The Ideal Microwave Keypad (and operation)

[MrAl]

The Ideal Microwave Keypad


Hello there,


Recently i had to buy a new microwave oven as the old one was having problems. The keypad was
all messed up in that some of the buttons would work sometimes and not other times. For example
sometimes the '3' would not work at all, and i was using the sequence "333" for a time setting of
"3:33" quite often, so that was a problem. Pressing harder didnt help either and i found out
why later.

After peeling up the keypad i found that pressing the key harder didnt help because it was
not that the key was not making contact with the substrate, it was because the key was
CONSTANTLY making contact with the substrate, thus the control circuit would not see the
key anymore. I was able to get around this for a while by peeling up the upper substrate itself,
and that would relieve the key 'short' and so the key would work for a while again.

After about a year of this, i finally decided it was time for a new one.

I ended up with a Panasonic microwave, one of the "Inverter" type ovens. This kind of oven
is quite a bit different than my previous ovens, in that it does not turn the magnetron on and
off in order to achieve low duty cycles which produces a lower power (like power 2, 3, etc.)
instead of the full power. What it does is actually lowers the input power in a pseudo linear
fashion such that although now the magnetron is always on, it actually gets lower power while
cooking on a low power setting. One exception to this is power levels 1 and 2 which lowers the
magnetron input power but also turns it on and off to get those two lower settings. I think
these two levels are a little strange, but i dont seem to have a problem with it.

So after saying that i have to say that the inverter feature is the best thing that ever happened
to microwave ovens. Being able to set the power at say 3 and have it constantly stay 'on'
seems to work really well for some foods. But unfortunately there are some other 'features' on
this particular oven that are incredibly ridiculous.
First, there is no way to turn off the 'beep' when you press a button. This gets very annoying
very fast.
Second and more annoying, in order to set the power you have to press the 'Power' button
several times to get the right power setting. In other words, if you want power 5 you have to
push the power button 6 times (6 not 5), and if you want power 1 you are in for a chore, you
have to press the power button 10 times! Yes that's right, 10 times, 10,9,8,7,6,5,4,3,2,1 !
But it doesnt take place that fast, it takes a few seconds just to press it that many times.
There is an option where you can hold the button in and it will count down, 10,9,8, etc., down
to 1, but it takes even longer that way. And if you accidentally press it one more (friggin')
time you have to start all over again because it goes back up to start at 10 again !

So here's where the "Ideal Microwave Keypad" comes in...

What im thinking of doing is taking control of the keypad with a surrogate keypad that is scanned
by a micro controller and then the micro controller 'presses' the keys according to what the
microwave oven really needs, while allowing just about any predetermined key sequence to do whatever
we want it to do. For example, for power level 1 i could just hit "1" and then enter the time,
or as more conventional ovens do, press the power key 'power' and then the number "1" to get
power 1, then enter the time such as "333" for 3:33 as before.
That's just one or two examples. The micro controller would then take over and press the
real oven power key 10 times and then press the "3" key on the oven three times for the 3:33
time setting.

So the actual concept itself is quite simple. Use a microcontroller to take over the keypad
which would allow a more conventional time and power setting.
And since i like to set the power every time now, i could skip hitting the power key and just
enter the power first (as one example above) and that would mean even less key presses to get
cooking
So that would mean only having to hit the key sequences on the surrogate keypad:
"1, 3,3,3, Start"
and that would it, done!

So anyone ever do anything like this before with a microwave oven? I know it would void the
warranty so i would not attempt this probably until after it expired anyway, but then there is
also the possibility of using actuators on some of the important keys. I'd need one for the
keys 0 to 9, and one for start, and maybe one for stop just so i could program several start
and stop times without needing the microwave's internal programming to do that with (which is
also limited).

Also any ideas on what actuators would be good for this? They dont have to be super strong
but have to simulate a reasonable human finger press on a key. In fact, this is probably the
most important thing as i think i am opting out of digging into the keypad electronics.

So my two questions are:
1. Anyone else ever do this?
2. Any ideas what actuators to use? I think i'd be very flexible on the price for these seeing
as this is something i use several times a day.

 

[Nigel Goodwin]

So after saying that i have to say that the inverter feature is the best thing that ever happened
to microwave ovens.

Depends on your point of view

Essentially it makes the ovens unrepairable - if the mag fails, or the HV diode fails, it takes out the inverter as well.

Every single one I've had for repair has has both mag and inverter duff - making them well past repairing.

 

[fernando_g]

Unfortunately Nigel, this trend seems to be a constant on newer equipment.

They may have better features or performance, but when they fail they are impossible or un-economical to repair.

 

[MikeMl]

It also occurs to me that Mr Al must have a lot of time on his hands...

 

[Mr RB]

Sounds like it!

I can huge problems in the interfacing, because the button pad will be matrixed and also time debounced etc. The matrixing you could eventually scope out but the debouncing stuff will have to be guessed.

The best option would be to just control the output stage of the oven's PCB, the bit that controls the magnetron and relay/inverter etc. It's a shame you didn't get a traditional on/off type microwave that would be so easy to interface.

 

[dr pepper]

Never done it with a microwave, but looked into it with something else, you used to be able to get a 4000 series cmos chip that was a crosspoint switch, you could connect one or 2 of these to the keypads matrix and address the switch you want to close.
But why you'd want to go to that length?
Might be better to look at the power supply board and see what signals go to it from the proc board, there might be a chip on there you can get a datasheet for and write yer own system, or it might be something simple like a voltage out.
As usual extreme caution advised, magnetrons run at thousands of volts.

 

[MrAl]

Hi again,


Some good points were brought up here some of which i dont want to think about


Nigel:
Yes, very good point. I am hoping it lasts a while so i dont have to repair it or
replace it or anything like that. I have no control over this though.
My viewpoint was from the cooking aspect. Based on my knowledge of power heating,
i would have never guessed that a constant lower power heating would be that much
better than a pulsed low power heating. Apparently the constant heating somehow
cooks the food more gently.
There's another interesting advantage however. That is with regard to the input
power to the oven (not the cooking power but the actual watts going into the oven).
For example, on a conventional oven that is rated say 1000 watts the input power
might be 15 amps for that 1000 watts of cooking power. So when i set it to a
power level of 5 it might pulse on for 15 seconds and off for 15 seconds. But
during those 'on' 15 seconds it draws the fully 15 amps, and i dont think there is
anyway around this except building our own inverter or using an expensive variac.
But with the inverter oven, when you set it to a power level of 5 instead of drawing
the full 15 amps on and off, it draws (roughly) 7.5 amps from the wall power line.
This is an advantage if we want to use other equipment at the same time as we save
7.5 amps for another piece of cooking equipment or whatever.
Here the line voltage drops lower by quite a bit too whenever something like this
is turned on, so it's an advantage there too as the line voltage doesnt drop as low.
Of course to someone else this depends on the kitchen wiring and how many circuits
they have and what else they want to run and how stiff their power lines are. But
here it helps quite a bit.
But yes, there's the reliability issue. I can only hope nothing goes wrong with it.
Perhaps you can suggest some tips for repair if it does konk out on me

MikeMI:
Well, i dont mind spending a little time here and there with something that i use
every single day. The things i dont have to use every day i sort of hold off on
usually.

MrRB:
Well, it's not as hard as it sounds. That's *if* i was going to go with a purely
electrical interface (see post after yours). I've done this with a chess computer
and that had 64 squares and several other control buttons. It's all about connecting
to the switching matrix as dr pepper points out. If there are say 16 keys, that
could be 4 scan lines a,b,c,d by 4 scan lines 1,2,3,4, and shorting 'a' to '1' might
be the same as pressing the "1" keypad number key. The 'on' time is found experimentally
but that doesnt take long, and something on the order of 50 to 100ms would probably
do it.
But alas i dont want to have to connect electrically, at least not right now. I would
rather design an actuator device which can press the keys mechanically. This would
mean i could do it soon and not have to wait for the warranty to expire.
Yes, sounds like a lot of construction, but i think it would be interesting.
Problem is, i havent yet looked for any actuators so i was hoping someone here might
have done this before or at least know of some that have enough power to act like a
human finger pressing a button. A strong solenoid perhaps? Any ideas?

dr pepper:
You are 100 percent correct
I did this with a chess computer a long long long time ago. The board was an 8x8 matrix
so i used one of those CMOS analog switches. It's the 1 of 8 type, where 1 line gets
connected to one of 8 other lines through the input data word which is 3 bits. Using
two of these analog switches, we can 'short' any matrix switch point and therefore press
any key on the 64 square board. I dont remember what the press time was nor the time
between presses as there is some minimum 'rest' time as well.
This was connected to a PC computer (a very old model) and the PC computer was able to
take control of the chess computer.

But for this project, at least to start with, i'd like to use actuators. This BTW would
also allow for migration from this particular oven to any other oven on the market just
in case this oven craps out and i need to get another one of a different make or model.

So i guess the main question then is what makes the best actuators for stuff like this?

 

[wiliswene]

I've had for repair has has both mag and inverter duff - making them well past repairing.**broken link removed**

 

[jrz126]

I also have a panasonic. Had to google it to figure out how to adjust the stupid power setting. I made the mistake of entering a time, then adjusting the power to medium. it didn't do anything when i pressed start, so i re-entered the time. Walked away and came back later giving it plenty of time to cook and then cool. It was still cooking when i came back and now burning. Turns out it cooked for X time on high, then X time on medium.
I have yet to figure out the "more/less" button.

Why not interface directly to the keypad first, then upgrade to actuators later? Would make it much simpler.

Also, can you add a popcorn function that determines when to stop by listening to the kernels popping?

 

[MrAl]

Hi there jrz,

They have manuals online for download.

I dont think i can make it any clearer than i did previously, in that i dont want to void the warranty. Taking apart the keypad would certainly void it so i wanted to start with actuators and i thought that would be kind of fun to do too. Also, this means i could transfer the mechanism to another oven at some point if something goes wrong with the oven and i have to get a different model or even the same model as i would not have to rewire that one too

Yeah the popcorn sensor would be nice. I suppose a microphone plus a bandpass amplifier to reject every noise except the popping noises. Thanks for the idea.

What i've always wanted in a microwave was to be able to sense when water boils in a container. Cooking spaghetti would be a snap.

 

[dr pepper]

Mr Al, not quite what I meant, I think you meant a 4053 or whatever which is a analogue mux, I meant a crosspoint switch, a different chip, heres one allthough its academic now though:

https://www.electro-tech-online.com/custompdfs/2013/01/MT8809_DataSheet_Sept11.pdf

Jrz, you can get 'auto' microwaves, I've even seen them with a popcorn setting, they use a moisture sensor, someone on this forum posted one a while back, it was a hardened moisture sensor with a built in heater to burn off the moisture.

 

[MrAl]

Hi,

Yes i see what you mean now. So by your definition then i ended up making my own cross point switch. Same basic idea though.

Mine has a pop corn setting but i havent tried it yet.


So for now i guess all i need are some actuators, electric, not hydraulic nor pneumatic.

 

[Mr RB]

You don't need a crosspoint chip. The existing matrix will scan by setting one axis lines HI or LO in sequence. And at the same time examining the other axis lines.

Your PIC could detect the scan lines going Hi or LO, and then makes the other line go HI or LO to match, which would fool the MWO micro that a button was pressed. It would not take that long to scope out the scan sequence and timings.

...
But alas i dont want to have to connect electrically, at least not right now. I would
rather design an actuator device which can press the keys mechanically. This would
mean i could do it soon and not have to wait for the warranty to expire.

Warranty issues, really? You could take that thing back to the store missing the inverter and magnetron and they would just junk it, fill out the form and give you a new one. Under warranty replacement they are not going to open it up and see if you connected wires to the innards.

As for connecting Frankenstein like solenoids to the front panel, it sounds kinda dumb. Even with large solenoids that are strong enough to get the high surface pressures for the membrane buttons it is going to be big, slow, ugly and noise, and STILL have to press 10 slow beeps to cook something. Would you really be happy with that solution?

I would open it up and have a look how the inverter is controlled. Or if still worried about the warranty try to obtain a schematic, and then we could ALL have a look!

 

[dr pepper]

Sounds like you've done keypad fooling before mr roman.

And I agree with the frankenstein electric hammers duct taped to the front not being ideal.

And yep unless theres great big scrapes all over it returning it under warranty shouldnt be an issue, by the time its got to the workshop (I think most just go in a skip or sold off cheap) its too late anyways.

 

[MrAl]

Hi again,


dr pepper:
Thanks for the heads up back in post #6 about the high voltage. Always a good idea to be aware of that.


MrRB:
Yes you're right. A little more investigation would reveal another switching method.
I'll consider that if i do decide to go inside and start modification that way.
And controlling the inverter itself may be an option too then, but im waiting to see
if i can find another way for the time being. No doubt though that will solve the
problem for sure.

I'll also think about that warranty again and if it matters that much. Some good
points to think about.

Ha ha, the Frankenstein mechanism would not be like solenoids mounted on the front with
duct tape (but that's a cool idea ha ha). It would be solenoids (or other actuators)
mounted on the SIDE of the oven with a series of lever arms that transmit the force
of the solenoid to the button. So the entire ensemble would stick out only about 1 inch
from the front, with the side protrusion about 5 inches or so.

But in my search for the perfect actuator now after coming across several nice models
that cost upwards of 30 dollars (USD) each, reality sets in. I've since then found
cheaper models, but that got me thinking...
The main problem is that the oven POWER button is a pain to use. So perhaps taking
control of that one button will be good enough. After that button is pressed the
required number of times, i can then enter in the time right on the original keypad
which isnt a problem. So really i might get by with one actuator which works with
the power button only.
I'd still need another keypad, with buttons 1 through 10, so that's 10 buttons,
but that's not too much of a problem as i have lots of momentary push buttons around
already.
So the sequence would go like this:
1. Push the secondary keypad button 1 through 10 for the power setting, then
2. Enter the time as usual.
I'd have to wait a small amount of time between 1 and 2 above, but it might be
good enough. With the full surrogate keypad i would not have to wait, because
while i was entering the time the uC could already be incrementing the power
setting.
So the drawback to using only one actuator is that i'd have to wait a short time
after entering the power setting on the alternate keypad. But i think it may be
worth a try before digging into the electronics, dont you? If it turns out to
work good enough i can do this on another oven too without too much trouble, just
mounting the one actuator on the side of the oven with one lever arm.

But you're also right about the beeps. Eventually i'll have to dig into it to
either eliminate the beeps (as i've done on other ovens in the past) or reduce
the output level. On some ovens you can adjust the beep audio level, so i'd
like to do that too. With my other ovens i never needed to hear the beep, but
again this oven is different in that when it is done cooking stage 1 before it
goes to stage 2 it beeps two times. Without that beep i'd never know (without
sitting there and watching the time) when stage 1 is complete and i can turn
the food or whatever. So i'd have to be able to turn the beep on and off at
least for times when i really do need it.

So eventually i'll have to dig into the electronics, i was just hoping it could
be after the warranty was expired. But i'll have to rethink this now.

I have an actuator coming via mail now, which i will experiment a bit with and
then evaluate my position on "to dig or not to dig" (into the electronics) at
that time. Should be here in about 3 days or so.

 

[KeepItSimpleStupid]

Mr Al:

I feel your pain. I have been playing with a microwave with a stuck keypad too. Mechanically pressing all of the keys used to unstick it. Then a hair dryer used to unstick it, but you also had to reset it by pulling the plug. Now, I think both the magnetron and keypad are shot, so not worth repairing or made from unobtainium. This does have the moisture sensor and the probe and also a broken carrousel motor, so it's basically toast.

I do have another microwave/convection oven that I do want to fix, but the schematics are unobtainium. It was built in 1985 and it's huge. I re-designed the fan bearing support, but the magnatron, capacitor, an opto and a resistor is popped for sure. Just havn't gotton that Round-TUIT to be able to fix it.

What I want to do is to build a thermister type changer, because I used probe cooking a lot. I think there is a micro-switch that detects probe insertion and the probe is way different in terms of where the contacts are on the phone plug. I did pick up a jack, but havn't got a nut so far. I got a GE probe, but getting info from GE is IMPOSSIBLE!

I'd also like to be able to change the range of the probe so the micro can proof yeast. Yep, heating frozen things to room temperature would be nice.

The keypad likely plugs in, so MAYBE you can intercept the keypad ribbon. A FET optocoupler will work without the possibility of damage.

The interface of this micro/convection oven is very simple.

You have modes of: micro temp, micro convect time, convection temp, convection time, micro time. TEMP in these means probe cooking.

Then you have a button called "Oven Temp/preheat". It defaults to 350 F, but it's used to set the oven temp or probe temp and is optional.

Then a button called "micro control" which is basically duty cycle or power level in %.

And the standard digits.

And start and stop. Start is on the left.

so: Micro time, 1,0,0, Start is 1 minute of micro cooking.

and Micro time, 1,0,0, Micro control, 1,0, start would be 1 minute at 10%

Very convenient and intuitive.

 

[MrAl]

Hi,

Wow, well at some point you have to pull the plug
Once they get too bad it's time to get a new one, i found that out too. But there are advantages.

One of the advantages is that i think the new ones are more efficient. So for 1000 watts input you get more actual cooking power. I noticed this because with 1000 watts input power on this oven im getting faster cooking and water boiling than with the old one with the same input power. And that's measuring input power, not input current, so it's all real. The difference is quite dramatic too. With the old oven maybe 3:33 would boil water, but with this one it's about 2:10 or something like that (8 oz of water in a styrofoam cup as the test sample starting from room temperature). On high i could probably cut that down to 1:40 or thereabouts, but i dont use it on high because it pins my watt meter (chuckle).

So have you experimented with using a thermistor to monitor cooking temperature or whatever? That does sound interesting so i'd like to hear more about this, and how or if you had to do any shielding of the wiring or any other problems that might have come up (or didnt come up when it seems like they would have).

I got so used to microwave cooking that i hardly ever do it any other way. I hate washing dishes
I used a crock pot a few times but i find that i eat more when i use it so health wise it's not as good as the microwave i guess. I keep stuff in it all day so i end up eating all day long when i use the crocker.

 

[KeepItSimpleStupid]

So have you experimented with using a thermistor to monitor cooking temperature or whatever? That does sound interesting so i'd like to hear more about this, and how or if you had to do any shielding of the wiring or any other problems that might have come up (or didnt come up when it seems like they would have).

The micr/convec oven already supports a probe, BUT it has about 3/8 of an inch of an insulated area which, I think, the micoswich rides on, so a standard 1/4" phone plug won;t connect.

I did, at one time, measure the R-T of the probe and even came up with the coefficients. The probe is intermittant and mostly intermittent. I can't find my data and I havn't looked too hard. It's original data hand written. It did correlate to a "Standard curve".

I bought this probe for $40.00
www.amazon.com/GE-Part-Number-WB20X5050-PROBE/dp/B003AFMV6S

Note that microwave temp probes are EXPENSIVE. I can'r get GE to tell me what the R-T characteristics are. It's probably a "standard curve"

"Newer" probe microwaves (Scare now - look at the price of the probe). This micro/convec oven was like $900 when new in 1984.

So, the 1/4" jack is already there, but won;t support any conventional 1/4" phone plug of a connector. The jack uses ceramic to insulate the contacts and is also pricey $20-$30.

The low temp of the probe is 105 deg in 5 deg increments. 105 is too high for yeast. So, I was thinking of a toggle swich on the back that might subtract 100 degrees, so an input of 200 F would correspond to 100 F.

The probe type microwaves can detect an unconnected probe without the switch, so that should be possible. What I don't know is if there are any sequences it has to pass. Thermisters are or were difficult to linearize over a wide temperature range, so measuring room temp will be harder.

PROBABLY, the sensor in this microwave uses say a resistor to +5 and then the thermister to ground or vice versa, but I have a schematic to another microewave that has a 58K resistor in parallel with the probe (a method of linearizing) and it looks like they can "turn on the probe" when needed.

I do have a parts placement diagram and parts list with values for the errent microwave. Just no schematic except a wiring diagram of the major components. So probably, some reverse engineering will have to be done. Diss-assembly isn't going to be a lot of fun. Something wierd happened with the triac and control circuitry. The current limiting resistor to the external magnetron triac toasted, so the resistor, optocoupler and traic are probably toast. The optocoupler has a very strange configuration and pinout.

So in one case, I would have to detect a 10K to 50K @ 25 deg. C thermister and linearize it and would have to simulate a resistance on the other end and also detect the insertion of the probe.

So, i have some "really course" idea at this point. I need to:
1. Fix the micro/convection oven
2. Get or measure the R/T curve of the intended replacement thermister
3. measure or find the R/T curve of the original thermister.
4. Reverse engineer the thermister conditioning circuitry.

Cobble together a design and test outside of the microwave. Be careful not to exceed the power rails.

Probe micro/convection cooking worked very well for cooking a canned ham and just about any meat. It almost could be used to proof yeast when making bread. Warming liquids to the proper temperature and even defrosting.

I think the idea now is to use Moisture sensors to determine when the food has been reheated.

When making pizza in a gas oven, I use 1/2 the time on the rack and 1/2 the time on a cookie sheet and test the center of the pizza with an IR laser sighted thermometer.

For your project banging the keypad doesn't seem elegant.

There are benefits to inverter microwaves and I even have a paper where someone made a controller that worked on a conventional one.

Your project in some respects seems easier. I know the optocoupler thing works. I used a Vactrol VTL5C1 which has flying leads to interface to the unknown. Inside the unit, i added a current regulator and a diode so it could use a 3.6 V to 30 V control signal.



BTW: the keypad has CLEAR button and "micro Control" of 0 makes you able to use it as a timer. It's a 3 x 7 keypad.

 

[MrAl]

Hi again,

Sounds like quite a bit of work there. If you cant find the data for the thermistor you can use ice water, boiling water, and room temperature for the three test points and calculate the curve from that.

The probe idea sounds interesting but i guess i'd have to create one from scratch if i did it with mine. Im not sure if i'd want to buy one or not, depending on price.

Yes banging the keypad is not elegant and it abuses the POWER button, but that's just the way it goes. It was a dumb idea by the designers, period. And the more i think about it the more i see that it could be quite easy electrically, as i could get away with an interface for only that one button. But it will be fun trying to get the actuator to do the work at first, just to see how it goes. I'd like to see how fast i can get this particular actuator to go in an out too just as a test.

I'll also have to look for the connector once i open it up, the connector mentioned in a previous thread that would interface the keypad to the circuit board. In my last oven the keypad board was part of the circuit board however, so i'll have to check that out. I have to stop that beep anyway so sooner or later i have to open it up. That is legal for this oven because there is a user replaceable fuse inside.

My older oven (two ovens ago) used a relay instead of a triac. There would be a loud 'clang' when the power setting was set low and the oven would turn on as the relay would pull in. That was nuts when i used a lower power setting and it had to turn on and off several times throughout the cooking period. "Clang...wannngggg........clang..wannngggg..." over and over again

 

[KeepItSimpleStupid]

Sounds like quite a bit of work there. If you cant find the data for the thermistor you can use ice water, boiling water, and room temperature for the three test points and calculate the curve from that.

Thermisters are very non-linear. The hard part is keeping things mixed. So, ohmeter in one hand, and room temp to boiling water in the other. Add, mix, measure. Lower than room temp? Add ice water or ice cubes.

Wanna go higher than 212, use peanut oil.

Hard part, keeping the broken thermister working. Been there, done that, but I used a fluke solid state probe, which is currently broken, but I have some other ways of doing the temperature.

One probe isn't too bad, but emulating is much harder.

Have fun. Stay in touch. I have too many other projects to do.