To handle the timing aspect of the Cat Feeder Unreminder I am going to use the Analog (It is actually a Linear Technology part!) LTC2956 timer component.
This is a wake up timer IC that has built in push button support for reset. Perfect for this project.
The timer itself is configured with just a few resistors and can range from 250ms to 39 days. Setting it to the required 22 hours is a piece of cake for this timer.
In shutdown, it only pulls 800nA while active and 300nA in shutdown mode which is one of the lowest current use for a timer that I can find.
To set this up for the Cat Feeder Unreminder we need to stay in RUN MODE. The device can’t leave SHUTDOWN MODE without interacting with the PB switch and that would make for a hungry cat. The AWAKE STATE with EN=1 is when the device will tell us to feed the cat. SLEEP STATE with EN=0 is in standby counting till next feeding
ONMAX to GND
This sets the T ONMAX time to the max
T ONMAX will stay FALSE
We don’t want the device to auto fall into the SLEEP STATE
SLEEP to a switch
We will use this to send the device to the Sleep State
Set LONG to max time
Tie PB HIGH
Makes sure we don’t fall into SHUTDOWN MODE
Set Period/Range for timer to activate in 22hrs
This should allow the EN pin to go high every 22 hours and will reset low when the push button on PB is pressed.
After chatting with Stephen Kraig on the podcast about hacking tig foot control pedals, I wondered what made the Vulcan brand foot pedal different. Most tig welding machines have either a 5 pin or 6 pin connector. Both of these types of foot pedals work the same. There is a simple SPST switch and then a potentiometer for how far the pedal travels. Farther the pedal goes down, the more amperage the tig welder will output. This requires 5 wires. The 6 pin connectors just have an unused pin.
The Vulcan Tig Welder (from Harbor Freight) has a 9 pin connector for the foot pedal! What do these extra pins do? Or are they just unused as well?
To open the foot pedal you need to remove this the panel on the bottom side with 3 screws. You will see the foot pedal spring inside.
The spring needs to be pressed down and unhooked from the bottom side (side facing upwards during disassembly).
Then use a 10mm socket and wrench to remove the bolt used for the pivot and pull the case apart. There are two spacers on the bolt that locate the top part of the foot pedal with the bottom side along the pivot bolt.
Inside the pedal we have a microswitch and a single potentiometer. Only 5 wires. Why Vulcan specified a 9 pin connector is lost to me. Would be interesting to see if there are any connections to the other 4 “unused” pins on the welder side…
Potentiometer is a SCR30-103-7/8C by HungYun. 10K resistance.
Here is a pinout for the connector.
I can probably get any finger controlled switch to work with this welder if its a 10K potentiometer.
Just wrapped up the schematic side for the AEM10941 portion of the Cat Feeder Unreminder.
I broke out all the configuration pins to their own pullup / pulldown combos so I can experiment with voltage cutoffs for the super capacitors. Most likely leave most of configuration resistors unpopulated!
I found these really neat push button terminals that I am going to try to use for attaching the solar panel to the PCB. Part Number: TBL009-254-02GY-2GY
In trying to reduce the power consumption of the Cat Feeder Unreminder, I am going to explore using some really low power comparators to build the AC drive voltage I need to run the TN LCD segments.
The MAX9019 is a dual package comparator fits the bill in the power requirements.
In the switching frequencies I am using (50-100Hz) it should only need a supply current of ~1uA.
We setup the first comparator to generate a square wave and then the second comparator in the package as an inverter.
The design breadboarded up. It drives the screen!
This is well under the resolution of my Siglent SDM3045X. Will have to wait to get the right equipment to measure the actual current the circuit is drawing.
Here are the two output drive signals on the scope.
Next step to work on is the power retention system. The largest draw on the system right now is the leakage on the super capacitors! I found some super capacitors made by Eaton, HSL0814-3R8106-R, that specialize in having low leakage. Slightly higher ESR then some super capacitors but that isn’t that important for this project.
For the Cat Feeder Unreminder, I am going to pivot from using LEDs to indicate the “feeding” status and use a TN-Effect Display instead. These displays are much lower power then illuminated LEDs but they require slightly more circuitry to drive.
TN LCDs run off low AC voltage from around 3VAC to 6VAC depending on the screen. The one I picked, Varitronix‘s VI-422-DP-RC-S, operates over this range. They are not particular picky about the quality of AC voltage, just that it has zero DC offset. Driving the displays with square waves seems common. Anything north of 50Hz should do as well.
I used a 0.1uF capacitor and a 100K ohm resistor. Should get the oscillator to jiggle around 70Hz. Then I fed the output of the oscillator into another inverter on the CD4049. This gives me two square waves that are out of phase which will give us the AC voltage we need!
This works great! However the downside is this oscillator uses ~120uA at 3.3V without driving the Display. The display takes sub 5uA to drive so this is a big part of the power budget!
My record player (Sony PS FL-1) stopped opening and closing recently. Only reasonable thing to do is to open it up and see what went wrong. The PS FL-1 is a “tray” style record player. Kinda like a CD player.
Well that is the problem. Looks like the travel switches for the tray broke apart. The actuator on one of the switches was stuck closed which probably caused the tray to break the switches.
Alps switch manufacture, part number is 205-11 403R. Google doesn’t turn up anything and I couldn’t find anything on Alps website about this part.
Fortunately I found someone selling the entire switch board for $15 on ebay. Will update when I reassemble the turntable.
For the Cat Feeder Unreminder, I originally wanted to run LED indicators but I think it would be cool to run a reflective LCD display like on solar powered calculators.
For a display I was looking at Varitronix‘s VI-422-DP-RC-S. I was thinking I can hardwire the display to say FEED. What is weird with these displays is they run on alternating current (AC).
The solar power subsystem provides 3.3VDC which won’t work for activating the segments. Google searching around shows that these displays run on AC square waves. Initial thoughts are to make a push/pull transistor circuit that can drive and source high and the low side can sink giving us a 3.3V AC drive source.
But after chatting with some folks I think using a 4049 Hex Inverter like this will work great.
Received the 3AAEM10941CPCX10 evaluation kit for the AEM10941 solar harvesting chip today.
Its possible it won’t light up right away. It takes sometime for the super capacitor to charge up. You can charge up the super capacitor with a power supply set to 3.3V and current limited to around 10-20mA. Make sure to not reverse bias the charging!