Control Unit for Dishwasher

[lodevaduh]

First thread post!

Over the years I've known that my Asko 1502 dishwasher (circa 1991) has some functions on it just flat out don't work. I believe they all used to work at one point in time, though. The "normal wash" cycle doesn't work when I press the buttons (this dishwasher has buttons, not a knob)... No response from any leds or anything either. But the "light wash" and "delicate wash" programs work, though they don't do very well.

So I thought it might be the "normal wash" button thats faulty, so I soldered in some compatible ones I found from an old racecar remote. But still nothing. This stumped me because when I put the dishwasher into a "diagnostic mode" to check its pumps and switches and such, whenever I would press the the "normal wash" button (which in this diagnostic mode would make the drain pump run), nothing happened. But I know the button itself still works...this led me to think that the processor only preforms some functions instead of all.

How do you fault a circuit like this? Could the little unit's processor be jammed (it's a 8952 microcontroller)? Are there memory chips that don't work? Or if the old processor is bad, is it as simple as buying a new 8952 and soldering it in there? I don't know a huge amount about electronics theory, I just know basics as well as how to solder.

So regardless of whether you know much about dishwashers or not, based one whats going on, what would you all say?

Thanks a lot!

 

[Resqueline]

If you can track the button traces back to the cpu and verify that they're not broken or something then the cpu itself must be faulty somehow.
It contains 8 kiloBytes of Flash memory so you can't just exchange it, and i'd expect the program to be locked so you can't just read it out.

 

[lodevaduh]

Thanks for the reply, Resqueline.

I actually looked at the traces, and I saw that the broken button in question (normal wash) went to another button that was faulty (pots and pans wash)...Could this mean the two buttons will only work if both of them work? Perhaps they depend on each other? It would make sense, considering the manual says that those two button should be pressed in succession for a particular wash cycle. Is this a common thing in circuits...linking buttons to other buttons? (Excuse me if it's a n00b question.) By the way, they're those kinda buttons that have four terminals in a rectangular shape, with a long skinny push coming out the top...

Thanks again!

 

[(*steve*)]

My guess is that the tracks to these switches are broken somewhere. You could try tracing them back as far as you can and using a multimeter to confirm continuity.

Whilst I think it's unlikely that a single pin on the controller has gone bad, it's not impossible. Two of them failing seems less likely unless there's a design fault. Being manufactured in 1991, I would presume the build quality is better than some stuff you'll see today.

If there are any flexible connectors, it's not impossible that if this is a well used device that vibration may have broken wires or tracks on them.

 

[Resqueline]

Yes, it's common to connect buttons in a matrix fashion.
The more buttons there are the more pins are saved on the microcontroller (4 x 4 = 16 buttons but uses only 8 pins).
If one button has a short of some sort then this can more or less lead to the cpu "hanging" or behaving erratically.
Those buttons have two wires passing through their body (so two & two terminals will always be a short).

 

[lodevaduh]

Thx again folks,

I replaced that second switch today, hoping that might cause the other one to work...But same thing... HOWEVER, this dishwasher has diagnostic modes for testing its components, so if you strike a certain key a certain amount of times, it will put the machine into a service mode. For one particular test, it involved pressing the pots and pans button 5 times, to get into a mode in order to test water level (blah blah blah). Here's the funny thing: I could press the pots and pans button 5 times, and it would go into its service mode alright; the button appears to function. But if I press pots and pans ONCE, like if I actually wanted to wash something, it wouldn't start the wash cycle.

So I'm thinking my buttons are fine...I even tested the new solder joints with a multimeter, and they were fine (although I will check the tracks, steve). And yes, steve, there are two sections to this unit connected by flex wires (soldered directly to the boards, and they kinda resemble IDE cables in old computers) There's small board with mostly buttons, diodes, and resistors on it, and then the main board with the ICs and power connections.

So, what are the chances some stray capacitor or diode or transistor has failed? Does any component pertain to buttons more than others? Well, i guess it IS a circuit, after all...every little piece matters.

Thx again for the help!

 

[(*steve*)]

My gut feeling is that if the button works for getting the unit into the diagnostic mode then it is working. I would imagine that the designers of this unit use the normal key scanning routine to detect the presses that perform the function or that enter diagnostics.

That does not preclude the possibility that there is a problem with the other button though. This is all speculation, but if the "normal wash" button is stuck down, and the logic for the "pots and pans" button requires that it be pressed *after* the "normal" button, the problem with the "normal wash" button will affect both functions.

Note that following the same logic, the normal wash button permanently open circuit would exhibit the same symptoms and would not compromise the scanning of other buttons -- and it's more likely (especially after replacing the original).

It's most likely that the buttons are a simple matrix, but it is possible that they might use diodes, especially if the same pin is used for more than one purpose. If you find other components connected between the switch and the processor pin, it may pay to try to draw out a part of the circuit.

You are likely (but not 100% certain) to fund pull-up or pull-down resistors so that when the button is released, the pin is pulled (say) to ground and when pressed it is connected to (say) +v. It also might be exactly the other way around. If you find one, check the solder joints and possibly the tracks too. If the resistor gets disconnected, and it is pulling against any internal pull-up or down on the chip, the button will again do nothing.

It may also pay to check that the pin on the controller is soldered too. If it's surface mount it might *look* soldered but not be. The problem is that even the pressure of a test probe may be enough to make a good contact again (temporarily).

 

[lodevaduh]

Re-do microcontroller?

Sorry I haven't replied in a while...

While looking around at projects around the web, I realized something (this is assuming it's a problem with the microcontroller)... I know its possible to burn software to a microcontroller, but is is possible to read software on an existing one and copy that to your computer, get a new chip, then copy the software to the new chip?

Is that practical? THx!

 

[(*steve*)]

In some cases, yes. For commercial products, probably not.

Most of these devices have some sort of protection which, when enabled, makes it really hard to read the software. For all practical purposes, for the average person it becomes impossible.

Commercial devices usually have protection enabled because it allows them to protect an important part of their design and thus to prevent cheap copies from being made.

I'd still be leaning toward a simpler hardware fault though. Did you check the traces all the way back to the pins?

 

[lodevaduh]

Steve, I haven't checked the traces recently, as I've been busy with lots of things, and in need of a dishwasher each night . So I'm just speculating and making a list the next time I open it up, which may be soon .

So with circuit traces, how should I attack it? Should I just use a multimeter and follow the switch's path?

I'll be sure to post here once I open the washer up again

 

[(*steve*)]

I would trace the path of the leads from both switches. One end may be common with other switches or may be linked directly to a supply rail. The other end may find its way back to a pin of the micro controller. (Of course more complex arrangements are possible.

In general, you want to make sure that there is continuity along the pathways which lead from the switch to a known good part of the circuit.

For example, with the switches open, I might expect to find that one pin of each switch has (say) a 10K resistance to gnd. If the "faulty" switches have this, then that connection is probably OK.

The other switch lead may go along a relatively long path, cross one or more interconnects between boards, before finally reaching its destination (fingers crossed, a uC pin). There should be continuity from the switch to this pin. It is probably best to keep one probe of your multimeter attached to the switch and trace along the track with the other lead to ensure that it maintains continuity. Potential trouble spots are places where the connection goes off-board or via another component (a resistor or link perhaps).