Thursday, 24 September 2009

When the toaster gets toasted

This is a couple of notes I made as an exercise while trying to determine whether I could fix my toaster. It had all of a sudden and quite unexpected stopped working, so I decided to take an hour or two to see if I could do something about it. These are the notes that documents the work. It's published to add to the amount of tester stories on the internet, as example of exploratory testing notes.

The toaster is used almost every morning to toast bread for breakfast. It's occasionally used at other times, but this is the main purpose of the toaster.
It stopped working on the morning of the 21st, which happens to be a Monday, and since we haven't been able to toast bread, unless using the oven, which works, but is overkill for toasting a couple of slides of toast.

Normal working procedure is to load the toasts into the two slots on top and push down the handle, until it clicks. The clicking indicates that the handle/lever is now locked and that heating has started, which usually can be visually checked after a few seconds as it starts glowing red down the slots. In addition you can feel the heat if you hold your hand still over the toaster. There's a timer on the toaster, which is adjustable, and when the time is up, the locking mechanism unlocks and the toast is moved up, by the work of a spring.
An observation done some time ago is that the toaster will not be able to lock if it's not plugged in - ie. no electricity, no locking.

The malfunction happens in the clicking part: it clicks, but it doesn't lock. Also, if one holds the lever down by hand, there's no heating done.

I decided to hold the lever down for a longer period. Nothing happened, so I was now sure that the heating doesn't start. I figured that this means that the lock and the heating contact was connected in some way. As the lever is mechanical and the heating is done by electricity, there must be a mechanical contact that works with the electrity.
By pushing the lever down until the clicking is heard and releasing it, and repeating this a couple of times, I saw that there's a flash of light from somewhere to the right of the lever. However I couldn't see exactly where it came from or what did it.
But my theory was that this was where the contact was.

Next I unplugged the thing and turned it around. It was full of bread crums, that fell all over the floor. I shaked it over the sink until no more crums came out, although it was visible not 'crum free'. I turned it over and tried to find a way of dismantling it.
Luckily there were four screws underneath, along with some metal fittings, that seemed to be twisted. Looking from top down again I realized that there was at least three major parts of the device: a plastic top, a metal core and a plastic bottom.

I had to take it apart and started with the four screws. All were very hard to loosen, but eventually all seemed loose. The top case came easily off in one end, but in the lever end it was stuck. By using the screwdriver to make a little hole I could see that the lever was in fact fitted into a place from where it could not be removed without destroying it. Although there was no trace of glue, the plastic fittings could not be twisted, turned, pulled or pushed without it would break. Tried all to a limit, and nothing showed that it was doable.
By pulling the entire cover away to the side of the lever, the timing device dismantled and it was possible to push the top cover sufficiently up, to reveal the electronics at the bottom.
It was slightly greasy, which I though was interesting, considering it wasn't normal operations to actually put grease on the toast prior to frying it, and there were small crums stuck in the grease. However, the electronics looked okay. I expected something like a black spot or a darkening where the 220V where transformed into something smaller, but the components where all looking like new, ie. no visually wearing or signs of execcesive heating and the board was the same.
I decided to use the air compressor to clean it, but the grease still made some crums stick.
I then turned to the springs and concluded that they all looked operational and well put together. I saw no 'loose ends', broken or bended plastic bits, or other indications of wear.

I decided to try if cleaning alone would do it. Used some paper to clean the electronics - especially the contact. The contact made me think of electrically induced magnetism: when the contact was connected, by means of a small piece of plastic mounted on the lever, it would connect with electricity which might hold the contact together in normal operation until the timer would cut the electricity. So I made sure no crums where in between the two sides of the contact.

However, putting the parts together showed no progress, except that a little heating could be felt, if the lever was held down, but only for a few seconds. There was no smell of burned plastic (=bad electronics) or even burned breadcrums.

Having cleaned the device and electronics I couldn't find more things to clean.
Having inspected the electronics visually I could decide upon measuring whether a resistor, or other device was working, to the limits my measuring gauge would allow it.
However, having no tools around that I could use to fix such a problem, I discarded doing that.
Having inspected the mechanical parts and found no signs of anything broken or misplaced, I reckonned that this wasn't a mechanical problem.
That put me back to the electronics. I got the idea that I left out inspecting the cord.
I went to take a look on that, but it looked brand new. I decided to measure the cord resistance, as it might indicate whether there was a problem with that.
It showed nothing: the cord was working well.

The next thing was that I figured (from my notes ;-) that I didn't really understand how the locking down thing worked. The magnetism idea was spawned more or less from the observation of the contact, which looked slightly coily, and the knowledge that without electricity, the locking mechanism wouldn't work. So I went to check on that.

Taking a closer look on the locking mechanism, I found a rip in the plastic I didn't notice earlier. It was well hidden behind a small piece of metal. However, moving the plastic piece, which was hinged in the left side, and could be moved down by the lever, and by the means of a spring restore itself to a tilted position, showed, that the ripped part had no mechanical purpose and looking closer at it, it appeared to be where the plastic piece had been mounted to the plastic frame it was shaped out of.
However, the metal piece, that had hid this part from me earlier turned out to be one end of another contact. So: there's a contact activated, when the lever is turned down, and another contact activated when the lever is 'locked' down. I noticed that the lever was fitted with a small plastic 'hook', that albeit rounded on the edges would lock the lever on the ripped plastic piece. Underneath the hinged plastic piece I saw a small piece of electronics, which I have no name for, but which consists of a coil with a rod in the middle, which will move up, when current is applied. (note: the right name is an "actuator").

So: I changed my model of operation: the first contact would simply connect shortly when the lever passes it on the way down. This will start the timer. The hook on the lever will lock on the hinged plastic piece, which will hold the toast down, and all the while the second contact will be 'on', providing the electrical circuit for the heating (and possible the timer). When the timer runs out, the actuator will push the hook on the lever upwards, thus releasing it and at the same time cutting the current.

Now - the problem is: what could make this system malfunction ?
Both contacts where Y-shaped, with the two ends pointing down. The lever/hinge would push on the one arm of the Y, and the other end would move towards the other pole of the contact. I decided to make sure, that there was nothing hindering the second contacts poles in touching each other, and also - using a small screwdriver - widening the Y's angle, to ensure a better contact.

I assembled it all again, and plugged it it. Yo and behold: it locked and heated!

I decided to make some toast and try it for real: it still worked!

This is a write up of the notes I made. I tried not to take things out, although I could have written all this as 'found a way to open it - cleaned it - mended the Y-contact and now it works', but that wouldn't show the process of despair I went through. I actually thought this was not going to work. Also, you could notice that I found a lot of stuff in this, but only one thing was part of the solution. However, you don't know where to go to find new stuff, do you ?

I'd like to point out, that without the notes and the process of going over the notes, I had probably thrown the toaster out and bought a new one without further ado. Only through the notes I discovered, that my explanation or model of the locking mechanism was inadequate. Besides showing how incredible stupid I can be (I mean, Really, Carsten - a magnetic lock in a household toaster, Really! :-), it also shows, that when no progress is made, you might start questioning your basic model. I think that correlates to many project problems I've encountered and heard of throughout in my professional career.
It's no good to be locked in your initial model - especially if it's wrong.

Now, I will go make me some toast.. :-)

Friday, 11 September 2009

Where's the 'on'-button ?

Twice I've recently bought a new, electronical device and twice the manual insisted on starting it by pressing a button, which does not appear to exist.

On one of the devices there's supposedly an 'on'-button, according to the manual. However, on the physical device there's a '1'-button.
On the other there's an extra button, which you have to operate, but it's not mentioned in the manual. And the drawing of the product does differ from the actual design of the product it was shipped with (but not so much that I suspect it's a mixup of product and manual).

I just wonder: if the producers don't care enough to update this 'manual', which is about 50 words long and has a small drawing, what else don't they care about ?