A thread in which we try to banish the carb.

The Something Awful Forums > Discussion > Automotive Insanity > Cycle Asylum > A thread in which we try to banish the carb.

Woolwich Bagnet: Apr 27, 2003

Yeah a wire could be used to bypass the trace pretty easily. As for the soldering, if you're not comfortable doing it yourself, an electronics repair place could do it easily I'd imagine. Those pins are plenty large to solder by hand.

Edit: There really shouldn't be enough current though going to serial GND to cause that kind of damage.. I'd be interested to see what the root cause is.

Edit 2: Check to see if pins 15 and 16 are shorted (top right two pins from the second photo). That cap off to the right of it I'm guessing is between the two. If those got shorted I could see too much current going through it.

Woolwich Bagnet fucked around with this message at 05:07 on Oct 16, 2018

# ¿ Oct 16, 2018 04:59

Adbot: ADBOT LOVES YOU

# ¿ May 11, 2024 22:09

Woolwich Bagnet: Apr 27, 2003

Slavvy posted:

Solder by hand with what? The tip of my (totally normal) soldering iron is like 4x the size of one of those legs but if the tools aren't too expensive I might try it. How do you do all sixteen at once? Or alternatively, how do you do them one at a time without melting the surrounding area? Literally never attempted anything like this before and am reluctant to gamble $400 on my shaky, clumsy-rear end hands getting it right on my first time ever.

I don't understand which pins you're referring to. 15 and 16 are the boot wire and something called accelLED, they aren't shorted but also aren't in the second picture. Or do you mean the pins on the chip itself? In which case my multimeter probes are too big to touch them individually.

Don't know what cap you're referring to, I can't see anything wrong in the area anyway.

Theories as to why it failed:

My laptop has a horrible chinese battery that could've surged while logging.

Serial cable may have gotten unplugged on the last run while logging and the earth ring got shorted on a part of the bike.

I've struggled with plugs getting petrol soaked and shorting, maybe voltage spikes from this somehow causing it to go wrong? This hasn't been the case in a good while though.

Yeah if you haven't done it before I wouldn't risk it. You can do those by hand, usually using flux and soldering 2 pins first, then applying solder. You'd need a hot air gun to remove the chip in the first place which I'm guessing you don't have.

Pins 15 and 16 on that chip (the max232) are Vcc and GND. You count starting where that plastic circle is as the top, pin 1 to the left of it. The top two pins on the right are 15 and 16.

It's possible that the chip itself isn't damaged at all if it was from a short. That tan rectangular component to the right of it (C31 I think based on the PCB) is a filtering cap that goes between Vcc (positive supply voltage on the board) and GND. Try putting the probes on it, one on each metallic side and see if there's continuity. The powersupply of the board itself is likely the only thing that could supply enough current to cause that damage.

# ¿ Oct 16, 2018 05:55

Woolwich Bagnet: Apr 27, 2003

Slavvy posted:

Thanks for the tips minnesotta, don't know where I'd be without this dead gay forum and it's pack of incredibly clever people. I really appreciate it!

I've tried to test continuity between the two legs of the chip and I'm pretty sure they're shorted. I can't say with 100% accuracy because I'm using actual pins to try to get the contact, so I can't be sure that I'm not accidentally creating a short because my eyes and hands both struggle at scales that small.

Testing the cap was much more straightforward and it looks like there's no continuity across that.

Can the chip short internally somehow out of the blue, is that a thing that happens? Somewhat concerned thinking about the current that chip likely deals with vs what it would have taken to burn up the trace like that, I honestly expected the cap to be the short which would I guess explain everything.

I've had a good check over the wiring harness and there's definitely no shorts there, the three serial wires peel off by themselves and are never part of the larger trunk.

I can't find any issues with the power supply for the ecu. I'll double check properly when I've got a chance but it looks like it's been wired so the MS is powered from the factory main relay that also powers up the coils, lights etc (separate circuits feeding off the one big lug) and from what I've read it needs a completely independent power supply so that's on the to-do list. OTOH the rest of the install is extremely tidy and professional, I want to assume the person who built it knew what they were doing but who knows. All efi-related earths are large and healthy and separately bolted to an earth plate which itself bolts direct to the battery -ve.

I'm running a quadspark so there's no coil voltage going direct to the MS, but I've already had one quadspark die prematurely from the aformentioned spark plug problem. There may be a deeper issue underlying this but I'm hosed if I can find it, it may or may not be me chasing ghosts. I've always felt that the traces in the data logs are 'noisy' and often have random spikes but I lack the experience to know if it's a real problem or if MS always looks like that when you've got a crappy tune.

Overall I can't think of anything I've done that could've caused this so I'm searching for causes that could be long term cumulative as there doesn't appear to be anything wrong with the bike itself. Can my lovely laptop cause something like this? Have a hard time picturing a USB port carrying the juice to cause that kind of damage without the laptop itself taking a poo poo.

As for replacing the chip. I've got a guy that fools around with consumer electronics, mainly phones and consoles and other popular stuff and see if this is something he can tackle. Searching more closely indicates the chip itself is a MAX3221I, does that mean the operating principle is different or is it just a size/shape/specs thing?

Assuming that cap is between pins 15 and 16 then there should be no short in the chip if there's no short there since it would still show continuity. I would confirm that that cap is connected to those pins. I think it's fairly unlikely that the chip would short like that out of the blue.

An easy way to tell if pins that close together are shorted is to place the probe between pins 14 and 15, and then press down with the other probe on the pcb and slowly move it closer to the top pin (16) until you feel it. It's easier to feel your way around things that small then to look at them while you're trying to do it.

It should be replaced with the I model since that one is rated for a wider temperature range compared to the other.

# ¿ Oct 16, 2018 21:27

Woolwich Bagnet: Apr 27, 2003

Slavvy posted:

Using the above method I can confirm there is definitely continuity between the two pins. There is also continuity between the inner end of the cap (as in the end furthest from the edge of the board) and pin 16 on the chip so I'm guessing it does do what you think it does.

Is the cap meant to have continuity normally? I get about 4.3ko across it in either direction and that seems to be the same as the other identical looking ones elsewhere on the board.

Pin 30 doesn't go anywhere, the wire is a five inch stub that's heat shrinked and insulated on the end. No continuity between it and the burned trace either. The burned trace has continuity to connector pins 18, 20, 22 and 23. All of which are marked as various flavors of GND so I'm guessing that's good and normal?

Capacitors block DC but allow AC to flow through freely (past a certain frequency). Most multimeters will show some value of resistance on them, it's fine. It's hard to say now if that chip is good or not. You could put a bodge wire to bypass that trace and see if it works normally I guess. If you want to be super safe about it you could use a resistor as part of that with a low value, 100 ohm or so to limit the current to try to prevent any future damage. It's really hard to say though without knowing the root cause, which could be difficult to determine.

Maybe try measuring the resistance between the + pin for the PSU that supplies that chip and ground to see if it's really low.

# ¿ Oct 17, 2018 01:05

Woolwich Bagnet: Apr 27, 2003

Slavvy posted:

Ok, how do I find and identify this stuff? Also see my edit above, the pins don't do what I thought they did and now there doesn't appear to be a short after all.

Looks like it uses a 5V voltage regulator LM2937. Based on the pictures you posted I believe it is this:

And this is the pinout for that:

Input should be from the battery, so 12 V. Output should be 5V when referenced against pin 2 (GND). If you have it connected to the battery be extremely careful that you do not accidentally short any of these pins with each other.

Output to pin 16 on that serial chip should show almost no resistance if it is the correct regulator.

You can try checking the resistance between pins 2 and 3 with the battery not connected to it. Also try measuring between pin 2 and the pin that broken trace goes to (pin 19 - SerialGND). It should be near 0 ohms.

# ¿ Oct 17, 2018 01:58

Woolwich Bagnet: Apr 27, 2003

I agree that edit 2 seems most likely... maybe when a serial cable was connected the ground on the rs232 chip was at a slightly lower potential and caused the current to flow to it, because the main ground was poorly connected or something? I say put a wire across it and see if it works then. I doubt you could cause any more damage than what may have already taken place. Oh and it's not voltage that would cause the damage, it's power (wattage). Power causes joule heating and the resistance of the trace is going to based on it's width/thickness. This is why thicker traces are used for paths that expect more power to flow through them. You can look around online to find the wattage traces can handle based on width/thickness.

Woolwich Bagnet fucked around with this message at 06:00 on Oct 18, 2018

# ¿ Oct 18, 2018 05:57

Woolwich Bagnet: Apr 27, 2003

Slavvy posted:

I've just remembered something that may incriminate me:

Shortly after I'd gotten the bike, I replaced the battery. Not paying attention like a dumbass, I connected all the earths except for the main thick gauge one and attempted to start the bike. I realised my error very quickly, but I imagine the starter trying to earth through the ms could've done the damage in a fraction of a second. I don't know why it would be the serial earth and not the main ecu earth though.

Righto, the search for a tiny soldering iron with a(n optionally) tiny operator attached begins.

You don't need a tiny soldering iron. Just run a wire between these two points.

# ¿ Oct 18, 2018 20:39

Adbot: ADBOT LOVES YOU

# ¿ May 11, 2024 22:09

Woolwich Bagnet: Apr 27, 2003

Glad to hear it's fixed!

# ¿ Oct 19, 2018 14:48

The Something Awful Forums > Discussion > Automotive Insanity > Cycle Asylum > A thread in which we try to banish the carb.