Go Down

Topic: Single digit, Seven segment Numitron program and pcb design needed. (Read 17021 times) previous topic - next topic


wvmarle

Just looking at the MAX1724 data sheet.

Suggested inductor size is 10 µH, but they also mention that bigger is better - with a limit depending on the current drawn. For an output current of 150 mA (which for this application should be more than enough) the formulas on page 10 suggest we could go to about 80 µH. The problem is however that such inductors become physically larger. So on the schematic I kept it at 33 µH.

Another thing: it's sometimes convenient to get an idea on how much power the battery has left in it, so I devised a simple indicator by adding the R9/R10 voltage divider and multiplexing it with SW1. This also means the 10k pull-up from switch SW1 is gone. Large resistor values in the voltage divider limit current loss (1.25MΩ leaks 2-4 µA, which should not have a significant effect on battery life). Those adafruit batteries come with protection circuitry built in so they will simply cut off when the voltage goes too low. C8 is to stablilise the voltage for the ADC (and it conveniently debounces the button in the process).

The idea: press SW2 long (1 second) and the display shows the battery level: one, two or three horizontal bars (roughly <30%, 30-70%, >70%) based on the actual battery voltage level.

I realised I had the LED wired wrong in the schematic, corrected this as I changed it to two separate LEDs.

SW2 can't have this cap for debouncing as it'll interfere with the MISO function of the same pin.
Quality of answers is related to the quality of questions. Good questions will get good answers. Useless answers are a sign of a poor question.

wvmarle

It still fits on the PCB. Didn't try to do any trace routing, though, that's another major challenge :-)
Quality of answers is related to the quality of questions. Good questions will get good answers. Useless answers are a sign of a poor question.

Daddyfredregill

Very nice. So the unite will still only have two switches? Where does that SWITCH label at the top of the schematic connect to? Or does that just imply a function of one of the buttons?

Also in the RTC is says diodes in series with the vbackup pin could cause improper results. So is the diode in your circuit going to be a problem?

Also having the desplay read the voltage would be a really neat feature.

wvmarle

That connects to the label SWITCH at the microprocessor.

For that diode: it depends on where it is. This one is between the supercap and the rest of the circuit, and is there to allow charging of the cap (with the 47Ω resistor to limit the peak charging current to 10 mA) but to stop the supercap from powering the processor and other components (or it would be drained very fast the moment the battery shuts off).

A diode in between the supercap and the RTC would not be such a good idea.

It is of course possible to have the actual voltage shown on the display but as voltage/charge is not exactly a linear and direct relationship it's a mere indication of available battery levels. A three-bar indicator is about the best you can get.
Quality of answers is related to the quality of questions. Good questions will get good answers. Useless answers are a sign of a poor question.

Daddyfredregill

Thanks for clearing that up. The three bar idea was what i was talking about, sounds great.

wvmarle

So one button for time display; one button for date display; and one of those long press for battery indicator. That's about all the functionality that can be squeezed out of this hardware.
Quality of answers is related to the quality of questions. Good questions will get good answers. Useless answers are a sign of a poor question.

Daddyfredregill


wvmarle

Good one.
Never thought about that part, would be a neat feature I guess.

We'd have a long press available on the other button to go into time set mode. Then one button to increment a number, rolling over at 9 or whatever is sensible for that field, and the other to go to the next field.

So when the time is say 2008-01-14 13:23, the first 20 can be skipped of course (whoever is around to use the thing past 2100 will have to amend the software), then go through it digit by digit. That's not too hard to implement.
Quality of answers is related to the quality of questions. Good questions will get good answers. Useless answers are a sign of a poor question.


wvmarle

Quality of answers is related to the quality of questions. Good questions will get good answers. Useless answers are a sign of a poor question.

Daddyfredregill

Sorry feel really dumb. I dont know what a P/A field is.

wvmarle

Quality of answers is related to the quality of questions. Good questions will get good answers. Useless answers are a sign of a poor question.

Daddyfredregill

So basically i need to make the current build to test and see if everything works.

wvmarle

Indeed. Then at least the circuit is confirmed good, and it's much easier to make changes if needed. It also provides a good test bed for the software.
Quality of answers is related to the quality of questions. Good questions will get good answers. Useless answers are a sign of a poor question.

Go Up