Kleenmaid / Brandt tx768 repair

[gizmo]

I have owned a nifty little h-axis top loader for a few years now. It is a Brandt TX768, badged in Australia as "Kleenmaid by Brandt." It was made some time between 1995 and 2000, in France.

I got it cheaply from a recycle centre, it worked OK except the heating part of the cycle was unreliable - most of the time it would skip through the heating phases of a cycle, doing a cold wash. (It only has a cold water fill.) Occasionally it would stop and heat as normal. I didn't put much time into it, but I could never figure out why.

I like the machine, it rated #1 in the Choice test when it was new, it does a good size load in a tiny footprint and spins at 1100 rpm, and is exceptionally water and energy efficient. When washing it uses around 100 to 150 watts, spinning is 200 watts at full speed which is unusually good for a machine of that age. Heating is 2000 watts, but it doesn't hold much water in the wash phase, so it heats quickly. (When it works.) It has "posistop" which means it always stops with the trapdoor to the drum facing upwards - it uses a magnet on a drum pulley, sensed by a reed switch, so it knows which way is up for the drum. It was stupidly expensive when new (over AUD$1800) but I got it for $15.

Ronhic from this website kindly sent me a link to the owners manual when I first got it. (Thanks.) Recently I brought this washer to the front of the shed as I wanted to use it and see why it would not heat reliably. I found repair manuals online, which detailed a diagnostic sequence. It all tested OK, and did a couple of loads properly. Then it started skipping the heating phase again - timer gets to the heat phase, heats about 15 seconds, then clicks to next increment and heating stops. The service manual tells me that this is the control module finding a fault in the temp sensor (thermistor) or temp setting potentiometer, so it skips the heat phase and completes a cold wash. The pot tested OK in the diagnostics so the fault had to be either the thermistor, wiring loom or the control module itself.

I tested the thermistor with my multimeter, and it tested OK. It misbehaved again, I retested the thermistor and it was open circuit. I removed it, and found if I bumped the thermistor, it would give an erratic connection. (It mounts in the perimeter of the heating element mount, and shares the same rubber seal.) New ones are available, but too expensive. I hacked away the plastic shroud where the wires connect, and found I could remove its innards. I was expecting a thermistor like I have seen on dozens of circuit boards, but it was something rather more crude. Two slender metal spring fingers pinched a tiny disc of heat sensitive conductive material. The electrical connection depends on the spring tension of the fingers against the disc. But the finger tips were tarnished and greenish, so I cleaned them off with fine emery paper and reassembled. I sealed it up with silicone sealant. It is drying overnight, test run is tomorrow. The repaired thermal sensor tested fine with the multi meter this time.

Photos of the machine and the sensor repair:

 

[jetcone]

Thats a nifty machine

And a pretty nifty repair. I ve done that with motor switch boards, its amazing how a tiny bit of corrosion or grease can just stop a motor or sensor.

 

[henene4]

So, is this an NTC or more like a bimetalic sensor? Never thought about what these looked like on the inside.

The early 2000s were the time of the most interesting machines in terms of controls IMO, especially on the cheaper lineups.
Combinations of mechanical timers and PCB control were a thing then, shortly before pure PCB control became cheap enough for any pricepoint.
That led to verry interesting and elaborate machines in general.

If it is a NTC it is more or less like a semiconductor and when semiconductors fail, they often behave extremly weired.
The place I once lived at had a Miele in the common laundry that would boil away once you wanted to go beyond 60C. Would continue to heat until the overheat protection activated or the timer errored out, throwing a heater error. Even managed to warp the detergent drawer!
It could just be that the corrosion on the surfaces inhibited the heat transfer.
If your fix works I'm really happy for you!

If it dosen't work and you can't find a cheap enough replacement, you could search for a datasheet for the old sensor and find one that has close enough characteristics.
The tabel with the specific resistance for specific temperatures might even be in the service manual.
Even if it might be different in size, some silicone should be abled to fix that.

Otherwise you could try to rig it up with some different system to select the temperature.
But that would be more like a DIY if you don't have anything to loose anymore.

 

[gizmo]

Hi Henehene4

It isn't a bimetallic sensor, they respond to a single temperature and "click" at that temperature. It is an NTC disc. I was expecting to find something like the first photo below, but it is more like the second photo below.

You are right about this being an interesting era for complicated interface between an electronic module and a mechanical timer. I have never seen anything quite like it. I am familiar with older machines with a mechanical timer, where the electronic module is just a motor speed controller. I have also seen hybrid machines where the electronic board is integral with the mechanical timer, and but they have still been that the mechanical timer is the "master" controller and the electronics were subsidiary. I am also familiar with newer machines with 100% electronic controls, and prefer them overall.
This Brandt machine is different. The timer dial still clicks its way around the cycle, but the decision to move from one increment to the next is made by the electronic module, and every peripheral component except the heater element is switched by the module, not by the timer. (so even the timer motor is controlled by the electronic module.) If the module detects a fault, it commands the timer to move to the next increment. (But it doesn't abort the whole cycle, it still attempts to complete a cycle as best it can.) I don't really understand why they were made like this, the mechanical timer must have added a lot of expense compared to pure electronic control, and every wire in the interface between the two is a potential bad connection.
in my case, there are two heating phases in the cycle, the first one heats to a max of 40 degrees C (or less if the temp dial is set to a lower temp),then there is an increment where it does a bio-soak at 40 if the bio button is pressed, otherwise it tumbles a few seconds then increments to the second heating phase, where it heats to the set temp if over 40, then increments to the long washing tumble.

the diagnostic cycle is strange to initiate - you select spin, wait till the fourth sequence of slow tumbling, then rotate the spin speed selector to maximum, minimum, max, min, and max again. This is detected by the module and it enters a diagnostic sequence, and you manually move through each step by moving the temperature dial from max to min or min to max. It seems to me like a very crude interface for a very sophisticated function.

Well I am going out now to fit the repaired sensor and cross my fingers it works...

 

[henene4]

I know washer and dryer NTCs as things like those pictured below.

General design is pretty consitent throughout brands, but specific values and sizeing differs.