alternatively you can put a battery in parrallel with your power supply(supply voltage needs to be set to the nominal battery voltage). Then any regen will go into charging up the battery. The excess charge will come back at once your motor loads the system down again.

You can send the Solo in and we will try to repair the Solo. If its repairable without significant time/parts cost we will only charge for return shipping. If its seriously damaged we will charge no more than 40% retail + shipping for repair/replacement.

evangifford
Posts: 16
Joined: Tue Jun 09, 2020 11:32 am

Re: SOLO no longer detected by Windows

Post by evangifford » Mon Jun 15, 2020 1:52 am

Ok, thanks for your thorough explanation.

Lesson learned - don't use a bench power supply without shunt diodes and power dissipation with a motor which can generate significant amount of energy (not inductive spikes, but generator).

Alternatively, I will put a battery in parallel with the power supply (at the same voltage as battery) to sink any generated energy and protect the SOLO (and USB port).

-

Caveat -

I wasn't expecting is this large amount of power generation from my setup - the motor I'm driving is a GEARED motor (12V, ~7A peak power draw, 1A typ) - with a 50:1 reduction - not a typical direct-drive motor with lots of potential inertia (like a scooter).

It's also free-running during the test with hardly any load. That's why I wasn't too concerned about dissipating the power and figured (incorrectly, apparently) that the fly backs would be sufficient.

Even in the final application with a full load, it would be physically impossible to move the motor in reverse as a generator without breaking the mechanics and gearbox.

As you explained, the bench power supply was not dissipating that power but protecting itself - so giving that voltage spike no where to go (I will measure this with an oscilloscope).

So there was a definitely a voltage return - but Is it still possible that what you described above applies to my case?
Last edited by evangifford on Mon Jun 15, 2020 2:30 am, edited 2 times in total.

User avatar
Basicmicro Support
Posts: 1387
Joined: Thu Feb 26, 2015 9:45 pm

Re: SOLO no longer detected by Windows

Post by Basicmicro Support » Mon Jun 15, 2020 8:52 am

1. You can also just set the maximum main battery voltage to 1 to 2v above your power supply voltage(use the voltage reading inside Motion Studio to determine the settings. This will in the majority of caases beable to prevent the regen by putting the motor into braking mode. Basically when slowing down if regen reaches that voltage limit it will brake the motor hard so its not perfect for all applications but it is the cheapest option(eg no extra cost).

2. I've got motors the size of my thumb. If I spin them up to speed and slow them down at the right speed I can cause my 12v bench top supply to jump to over 60v. I was using a high voltage driver for the test of course but it is very easy for regen to cause a prolonged overvoltage. You have to remember power supplly will not sink ANY power. So any regen above internal resistance will cause a rising voltage and that rise goes up exponentially. Catch it early and its not problem wait till it doubled and trippled the power supply voltage and even the electronics arent fast enough to stop it.

If the controller is damaged and you are using a power supply this is basically 100% the cause. I've heard of special power supply that will sink current(4 quadrant supplies) but I've never ever seen one. If you have one you would have paid an order of magnitude(or 2) more for it than for a regular supply. If your supply tripped to protect itself you know you went over its voltage limit. Since you have no voltage clamp circuit and you have adjusted the maximum main battery voltage limit yes. This is what happened.

Post Reply