Roomba 531 circle dance - BiT 13 failed

[ji101]

Hello, im new to this forums.

I am the owner of a Roomba 531, and it started dancing un circles. It gets off the home base (about half a meter) and then starts going un circles, clockwise, without moving forward or doing it very rarely.
I did some bulit in tests, and it seems it is not a sensor fault, because it recognizes my hand and slows down. And it passed tests 2 to 10 without problems.
When i ran wheel tests, BiT 12 is ok, it moves left wheel forward. But BiT 13 moves right wheel backwards, and i think this is wrong. Test 13 is ok as it moves both wheels backwards.
I replaced right wheel module but it made no difference at all.
Any ideas on how to proceed?

Thanks in advance

[vic7767]

Your Roomba is possibly suffering a failed H-Bridge transistor for the right wheel drive. You can test the motor by using a 9 volt battery. To test both directions of the motor just reverse the battery polarity. If the motor works in both directions then the problem is within the H-bridge.

[Gordon]

Hello, im new to this forums. ...I am the owner of a Roomba 531, and it started ... going un circles, clockwise, ... . I did some bulit in tests, and it seems it is not a sensor fault, because it recognizes my hand and slows down. And it passed tests 2 to 10 without problems.

FYI, BiTs 2 to 10 don't relate to wheel-motor operation, other than using certain data from that sensor set to help Roomba navigate while in Clean Mode.

When i ran wheel tests, BiT 12 is ok, it moves left wheel forward. But BiT 13 moves right wheel backwards, and i think this is wrong.

Quite "wrong".

Test 13 is ok as it moves both wheels backwards. ...I replaced right wheel module but it made no difference at all. ...

The three BiTs demonstrated the wheel motors to be operational. If, with the RH Wheel Module dismounted, you had applied a two polarities of dc test voltage to the RH wheel motor you would have seen it properly reverse wheel rotation. I.e., no need to replace the RH module.

Your 531's problem has to be a fault on its main printed circuit assembly (main_PCA) in a sub-circuit referred to as an H-bridge circuit, composed of many components. To date, you will not find a schematic diagram of the 500 series wheel-driver H-bridge on this rr site, since no one has taken time to do that work; and, that means we can't even surmise whether the fault is actually in the H-bridge section, or in the robot's micro-controller unit. However, that is a moot point since a single process is need to effect repair of the robot.

I think your only repair recourse is to procure a replacement main_PCB, and then R&R it into the 531.

[ji101]

Thank you both for your help.
I managed to open Roomba and took pictures of the board. Do you see some component broken or something wrong?

https://docs.google.com/folder/d/0B8vRx ... E4NGc/edit

[Gordon]

Thank you both for your help.
I managed to open Roomba and took pictures of the board. Do you see some component broken or something wrong? ...

Welcome. There is little chance that visual damage will have resulted (unless a component had been severely overheated), or seen in the photos you provided (especially at the scaled down resolution invoked by the pix-host). Your scenes ...121106-212308 and ...-212634 cover the bottom & top areas that include the Right Wheel Motor Driver's components. Anyone sleuthing this problem would need to get up close with optical magnification while looking for physical damage clues. Here are a couple pix (2nd is from your ...634 image) enclosing H-bridge components on bottom & top PCA sides:

Referring to that image, find on your main_PCA power transistors (xstrs) Q59 & Q60, Q75 & Q78, then include surrounding diodes D5? & D53 and D63 & D64 to define the power handling portion of the H-bridge. And then include four TO-92 cased xstrs that are left & right of the H-bridge (they interface the power section to the MCU ( in concert with some active top-side components)).

The 12 solder pads for the four power transistors are easy to spot, and tell you where the middle of most H-bridge components are located.

You will not find a a schematic diagram of 5XX wheel driver elex. AFAIK, no one has created one. ...{REVISED on 121215.} That is, none created prior to 121129. After that date it became possible to view R3-Roomba's right wheel H-bridge circuit. The link to that drawing is: viewtopic.php?f=1&t=16618&p=112091#p112091
By gaining an understanding of how U6 (quad NOR gate) controls power transistors of the motor-bridge it might just be possible to figure out whether one or more NOR-gates can fail in a certain way to always cause reverse wheel rotation no matter that forward rotation was commanded.

[ji101]

Thanks for your time and help. I guess i'll go for a PCB replacement. Do you think this would be ok?

http://www.ebay.com/itm/ROOMBA-510-530- ... 3f1d83789e

Or maybe I should go for a new one like this:
http://www.ebay.com/itm/NEW-iRobot-Room ... 3a7bf7ad29

I'm in Argentina, so my only choice is ebay i guess... Thanks again

[matofill]

I have exactly the same problem, but right wheel was death and i not made build it test.

I open the motherboard and found that 2 transistors in PCB are defective one NPN and one PNP, i change it from one another defective PCB (water problem ...) and robot 581 now works perfect !!! So there are only 4 transisitors and still 4 , and cca 4 diodes and if some electronic can check that, you can repair it !!!

I wrote all this because this post help me to solve my problem !

[Gordon]

I have exactly the same problem, but right wheel was death... I open the motherboard and found that 2 transistors in PCB are defective one NPN and one PNP, ...

matofill, please write what "Qnn" transistors you replaced. Were they: Q83 and Q63?

[Gordon]

Thanks for your time and help. I guess i'll go for a PCB replacement. Do you think this would be ok?
http://www.ebay.com/itm/ROOMBA-510-530- ... 3f1d83789e

That seller claims to have repaired the mobo he has for sale. He does not report what was wrong with the board, nor what components he replaced. I would not care to do business with him. OTOH, his listed price is lower than you will find elsewhere. If you are a gambler, then give this guy a trial.

Or maybe I should go for a new one like this:
http://www.ebay.com/itm/NEW-iRobot-Room ... 3a7bf7ad29 ...

Check the Buy / Sell / Trade forum to see if this eBay seller appears on the 'acceptable' or 'unacceptable' lists. The one seller that rr members have found trustworthy is "casino187". Here is his offering for some 5XX main_PCAs: http://www.ebay.com/itm/Roomba-500-Seri ... 27cc4a370e
Relatively expensive however!

[matofill]

I have exactly the same problem, but right wheel was death... I open the motherboard and found that 2 transistors in PCB are defective one NPN and one PNP, ...

matofill, please write what "Qnn" transistors you replaced. Were they: Q83 and Q63?

No, the death transistors was Q59 and Q75, i also tested another 6 transistors , 4 diodes and some resistor in another side of PCB, but all another elements was OK, only this two transistors was death.

And robot now worked already 6 cleaning missions and works perfect with the same wheels like before !

Soory for my bad English.....

[Gordon]

matofil, thanks for the "Q59 and Q75" report. I am surprised to see a mixed failure mode between those two!

Member latzka reported base-voltage measurements in another thread to demonstrate that: a) when PNP Q75 is switched into conduction mode (along with its NPN partner Q60) the wheel-motor shaft rotates in a direction that makes Roomba drive forward, but when NPN Q59, and its partner PNP Q76 switch on together, they conduct motor current that drives Roomba in reverse.

Something really weird caused two power xstrs on the same side of the H-bridge to fail!

... Were they: Q83 and Q63?

No, the death transistors was Q59 and Q75, ...

Today, I am wondering why I quoted left-wheel example transistors(?).