Wednesday, July 25, 2012

Broken Logitech G25 Part II

Finally got the parts in the mail, I left the G25 disassembled the whole week and it was lying on my room floor, starting to get a little obnoxious. Here’s a shot of the broken FET.

IMG_5079

IMG_1163

I ordered 2x both N and P channel dual FETs, although only the IRF7103PBF was actually shorted on mine.

IMG_1165

IMG_1167

Proceeded to solder the replacement in.

IMG_1169

IMG_1168

I just held it on top of the pads and tacked on corner on. Then adjusted the whole IC until everything lined up and completed soldering it on. Not the cleanest job, because some of the pads on the board have no mask, so it looks like a couple of pins are shorting, but electrically it’s alright.

I then proceeded to re-assemble the motor and linear gear together.

IMG_1180

I left the wheel installation out till I could see if the wheel would fire up and calibrate. During re-assembly I just put the linear rack gear somewhere in the middle and ta-da, it worked during power up and could calibrate properly.

There was a little bit of arcing and smell from the cleaned up motor though, I suspect the brushes haven’t bedded in properly yet. I then plugged and unplugged the wheel a couple of times to get it to cycle the calibration process more and that fixed itself. There was also some cogging present initially, that’s also fixed.

Proceeded to finally play some GT5. Open-mouthed smile

Tuesday, July 17, 2012

Broken Logitech G25 Part I

Wanted to play a little GT5 over the previous weekend. Hooked everything up and turned the PS3 on.

The G25 normally does a little calibration routine at the start, but this time it did a little quarter turn and sat there motionless. Power LED on the shifter just kept blinking, I thought WTH and decided to try rebooting the PS3. Nope, this time it wouldn’t even move and just kept blinking.

Plugged it into my PC, and everything works, except the wheel. I’m guessing without the calibration routing, the wheel axis will not work.

Measured the power coming out of the adapter and it was a stable 24V, as per the spec on it’s label.

Next I tore into the wheel itself to do a little poking around.

The 1st thing I suspected was the motor driving circuitry. Disconnecting both motors and probing around the vicinity confirmed that I was right, one of the dual MOSFETs in the IRF7103 was a dead short.

Proceeded to remove the FET and see if it’s really dead, or something else on the board had died.

IMG_5072

Yes it was dead. Removed and re-probing with my trusty DMM confirmed that. I put both the FETs on order (IRF7342 and IRF7103) thru RS-Components (actually only the IRF7103 had fried, and it was a single FET in the dual FET package).

I then decided to find out what exactly killed the FETs, wouldn’t want my replacement to go up in smoke the second I powered it on, why did the wheel turn 1/4 way before coming to a stop?

Removed a couple of screws and disconnected a bunch of stuff to get the gearbox and motors out where they could be comfortably moved around. I then decided to probe the resistance of both motors.

IMG_5075

IMG_5076

Here you can see that one motor has an encoder (where I guess the wheel actually gets it’s axis signal).

A pic of the reading I got from the motor with the encoder on the back. This is the min resistance (I spun the motor slowly by hand and used the DMM’s min function).

IMG_5081

One of the ‘other side’

IMG_5082

I pulled the motor leads out of it’s connector and tried using the power adapter itself to power the motor with the encoder. Bad idea, a whole bunch of smoke came out of the commutator area and the motor just jerked a little and sat there pretty, whilst more smoke poured out. I quickly disconnected the wires.

IMG_5083

Can’t really see the commutator in this pic, I then disassembled the whole motor so I could see what was going on inside, initially I suspected that maybe a winding was somehow shorted or the brushes were somehow shorting together.

IMG_5085

Here’s what I saw after prying the back out. Look at the gaps in between the comm segments. A mixture of oil,grease and carbon has built up inside and I’m guessing is the cause of all the problems I’m having.

I dug through my drawer and found a little pin that was small enough to fit into the gaps and used that to clean it out. Together with some contact cleaner.

IMG_5086

This is the amount of crap that flowed out.

IMG_5087

All that black stuff came from the commutator. After cleaning I measured the resistance again.

IMG_5088

Not too bad I’d say, pretty close to the other motor. For that motor I just sprayed a little cleaner into the comm through the side gap and called it good.

Piecing the thing back together. I had gouged the end ‘tabs’ a little during the disassembly, and ended up just using a small punch to wack 4 tabs back on afterwards, seems to be tight enough to me.

IMG_5089

I also noticed a little crack on the encoder wheel and used a little bit of CA glue to stick her back together. The wheel just slides onto the shaft and is held there with friction.

IMG_5090

Look forward to Part II where I replace the fried MOSFET and try powering her up! I have the parts on order right now and am waiting for them to arrive.

Sunday, July 8, 2012

Gigantic Hiberfil.sys in Windows 7

You’re reading this coz you have a tiny (40GB) SSD like me for the OS drive and there’s a gigantic (mine was up to 12GB before I had it removed) hiberfil.sys file and you want it gone to free up some much needed space? Fear not, this post is for you.

As far as I can tell, hiberfil.sys let’s your PC Hibernate, Windows will use hiberfil.sys to store it’s current state, and since windows is managing it, you can’t delete the file. Until you do the below steps of course.

Open up an administrator mode command prompt, and put this in.

powercfg -h off

You should immediately notice the Hibernate option gone from the shutdown menu, as well as, TADA, a bunch of free space on your C: drive (the file hiberfil.sys should also be gone from the root directory).

Note: I forgot to take screen captures of the process, sp yes you’d have to put up with a wordy post. Sorry.