Canon 400D IR mod

[bojan]

Right now I am in process of IR filter removal, everything is disassembled down to sensor.

Question: do I have to put back the filter assembly together with ultrasonic shaker at all ?
I will never use it as terrestrial camera.
It seems that the only purpose of this frame from now on could be to keep the dust away from sensor.

[glend]

No mate, you don't need to retain the piezo 'shaker'. Removing it does not generate any camera error 99s. I have a pile of them from my 450d mods. However you should pay some attention to camera setup functions and turn off "sensor cleaning" functions, just to be sure you don't get any errors if it tries to use it. Some models can be set to clean on power up, so you want to stop that.

[bojan]

Thanks for this Glen..
Camera is being put together :)
BTW, error99 was probably caused by short circuit - one of the pins (VCC? it was one of longer ones, in the middle of connector) from CF connector was pushed out of plastic frame and was wandering around inside the camera for some time.. It fell on the table when I removed the main board. I managed to repair this (I have stereo microscope and steady hand).
Now I hope that nothing was fried inside (i have a feeling the power supply circuit is below this shield.
Interestingly, there was no problem with memory card... at least not when I was playing with it.

[bojan]

:mad2: Brick...

[glend]

What does that mean? I believe the Digic processor could be under that shield, and they can usually be popped off. I am not really familiar with the 400 internal but that is the location of the processor board on the 450d.
Does it power up? I have often found one single connector can be the problem.

[bojan]

It doesn't power up any more.
I disassembled and re-assembled the whole thing and the result is the same.
Glen, under that shield is power supply unit (switch mode), Digic processor is above it.

Started to look into circuitry, batt voltage is there, I found the location of the FET switch which I earlier suspected to be faulty (based on symptoms).
The small part marked as "P" might be the fuse.. there is no voltage on other side, towards the main board.
According to this (http://www.aliexpress.com/item/50-Pcs-Per-Lot-Littelfuse-SMD-0603-Fast-Acting-Fuse-1-5A-32V-046701-5-Marking/32309803133.html) website, it is 3mA.. not quite right! Or 3A? but I am not sure if this is the standard marking for fuses...


It will take a while before I start imaging in Halpha and Mono, it seems..

[luka]

The part marked P is a fuse. However, you will be struggling to find the current rating without knowing the manufacturer. See for example this (http://www.digikey.com.au/techxchange/message/8667) post.

[rcheshire]

https://www.dropbox.com/s/z42vyepqu3eu12z/canon-service-manual-eos-kiss-digital-x-eos

If you don't already have it. Parts catalogue toward back of document. Fingers crossed.

[bojan]

Rowland,
Thank you !
Yes, I did have it already, but still thanks heaps.
Actually, I should have posted the direct link I found by googling earlier...
http://thydzik.com/downloads/canon-service-manual-eos-kiss-digital-x-eos-digital-rebel-xti-eos-400d-digital.pdf

However, this service manual is not designed for circuit-level repair, components are very hard to cross-identify and part numbers are Canon internal with no reference to source...
So I will have to try to figure out the apropriate fuse value by replacing it with jumper and then measuring current (DC and peak).

BTW, I must be getting silly recently.. last night I obtained 450D... also with a problem (doesn't read memory card).
So, once I am finished with 400D, I will start work on 450D. One of them will surely become mono.

[glend]

I have found that Memory card problems in the 450D can be due to the top connector on the main processor board. That card module ribbon is designated by the letter (I) in Gary Honis' mod guide, this pulls out on one side slightly due to the hard bend it takes to the memory card board connector underneath. The problem may simply be a bad ribbon connection there. I have some spare 450D parts, let me know if you think you have a bad board we might be able to work something out.

Btw debayering a sensor is not an easy task, you may ruin a couple before you get a good one, i certainly did. Feel free to ask anything.

[rcheshire]

This (http://www.dutchforce.com/~eforum/index.php?showtopic=36451#)might help. Fuse value is elusive but it must be listed somewhere.

The 450D is quite hardy as is the 1000D. They are easy to adapt for cooling which is a bonus. Heat or freeze, they just keep going.

[bojan]

Here (https://canonrepair.wordpress.com/readers/) it says Canon uses only two type od fuses: "P" and "K" (3A and 1.5A respectively).
Here (https://canonrepair.wordpress.com/2011/01/29/smd-biztositek-dc-panelon-smd-fuse-on-the-dc-board/) as well..

RS components have them: http://au.rs-online.com/web/p/non-resettable-surface-mount-fuses/5413619/

Replaced the fuse and the camera is back to error 99, same symptoms.

[bojan]

It seems that P-MOSFET is dead... I can see the control signal, but it remains open circuit.
It must be replaced with a new one, IRFTS9342PbF may do.

While waiting for parts to arrive (1 week or so) I was wandering if I could simply short this FET to determine if the rest of the circuitry is alive (ohmic value is ~20kohm.. meaning there is no short circuit for test voltages used in my Fluke multimeter).
Of course, this means the battery life will be significantly shortened, but I am not concerned about it at the moment.

[bojan]

Faulty FET is replaced with new one and supply voltage to the front board (which controls shutter and flash) is now present.
However, error 99 didn't go away...
Upon pressing the shutter button, I think I felt something was trying to move (mirror motor?), and there is electrical activity on motor terminals, voltage drops from VBAT to GND, and after that - error 99.
It seems the logical next step is checking the motor and overall mechanical status of the mirror box. Manual manipulation of mirror looks normal...

Another symptom surfaced now (I haven't check that before): Autofocus doesn't work when lens is in place, and there is no activity either... but now there is no error 99.
It seems the problem (which also caused FET to malfunction) is further down the track.

[glend]

Don't worry about auto focus not working as removing the filters negates it anyway , changes internal focal length. Not an issue for a telescope mounted camera, and you can always focus a lense manually if you must. Re Error 99, there are certain steps that should be taken to clear it. The partial lockup and then an Error 99 i have seen before, if you have ckeared the fault and it still shows try this: removed the memory card, and the battery. Sometimes you will hear the shutter park itself when you open the card door or remove the battery. This indicates the shutter mechanism is reset to park. You can then reinstall card and battery and try again. Also there is a big cap in there and some models have a memory battery, you can let the camera sit fir awhile without the card and battery and it may drain any extraneous voltage held. There are a number of suggestions online for clearing Error 99.
Good luck.

[bojan]

Glen, thanks for your suggestions. More or less I tried all of that...
I am not worried about autofocus fault per se of course, but it seems to me it is the symptom that can indicate the location of actual fault(s) more precisely...
Anyway, what I really need now is the circuit schematic.. and this is what is probably not (easily) available.

BTW, opening and closing the memory door controls the FET (via main board processor) that I replaced.

Most likely the remaining fault is still on "power board"
http://www.dhcameras.com/canon-eos-400d-digital-xti-dc-dc-power-board-replacement-part-new/

I will also check the other fuses on main board as well (there are 2 more), FU02 is most likely to be blown because of the previous problem with CF card connector (now fixed)

[rcheshire]

That's another chapter in DSLR know how... glad it's all coming together.

[bojan]

Yesterday I was digging deeper into the circuitry..
Found out that there is no power supply to the lens connector (it should be there when lens is attached (pin 3, red wire).
N-MOSFET transistor provides this voltage, it is controlled via LV8015 from processor board by signal VB2ON.
And that signal is not there when shutter button is half pressed, and consequently there is no power supply (4.75V) to the lens and some other circuitry, and this the reason for error 99.
Now I have to check ribbon cable, associated connectors and source of VB2ON signal (it is coming from pin14, connector on main board closest to shutter).
Of course, the real reason for N-MOSFET not being ON could well be outside the processor board and could be because some other signals are not as expected by firmware.
That wandering pin from CF connector must have done a lot of damage...

BTW, ElectroTanya (http://elektrotanya.com/canon_eos_400d_eos_kiss_digital_x_r ebel_xti_sm.pdf/download.html) has much higher resolution copy of service manual (35MB).

Stay tuned..

[glend]

I would suggest it is likely a ribbon cable problem, as 90% of my 450D issues are traced to bad ribbon cable connections. A ribbon that is not perfectly aligned to its connector and with a good lock bar can give open circuit conditions. I have had to replace a processor board simply because of a broken hinge on a connector locking bar. I have developed a way to change 450D sensors without disconnecting the main problematic ribbons on the processor board and error incidents on reassembly are virtually nil now. Ribbons the make tight turns behind boards are the ones under the most stress of being open circuit.

[bojan]

One more check:
The encircled ICPWR3 ( marked "G2" or similar, it is not clear) is 2-input gate.
With shutter button half-pressed, there is a VCC on pin 5 as it should be, and HIGH level on pin 2, LOW on pin 1 (testpoint VB2ON). Output (signal TIN) is LOW
If this is AND gate, then there should be LOW level.
If this is NAND gate, it is blown.

I am giving up for now..

[bojan]

Details on signal path. VB2ON is not where it should be.
Or not.. all depends on what kind of single gate it is (http://www.ti.com/lit/sg/scyt129f/scyt129f.pdf), and if gate IC (ICPWR3) is healthy or not.
LPWRDET is Low power detector.. probably from Motorola.

[bojan]

The flex cable carrying VB2_ON and connections are OK, confirmed...

Then I cut the VB2_ON track, and shorted ICPWR3 pins 1 and 2.
Output (pin4, TIN) went HIGH... but TOUT remained LOW. And according to LV8013 (http://www.onsemi.com/pub_link/Collateral/LV8013T-D.PDF) data sheet, when EN and TIN are HIGH, TOUT should also be HIGH.

There is another strange thing with this LV8013... it has charge pump, output from it is ~13V, but motor is definitely powered from battery (via _-MOSFET that I had to replace earlier as faulty). At this moment, I am not sure where this voltage is used, and I can't imagine someone just blindly followed the data sheet and left all recommended components on bard, only to use only a fraction of the the functionality of the IC.

[glend]

Bojan, I admire your persistence in debugging these problems as I don't have the skills or equipment to do it. Hope it works out for you. I would have bought a cheap donor camera by now and just swapped the board. I know through my 450D work that the cost of replacement boards from reputable sources is higher than just buying a beatup working camera body off ebay, and building a parts bin.

[bojan]

But I really enjoy all this debugging :P
Business-like approach was simply not encoded in my DNA, that's it.
And yes, I am watching ebay for a donor(s)..

[rcheshire]

Just a guess could the 13V be for the flash?



This is great stuff and worthy of a book mark.

[xelasnave]

Do you have a good hammer?
Great thread.
Alex

[bojan]

Hammer is one thing I don't have... not for suggested application that is ;)
My better half says I am a hoarder :shrug:...

[bojan]

After replacement of LV8013 (motor driver + power supply for lens) my Canon 400D is repaired, error99 is no more.
On the attached image (showing layout beneath the LV8013) it is obvious that +13V is used for lens power supply (one via leads to the bottom of the power board, which I didn't have to remove to solder in the new IC).

Couple of days ago I also received the UV-IR CUT filter (Rocolax 48mm, for my Canon 400mm 2.8L drop-in adapter), so it will be possible to do some real testing tonight if weather in Melbourne permits.

Pictures attached: two at the top row are taken with same lens, with "normal" (left) and modified camera (right).
The bottom image is taken with modified camera through Rocolax UV-IR-CUT filter.