[steveo]

this is a long story, but might be interesting to someone, especially a cheapskate who might want to get their knock-off opti working again... or just someone interested in electronics

i pulled apart my dead opti recently, which was a reman gm with a knockoff optical sensor in it, which obviously have a pretty high failure rate.

attaching the opti as a unit to a scope, and running it with a drill revealed intermittent signals, especially when tapping gently on the metal part of the opti.

in the future we might be stuck with this thing; as mitsubishi isn't making an opti sensor anymore.

as this seems to be the sensor used in all the recent GM remans, i examined it closely to find out what the deal is. i believe the sensor in the "$100 ebay special" is the same unit, but i haven't taken one of those apart (yet)

the majority of the surface mounted components on the sensor board are of reasonable quality. the soldering of the components on the upward-facing side of the board is all done by machine. the board itself and etching are good.

this doesn't suprise me -- chinese electronics factories are actually pretty good at board layout and etching, and they have tons of good robotic soldering rigs.

the board is marked with numbers that don't mean anything (probably internal production codes..)

there's a pair of LM2903 of unfamiliar brand, which i suspected as a failure point.. but i desoldered and tested them against a texas instruments LM2903 i had laying around, and they appeared to be a good enough clone.

the internal electronics are well sealed with silicone, and are likely to be fairly water resistant, meaning any water intrustion problems even on a cheap opti are probably with the slotted wheel, not the electrical part. all the same, with the design of the casing, water instrusion is possible.

the control board is hand-soldered to the 4-pin opti connector. the work is sloppy but strong. the sealing of the connector pins through the case is perfect.

the big nasty kill-your-opti fault i found was in the stage of assembly where the optical pickup (looks to be a standard photo-diode) are hand-soldered to the board.

of the four connections made between the two sensors, none of the solder wicked to the photodiode at all. gently vibrating the photodiode while monitoring a scope, as it was originally mounted to the board, resulted in lost conductivity. i'm certain this caused the intermittent failure i was experiencing

the photodiodes were attached so poorly, they could be removed from the board by simply pulling them.

interesting part is, the photodiodes were siliconed into the casing, then the board was attached. there is quite a bit of silicone in between the photodiode and the board.. in fact, it's everywhere.

i can only draw the conclusion that the silicone contaminates the contacts, preventing the solder from adhering

i cleaned the contacts and reflowed the solder, and the opti seems to be behaving itself in a simple 'scope test driven by a hand drill

 

[ShowNoMercy]

So I guess QA did not catch the over zealous attempt to ruggedize the diodes I guess, interesting. Or, they were somehow attached and ruggedized in one step?

Any pics pre fix?

 

[steveo]

i do have some pics, they're at work, i'll try to remember to post them

i'm very sure it's silicone contamination due to the order in which they're assembled, and they wouldn't be a bad sensor otherwise. if they'd just leave the diode side alone and just pour silicone in from the top after assembly....

 

[cocobolo95]

This is interesting, it's been a long time since I took an opti apart. From memory, I remember the whole optical unit being encased in plastic. Is this correct? If so, how did you remove the plastic case, how do you reseal the optical board, are all the components on the board available as spare parts.

I ask because, if components are available, circuit board etching is not a process that can't be done by someone with some electronics knowledge.

With the ability to make the circuit board and encase it, and obtain a new bearing for the opti, there's no reason why an enterprising person couldn't rebuild the opti.

I imagine even the optical wheels, if replacement is needed, could be made by a good machine shop.

 

[steveo]

they went with the approach of pouring silicone sealer over everything inside the sensor, but leaving a cover you can pry off easily. after working on it, you could just pour silicone in it and snap the cover back in (maybe with a few dots of epoxy to hold it)

im pretty sure the mitsubishi sensors don't come apart.

yes, everything on the board is a common off-the-shelf component, and can be replaced.. that is, except for one 'mystery IC', but i could find out what it is pretty easily.

the optical wheel could not only be replaced, but improved. it could be made out of stainless and thickened to almost 0.75mm and still clear the sensor. right now it's flimsy and prone to corrosion

 

[bobdec]

Got towed home 3 years ago when mine stranded me (DTC 16 low res) , it started back up and ran next day. Scoped high resolution and saw 4 volt saw tooth instead of 5 v square wave on high res signal, when running. Opened it up and the wheel was spot welded onto the drive. My old wheel was rusted across about 30% of the surface outward to the slots from water intrusion and HV ozone. So I never dug into the real problem and replaced the unit with a Delco. And like a fool threw out the old one. Regretted that many times as it would have been beneficial to analyze the failure. I see that LM2903 IC's are only about $1.00 each in packs of 5-6. Interesting if they are accessible to be replaced (some circuit pics would be great) . May be of value to reverse engineer the internals of the opti and come up with a schematic and parts kit of components. It would seem the electrical pieces would be much lower cost than the actual casing and drive components. Putting new electronics into a solid GM casing/drive as a rebuild would be interesting. Still would need a new rotor and the mechanical alignment to fractions of a degree on the slotted wheel, that's a challenge.

 

[steveo]

everything is surface mounted except for the photodiodes. there are no pins through the board. that makes everything really easy to replace

here's some shitty pics

2013-02-26 08.24.44 | Flickr - Photo Sharing!

2013-02-26 08.25.06 | Flickr - Photo Sharing!

2013-02-25 08.46.08 | Flickr - Photo Sharing!

 

[bobdec]

Looks pretty straight forward, except for the epoxy on the board. Think I try to scrap up a couple of old defective opti's and play with them. How is the wheel attached to the drive shaft ??

 

[kwacasey]

I see alot of people on here knocking the chinese opti's, sounds to me like the real issue might just be poor quality assembly. This is really not surprising to me at all, however might be a good thing if we can identify the issues and log them for others to reference. I think this is a great post Steveo.

I might pick a defective one up sometime soon and see if there is anything i can do with it.

 

[steveo]

yeah i want a complete chinese opti to dissect too, so i can figure out the other issues. this was a reman gm with a chinese sensor.

 

[kwacasey]

Most if not all of these electrical components come from asia, china happens to be a very large country with alot of manufacturers of electrical components. I have alot of experience with parts from this corner of the world, some are good others are not. If we can identify the weak ones and find suitable replacements I am sure we can turn some of these lemons into lemonade.

I hear that the OEM part has now been listed as EOL (End of life). Now would be a good time to identify good alternatives for that part.

 

[BLACK HAWK]

simple'ish solution, upgrade to the EFI connections x24 kit. no more opti and the tune ability of an LS[enter number here]

 

[steveo]

yeah instead of figuring out how/why things work, lets just change to a different system entirely

'cause that's more fun

:craz28:

 

[cocobolo95]

Even the efi connections 24x kit has electronic parts that can and will fail in time. And with planned quick obsolescence being the in thing, sourcing spare parts for that will be a pain in not too many more years.

 

[steveo]

...I have alot of experience with parts from this corner of the world, some are good others are not. If we can identify the weak ones and find suitable replacements I am sure we can turn some of these lemons into lemonade.

i have a lot of experience with chinese mfr's too

the funny thing about chinese manufacturers is that if someone is paying them to do it, they can get really good at designing stuff.

they aren't idiots. they're just better at price scaling than we are..

if you say "make me a $0.50 opti sensor, times 10,000", they will gladly do it

....with the appropriately skilled employees, and the appropriate corners cut to make it that price.

if you say "make me a $25 opti sensor that's close in quality to this mitsubishi one", they'll gladly do that too.

the trick is, the first lot off the line is going to be half-developed and have issues. you'll have failure rates through the roof. they'll try to cut corners at first, no matter what, to see if they can get away with it.

you get back to them, tell them the issues, say they need to be resolved in the next batch to continue your business relationship. usually they are

that's not unique to chinese mfr's.. most stuff you have built in the USA or mexico is the same way.

 

[BLACK HAWK]

yeah instead of figuring out how/why things work, lets just change to a different system entirely

'cause that's more fun

:craz28:

No comment comment :lol: .. this is one Chinese LOVING thread though! :thumbsup:

 

[ShowNoMercy]

Even the efi connections 24x kit has electronic parts that can and will fail in time. And with planned quick obsolescence being the in thing, sourcing spare parts for that will be a pain in not too many more years.

Only sensor is a mass used crank position sensor and in my case, a mass produced vortec cam sensor.

Not defending the system, but its def ahead of the curve in terms of finding replacement stuff :lol:

 

[kwacasey]

My real take away here is that poor craftsmanship was the likely cause of failure for your system, with that being said we might be able to fix and move on. I was under the impression that the weak point was the optical sensor itself, that is the only magic part of this system, everything else in the module just supports that sensor setup. If the optical sensors are decent quality (not the cause of most failures) we can identify the weak points and maybe come up with our own solutions that will help us save money and have some amount of faith in these "lesser quality" opti's.

 

[BLACK HAWK]

I'm sitting here wondering what a chinese opti is, is there a part number for it or a link to one?

 

[cocobolo95]

Chinese optis are the ones sold on ebay and other such places. Chandler motor sports opti is one example.