Roomba Red Charging Circuit Repair/Troubleshooting

[jwelser]

Hi all,

I am a long time lurker, long time Roomba tinkerer/fixer, but this is my first post because I am stumped.

I have a Roomba Red that has a bad internal charging circuit. It is out of warranty, so I am on my own with this one. Basically what happens is that when it is hooked up to its wall charger OR the charging home base, the battery light will flash red forever and never turn green. The battery is good, because I can put it in another Roomba Red, charge it up (and then even put it back in the "original" Roomba Red, which will then work fine.)

When I ran the diagnostics as detailed here:

http://mysite.verizon.net/gsplews/Diagn ... .v.0.4.htm

....I learned that tests 15 and 18 (verify charging voltage) PASSED, but tests 16 and 19 (verify sensing of charging current) FAILED. Since tests 16 and 19 failed, I wouldn't expect tests 17 and 20 to pass either, and they don't.

I am an electrical engineer, so I'm not afraid of tackling these kinds of problems. I read another post in this forum where someone found their charging problem to be caused by bad TVS devices. So I pulled the PCB and and took a look. The TVS devices looked physically OK (not cracked/discolored/burnt like those of other people who had problems,) and when I tested them with a DMM, they read in the 10s or 100s of kohms, which seems reasonable. I also tested all of the discrete diodes and transistors, and they seemed to test OK.

Now, I'm stumped. Without a schematic, I'm really not sure what to try next. Does anyone have any ideas about what I should try? Does anyone have a schematic (at least for the charging circuit?) I'd really appreciate any help/advice that anyone can give.

Thanks a lot,

Joe

[vic7767]

There is no charging circuit schematic available for the public. We Roomba owners are on our own there. Something you could try would be to reset your Roomba by removing the battery, hold down the pwr button for 30 seconds, thereby discharging the capacitors on the PWB. Install the battery and perform your charging setup either with the homebase or just the charger. Has your Red been OSMOed ? From what I understand, there were some minor software changes to the charging intelligence that are applied via OSMO.

[alepplew]

I think I figured out what is wrong with a SE that thinks it is charging (pulsating power LED) but does not. Some oscilloscope diagnostics clearly showed that the mosfet used as a charging current switch is shot in my case. This is also the part most likely to fail. If you take out the mainboard (lift it a little up together with the bumper sensor arms, then you can pry the arms out sidewards)
look at the right side of the circuit board. You will see two fat diodes at the right lower corner. Another fat diode is located some 5cm to the left of those. Below that diode is the mosfet. It is a SMD part with 4 soldering points, part no. STN3PF06. It can probably be replaced using a TO220 packed type, that is easier to solder and may accept more current. Make shure that it is a P type and the source drain resistance is below 200 mOhm. Here the data
TYPE VDSS RDS(on) ID
STN3PF06 60V <0.20Ohm 2.5A
The data sheet is available at http://www.datasheetcatalog.com/

The drain of that mosfet is connected to the positive lead to the battery through that fat diode that is polarized in the "charge" direction. The source is fed with appr. 20 V. THe gate is held at~20 V as well and pulled down to 0V every couple of millicseconds for 5 ms or so (charging pulse). The drain should show some fluctuating DC voltage slightly above the battery voltage with a multimeter. If this voltage is lower than the battery voltage during a charge cycle the mosfet is shot.
The design problem here may involve connection of a completely drained battery that draws more than the 2.5 A the mosfet is rated for. Thus, a completely drained battery should be precharged with some car battery charger or similar until it has at least 12 V before charging it inside the roomba.

Hope this helps and good luck
Al

PS please see my post below. Gordon found another mosfet on the opposite side of the board that may cause the problem

[mfortuna]

Good job!

There are multiple failure points. I never finished debugging a broken red I have but the mosfet gate was never active.

I since bought an external charger and use a rebuilt battery pack to power this roomba. It's sort of my on-call robot. Slap the battery in and send it on its way.

Mike

[Gordon]

Great sleuthing Al! I have a couple pix that can support your words; see attached pix. They both show the RH-end of the PWB. One shot is of the stuffed PWB (which has the 5V-Reg hiding some things you point out), and the other shot is of the same area stripped of large components.
I've inserted some part labels for the parts you point out.

...look at the right side of the circuit board. You will see two fat diodes {D30 & D32} at the right lower corner. Another fat diode {D31} is located some 5cm {3cm} to the left of those. Below that diode is the {U4} mosfet. It is a SMD part with 4 soldering points, part no. STN3PF06.......
Hope this helps and good luck
Al

That info ought to help numerous Disco owners!
BTW, did you notice that U4-S takes a via to the front side and connects to another N3PF06, U2-D? i'm wondering if U2 might also be susceptible to failure.
---Gordon
{Edited 070206: "D31" WAS D11}

[alepplew]

BTW, did you notice that U4-S takes a via to the front side and connects to another N3PF06, U2-D? i'm wondering if U2 might also be susceptible to failure.

Hi Gordon, thanks for the hint to U2 on the opposite side of the board. I missed that part. It is actually switching on the charger voltage inbetween the D32(coming from the charger jack) to the source of U4. Now an oscilloscope would be reqired to see which one is broken. If the source of U4 is keeping 20 V during a charge pulse the culprit is U4, if its source is dropping below battery voltage during a charge pulse the broken one should be U2. However, since these parts are around 2$ or so, I would recommend just to replace both.

Thanks for the pics!
Al

[alepplew]

I finally got a replacement for the charge controlling mosfet from www.rsonline.de at 0.70? a piece.
This is a Philips BSP250. I think any P channel mosfet that can do appr. 3 A and 25+ V will do. Package of the mosfet is called SOT223.
The roomba is charging perfectly, he is happy as am I following the replacement of the U4 mosfet

Good luck to all
Al

[Gordon]

It seems to me it would be appropriate to provide this thread with a link to the schematic (albeit a working draft version) which I announced on March 10th.

Interested parties may visit this thread to gain access to it.

{edit}While I'm here, I want to correct the link to Disco's diagnostic tests procedure that is shown in the top post; which won't work because a version bump impacted the filename:

WAS:"http://mysite.verizon.net/gsplews/Diagn ... .v.0.4.htm"

IS: http://mysite.verizon.net/gsplews/Diagn ... .v.0.5.htm

[Snake98]

I finally got a replacement for the charge controlling mosfet from www.rsonline.de at 0.70? a piece.
This is a Philips BSP250. I think any P channel mosfet that can do appr. 3 A and 25+ V will do. Package of the mosfet is called SOT223.
The roomba is charging perfectly, he is happy as am I following the replacement of the U4 mosfet

Good luck to all
Al

Was this the only part you repalced. I'm not a EE, but I'm handy with a soldering iron. My Discovery Se is doing the same thing, and wouldn't mind trying to fix it.

[baroom1]

I am interested too. It is a shame to see my Discovery SE wasted for just a bad charging circuit. A rapid charger cost $50+ off eBay and is way over priced.

[Gordon]

...Was this the only part you repalced. ...

Obviously, Al is not responding. I wanted to give him a few days to do so, before jumping in here.

His first post is fairly clear about which part (U4) he found duff, then replaced. Then, in a later post, after learning that U2 (an identical MOSFET, which is in series {have you looked at the charging-controller's schematic?} with U4, and mounted on the forward face of the main-PWB) could also halt charging (if failed), he suggested replacing it along with U4.

As I recall, U2 is more hassle to remove because its case is bonded (like 'epoxy-bonded') to the PCB. I would recommend such bonding for both U2 & U4, upon renewal. They need all the thermal dissipation enhancement one can conjure!

[bitbuster]

After I had the same problem as everybody else in this thread I just replaced U4. At first I was very happy, because my Roomba Scheduler finally started charging! But then I was a little bit confused when he stopped charging about 15 minutes later.
So I powered him up and he started cleaning as always but stopped about 5 minutes later with a blinking red Power light->battery empty.
Any idea what the problem could be? Do I maybe need to replace the U2 also to fix that?

[Gordon]

bitbuster:

I don't know. You did not provide enough background info, IMO. Here is a story I can make out of what you have given us:

1) The robot had its U4 part replaced, which then permitted battery charging to start.

2) The 'bot's battery, which had been lying about for some time, was in in uncertain state of charge, but probably low, since there was no way to re-charge it without Disco working 100%.

3) When the highly discharged battery was placed under charge, the controller opened the gates wide to allow the Fast Charger's full current to feed into the battery.

4) Both U4 and U2 felt the 'heat' from the high charging rate. In about 15-mins, U2 said to his new co-worker, "new-U4", man! I'm throwin' in the towel--I got racked pretty hard when the old U4 crapped out a while back, and I never felt right since. I can't take all this work! Bye. Charging halts. U2 had suffered a collapse due to a combination of a latent defect which developed as it and old U4 were being over-worked in the recent past, and the new, heavy load of charging an empty battery.

5) Next, the weak battery is called on to power the robot in cleaning-mode. The battery complains "cripes, I'm starving man, and they want me to do something!!". Roomba says "I hear you man! I see you're kinda low, I'll throw out the red flag for you. Red-flash, red-flash... etc.

The question remains: Is the battery charge depleted as Disco infers, or is it Disco that is unable to measure/assess battery voltage? I'm voting for the former.

Make sure you have an operational battery, then check cleaning performance.

What else can you do, but to assume U2 is duff? May as well replace it. You bought extra FETs, right? Once the PWB is out of its slot, the task of replacing U2 is a lot easier than it was for U4.

I think it would be prudent to do trial charging with the PWB out of its chassis-slot. Not too certain what it is you can do (w/o an O-scope), other than put a finger-tip on U2, maybe U4, to see how bloody hot they get. Just be certain battery voltage is not lower than 12Vdc when you begin the trial.

[bitbuster]

Gordon,
thanks for your quick reply.
I don't think that the battery is faulty, because my roomba stopped working just about one week ago. And before that the battery was working just fine. Also my Disco is just 9 months old.
But I can't verify if the battery has a problem because I don't have a second battery to test...
Yes, I bought extra FETs just in case...So I think I will try to replace the U2 today.Damn...I should have done that together with U4...It's really a pita to get the PWB out.
I will try that and post my results.

[finfin]

My Roomba (Marvin) is also not charging it's batteries It is a Scheduler that I bought a little over a year ago. The power supply is reading 22.63 VDC but the Roomba battery connections are showing 4.98 and 4.87 VDC with no battery connected. I haven't seen a lot here about voltages but I expected to see more like 7.5 or higher. any thoughts on how I should proceed? Are there any guides on trouble shooting this board? Are you commonly seeing replacement of U4 and U2? I've got a scope I could put on it If someone has so expertise on where to look and what I should see. I keep seeing the mention of a schematic but all I get at the sight is a txt document. Is there really a schematic somewhere?

[finfin]

Sorry!!! I see the link to the gif now... Good job on the schematic too!

[bitbuster]

Oh, I forgot to say: After I replaced U4 and still had problems I charged my scheduler about 5 times. Every time I charged him the battery lasts longer. So now he is back to his old shape! Right now I see no reason to replace U2. My roomba is working fine now.

[finfin]

So Bitbuster, does mine seem to have the same problem that yours had? What should the charging terminals put out with no battery connected?

[bitbuster]

I'm at the office right now so I cannot answer your questins regarding the voltage. But the problem you are describing is the same that i experienced.
So I suggest, that you replace the U4. I'm pretty sure that will solve your problem..

[finfin]

Did I understand that you can replace U2 and U4 with the same MOSFET?