Dishwasher: Hardware Modifications

Sun, 09 Aug 2026 00:00:00 +0000

Adjusting The Body

After I had forged a new brain for my broken dishwasher, it needed some modifications to its hardware, or its body, if you will. In this post I will describe what changes were made. I’ll spill the beans on how to sample two switches on one input, what’s behind the control panel and how I plan to shave a few pennies off my electricity bill. Let’s get into it!

Automaton Anatomy

As the overall system of the dishwasher was thus far only described in text, we’re due for a graphical overview of the system.

System block diagram

Originally the device also had a rinse aid tank and a solenoid valve to dispense it during the final rinse. Since the new controller has an insufficient number of outputs, I dropped this functionality. I plan on drying the dishes by hand after washing, so rinse aid is not essential anyway.

Senses

The HEDW44004 uses a temperature switch (also known as a Clixon) to control the water temperature. When the water temperature rises to approximately 60 degrees Celcius, the switch closes. When the temperature has dropped to about 45 degrees, it opens again. In the new design the temperature switch will be switching 5V DC from the control board instead of the 230V AC.

Interfacing the float switch became a bit more complicated than it used to be. The float switch consists of two dual throw switches which together detect three discrete liquid levels. The thing is, most of the microcontroller GPIO was already assigned to other functions. I also wanted the board to have a spare input for future upgrades (leakage detection was especially high on my list). With the board only having three inputs and one already taken by the temperature switch, there was only one input left for the float switch.

Staying afloat

How can a float switch with three discrete states be read by one input? The answer is: analog multiplexing. The three discrete states are converted to a voltage by changing the values of a resistor divider network. This gives each level its own voltage. The ADC then reads this voltage and the firmware converts it back to one of three states. If the voltage levels are far enough apart, the three states can be de-multiplexed without too much trouble.

The circuit below depicts the switches and voltage divider. The lower resistor is of a fixed value, and the upper is determined by the state of the switches. The voltage across the lower resistor changes in discrete steps, depending on the liquid level. If you look at the picture above, you can probably envision the process.

Three in one

The switches used here are the ones that came with the dishwasher. They are intended to switch high voltage AC instead of the 5 volts they will switch in their new job. Low voltage signal switches are gold plated internally to prevent the contacts from oxidizing. High voltage switches don’t get this (more expensive) surface treatment. Their higher voltage can easily break through any contact oxidation. This lack of plating makes them less suitable for low voltage signals. This can be overcome by increasing the current through the switch. In this case, the total switch current is at least 10mA (which is high for an analog signal).

Frontal Lobe

The old mechanical brain used to be mounted to the frame behind the control panel. The new electronic one is quite bigger. However, it can still be fitted to the frame. It does need new mounting points, so I drilled holes in the frame to fit threaded stand-offs to hold the board.

The connector for the mains powered components is near enough to where the old timer used to be, so the wiring did not have to change much.

Brain mounted to the front panel

(inter)Face

The original control panel had two button switches, two incandescent lamps and a timer knob to set the program. The light bulbs have been replaced by LEDs and one of the switches has been replaced by a tactile switch. The second switch remains in place and keeps its original function: turn on and off the power. The picture below shows the control panel without its cover. You’ll also notice the speaker that I once wasted so much time on. It remains hidden behind the panel, but won’t be making any sounds.

Front panel au naturel

The timing knob landed itself in a new job. In order to safely test and update the firmware, I needed a programming port in a location where no live wires live. The control panel is just such a place. The hole for the timing knob shaft is just the right size to fit a piece of perf board with a flat cable connector. After cutting off the shaft, I had a perfect (removable) cover to hide the programming port from view.

Timer Knob…

…turned programming port

Fur

The dishwasher being in a partially stripped condition on my workbench offered the opportunity to add another upgrade: thermal insulation. The machine originally didn’t have much of it, and wasted a noticeable amount of power on heating up my kitchen cabinet. I wrapped the dish compartment in a material called SuperWool, which - I assume - grows on some form of bio-engineered sheep. It won’t catch fire up until temperatures of 1200 degrees Celsius. It’s advertised for use as thermal insulation of chimneys. It is also, as opposed to some other thermal insulators, non-carcinogenic. That’s definitely a plus if you plan on using it indoors. Or anywhere, really, in my humble opinion.

Looking sheepish

Respiratory System

After running the pump for an hour or so, my nostrils met an odour that was uncannily similar to that of an electric motor running hot. This happened to be an unwanted side effect of wrapping the dish compartment in artificial fur. The dish compartment wasn’t the only part of the machine where heat accumulated. The thick layer of insulation inadvertently blocked the airflow from the motor to the ventilation slits in the rear panel. There are two such openings on the device; one near the top of the panel, and another at the bottom, right behind the motor. With the motor being actively cooled (it has fan mounted onto the back), I naively assumed the topmost slits could be dropped. They couldn’t. With the insulation in place, the airflow got constricted. Luckily, the problem was easily solved by drilling some extra holes with a step drill.

Inhale… Exhale…

Will it Wash?

That’s the question to be answered in the next and final post of this series. Then I’ll discuss the simple program that makes the device do its dishes.

Thermal Insulation on project392.net