Hardware diagnostics

pleriche

6

Hardware diagnostics

« on: 30 / August / 2014, 07:44:17 »

Advertisements

I'm trying to fix an SX200 IS which initially appeared to have a stuck lens (common problem with ingress of dirt), but having completely disassembled and reassembled it (see https://www.ifixit.com/Guide/Disassembling+Canon+PowerShot+SX200+IS/28471), I'm convinced that isn't the problem. In fact, with the lens assembly on the bench, I could extend, zoom and retract it with a 3v battery applied to the zoom motor.

Either something has slipped out of place inside the lens and I haven't managed to put it right, or whatever detects the lens zoom state is faulty. I'm having trouble identifying any such sensor.

It would be interesting to run a script to report the lens extension and zoom state as seen by the camera, but I don't immediately see anything in the command reference that'll do that for me. Am I missing something?

Regards - Philip

Logged

ahull

634

Re: Hardware diagnostics

« Reply #1 on: 30 / August / 2014, 09:22:03 »

It sounds more likely that one of the optical interrupters is either out of place, or has some obstruction, or the FPC (ribbon) cable is damaged.

If that is not the case, then perhaps the driver circuit for the motor is damaged (blown transistor), or there is a current limit fuse somewhere on the board which has "popped" as a result of the lens being jammed.

These cameras are tricky to fix when you start digging down to this level, partly because the components are physically very small, and partly because there are no circuit diagrams.

I have changed fuses on a couple of Ixus 50/60 vintage cameras, (in those cases the fuse is on the FPC cable that runs to the power switch assembly, and when it blows it leaves the camera completely dead), but I have never had any dealings with an SX200.

Logged

srsa_4c

4451

Re: Hardware diagnostics

« Reply #2 on: 30 / August / 2014, 11:02:23 »

You did not specify how the camera and its lens behaves after a starting attempt.

In addition to what ahull wrote above: you may have a problem with the focus lens. It can be stuck for some reason, for example (sorry for the poor English)
- dislocated lens guiding pin
- lens overran its end position

If you are able to start the cam in playback mode (and use CHDK), you can get the camera's error list to a file:
Create a plain text file with the following lines:
call_event_proc("UI.Create") call_event_proc("UIFS_WriteFirmInfoToFile",0)
Name it firminfo.lua, copy it to the CHDK/SCRIPTS directory on card. On camera, go into the CHDK menu: Miscellaneous stuff -> Enable Lua native calls [.]
After executing firminfo.lua, you'll find a firminfo.txt on the card. The error messages might give some hints.

Logged

pleriche

6

Re: Hardware diagnostics

« Reply #3 on: 30 / August / 2014, 15:15:44 »

Thank you srsa_4c - I'll definitely try getting the error list though I'm not sure yet whether I can get the camera to boot and execute a script before throwing the lens error.

Unfortunately it takes quite a while to completely dismantle the camera in order to retract the lens using a battery on the zoom motor, and I've only done that the once. If my powers of observation are reasonably accurate, on powering on it extended the lens to the normal position and then seemed to try to inch it further forward before giving the lens error. Powering off and on again simply caused it to try inching further forward on each power cycle until it was fully zoomed. The camera came to me in the fully zoomed state.

Next time I dismantle it and retract the lens I'll try switching it on in playback mode, hopefully running the script.

My best guess is that there's a sensor which detects the lens home position (a connection behind what I take to be the macro lens just in front of the image sensor), and that it navigates the zoom in and out by dead reckoning, counting revolutions of the zoom motor (with a couple of connections at the spindle end of the motor). Either the home position sensor or the spindle sensor is maybe playing up. Both the motor and the electronics driving it seem to be ok.

Frankly, I'm less interested in fixing it (though I'd be dead chuffed if I did) than in understanding the problem. Having got so far, it's bugging me!

Regards - Philip

Logged

ahull

634

Re: Hardware diagnostics

« Reply #4 on: 30 / August / 2014, 16:24:55 »

Well if you are not too bothered about fixing it, you might consider purchasing a replacement...

Try the usual suspects, for example... http://www.ebay.co.uk/itm/CAMERA-LENS-UNIT-REPAIR-PART-FOR-CANON-SX200-SX-200-IS-CCD-Silver-/121406310726?pt=Digital_Camera_Accessories&hash=item1c44614946

.. however if you want to know how it works then this thread will give you a lot of detailed information...

http://chdk.setepontos.com/index.php?topic=8801.0

Over the last few years I have fixed quite a few of these little point and shoot cameras, for fun, rather than for $$$, and I can say that in the main, most of the Canon lens mechanisms operate in a fairly similar manner, so what applies to the S90 will most likely apply to the SX200.

Logged

pleriche

6

Re: Hardware diagnostics

« Reply #5 on: 30 / August / 2014, 17:36:29 »

Thank you! I had considered one of those replacement lenses from eBay but put that idea on the back burner at least until I finally dismiss this lens as unrepairable. And interesting the thread in your 2nd link indicates that some such replacement lenses are actually factory rejects. I might do better to wait for a non-functional SX200 with a different fault to come up on eBay, and then I should be able to get one working camera out of the bits.

Wow! That thread is impressive! My lens is substantially different, but as you say, the principles are very likely to be the same. In particular, my guess that there is a stow switch to indicate when the lens is retracted seems to be correct, but I'll have to read it much more carefully next time I have my lens out.

One particular point you can probably help me with: it says some opto switches are held in with rubber cement. I seem to have several of those such as in the attached image. I wasn't sure whether you could remove them or if so how, with the result that I've been putting more strain than I like on the ribbons, so farwithout obvious damage. Presumably you just scrape the cement off. But what sort of glue do you use to refit them?

Regards - Philip

Img_1529a.jpg (27.63 kB, 486x792 - viewed 33 times.)

Logged

ahull

634

Re: Hardware diagnostics

« Reply #6 on: 30 / August / 2014, 18:02:38 »

Quote from: pleriche on 30 / August / 2014, 17:36:29

Thank you! I had considered one of those replacement lenses from eBay but put that idea on the back burner at least until I finally dismiss this lens as unrepairable. And interesting the thread in your 2nd link indicates that some such replacement lenses are actually factory rejects. I might do better to wait for a non-functional SX200 with a different fault to come up on eBay, and then I should be able to get one working camera out of the bits.

Sadly lens damage is the most likely fault (with LCD damage being the second most likely one), and therefore you may have to wait a while for another SX200. Having said that, you may well pick one up with a damaged LCD for not very much, so long as you are prepared to wait. I set myself a goal of never paying more than £4.04 (+P&P) for any camera, and I have collected a substantial collection over the years.

The lenses you see on ebay may well in fact be "recovered" from e-scrap, I've bought a few over the years, and only had one dud (it had dirt in it).

Quote

Wow! That thread is impressive! My lens is substantially different, but as you say, the principles are very likely to be the same. In particular, my guess that there is a stow switch to indicate when the lens is retracted seems to be correct, but I'll have to read it much more carefully next time I have my lens out.

The optics and sensors vary, but the basic mechanics and position sensing electronics are typically very similar. The pinouts of the ribbon cables are also generally very different in their layout, since they are designed for easiest assembly of each particular model.

Quote

One particular point you can probably help me with: it says some opto switches are held in with rubber cement. I seem to have several of those such as in the attached image. I wasn't sure whether you could remove them or if so how, with the result that I've been putting more strain than I like on the ribbons, so farwithout obvious damage. Presumably you just scrape the cement off. But what sort of glue do you use to refit them?

Regards - Philip

Fixing the lenses is a bit of an art. I can't give much specific advise, apart from the fact that cyanoacrylate (superglue) is not a particularly good idea as it tends to fog optics. Bathroom silicone sealant applied very carefully with a syringe or needle makes a good substitute for the rubber glue, ( mix it with a little talc or cornflour if you want it to set quickly). Bathroom sealant however produces acetic acid, so use it sparingly, and try to avoid getting it on metal parts. The plastic of the FPC ribbon cables wont care, but copper contacts probably will.

Removing the various glues takes patience, and if you have a good magnifier or access to a microscope that will also help. The magnification doesn't need to be particularly powerful, x 4, x10 or x 15 will be more than adequate for most tasks. Also look out for plastic anchor pegs that go through small holes in the FPC cables, you will need to lift the cable off these pegs before attempting to extract the sensors.

One other trick that Canon use in some models is to solder sensors and sections of the FPC cable at different parts of the assembly process, to allow for more complex folding and tighter cable runs. This means that some lenses can only be disassembled by un-soldering certain contacts. Soldering these cables requires a very fine tip on your soldering iron, and careful handling.

As far as removing the glues goes, I tend to use a very small flat bladed screwdriver, or a scalpel blade, and light pressure, you don't want to cut through the tracks on the cable.

One final point, I spent a lot of time trying to figure out why one of my lenses didn't work, only to discover that there was dirt in one of the connectors on the main PCB, and nothing actually wrong with the lens at all, so check that your connectors are clean and that the cables are straight, and correctly seated in the connectors.

« Last Edit: 30 / August / 2014, 18:07:53 by ahull »

Logged

pleriche

6

Re: Hardware diagnostics

« Reply #7 on: 03 / September / 2014, 08:01:18 »

Thanks srsa_4c - I eventually managed to get the firminfo script to run. I had to set the Autorun attribute, and was confused for a bit until I found the report file in the root of the SDCard. On my own camera (same model, fully functional) it reported several historic lens errors (probably from when I pulled the battery with it powered on to see how it was meant to recover) but on the faulty one (my cousin's) it reported FocusLensError - so your guess was right!

I take it the focus lens is immediately in front of the image sensor - I haven't tried removing that. There seems to be a stepper motor, a spring, a very fine leadscrew and a guide pin parallel to the leadscrew - how they work will hopefully become evident when I get in there. Why a fault here should prevent the lens from retracting is slightly puzzling, but I suspect the lens-stowed sensor is part of the focusing mechanism and is continually reporting True. On the other hand, it could be the GPIO pin on the processor chip stuck high (or low).

Your tips, ahull, on removing the glue are very relevant - thanks. I have a pair of 3.5 dioptre readyspecs which I've found essential for this sort of thing ever since the autofocus on my eyeballs got stuck on infinity some years ago, plus a jeweller's eye loupe (around 20 dioptre) which gets quite a lot of use.

I think I'll have to apply your tips in order to free the back plate from the main lens shell so as to facilitate dismantling the focussing lens. As for refixing them, silicone sealant takes a while to go off, and every tube I've ever had has set solid a few months after opening - very annoying if all you need is half a ml. I'm inclined try hot melt glue. A glue gun may well deliver far too much, so I might try spooning a little from the glue gun tip with a matchstick, or maybe heating a bit of glue in a metal lid over the soldering iron.

Regards - Philip

Logged

ahull

634

Re: Hardware diagnostics

« Reply #8 on: 03 / September / 2014, 10:27:10 »

One other thing you might consider as a "glue" is nail polish.

This works well for most plastics (except perhaps polystyrene, which dissolves in the acetone generally used in nail polish, so do a test first), and is very simple to apply. It doesn't fume the way cynoacrylate tends to, so is unlikely to fog lenses.

I have used this as a "screw lock" to stop small screws vibrating loose, but it should work well in most situations where you may need to re-fix more than once, and where high bond strength is not required.

The choice of colour I will leave entirely up to you.

One other tip, if you have a can of air-duster or a compressed air source, *gently* blow in to each of the sensors, to see if perhaps some foreign body has lodged in there. Don't be tempted to go in guns blazing, as some of the components wont survive 40psi.

Lung power might also work, but is inclined to leave some organic residue (spit and bacteria) which might encourage mould in the future.

« Last Edit: 03 / September / 2014, 10:34:55 by ahull »

Logged

pleriche

6

Re: Hardware diagnostics

« Reply #9 on: 04 / September / 2014, 13:31:46 »

Well, it seems I've found the problem. The focus lens has a threaded brass insert which engages with the focusing leadscrew, and the thread in the brass is worn out, with the result that the focus motor is unable to move the focus lens against the retaining spring. So unless I could find a friendly amateur watch maker to make me a new brass thingy or a camera with a different fatal error to cannibalise, this would seem to be the end of the road. And I can put off the vexing question of which colour nail varnish to get

Thank you for suggestions, one and all. I've leant a lot.

Regards - Philip

Logged