jonathanjo

@NickAlexeev Very curious about what unofficial secret mark you'd put on your diagrams!

@rich.g.williams I think it's less a case of being fussy and more a case of adhering to common practices so the output is a team output, regardless of which member did a particular part.

Depending on ALU_FUNC_SEL the output is either the XOR or a constant value 0x01. With bit 4 tied to positive, you get a constant 0x11 which would be an extremely constant to generate. (This is the only constant on the whole circuit.)

Bug: you were close, but no. The bug is bit 4 should be GND.

then you can do the following

You have to see that the yellow and orange symbols each go to only one place each, and that each yellow goes to the same chip as its matching orange. If you do,

Optimisation I think is hard to spot on mostly-symbols diagram:

@rich.g.williams David's bug is on this portion: see if you spot it! This is part of a very ordinary 8-bit ALU.

@rich.g.williams From your viewpoint, what do you think of David777's diagram?

I Paul Horowitz's bad diagram too.

Also on EE.SE as electronics.stackexchange.com/questions/28251/…

I'd have called them kind of "hybrid" style, certainly you can see elements of the "mostly-labels" style you're speaking about.

These are Bosch ECU from autodtc.net/ecu-item/database/ecu-circuit-diagram

Hi ... I've been looking for ECU schematics to see how they look from the point of view of this discussion. I've found a few, not sure how I'd categorise them.

(The question rests on how much current you can use per MCU pin.)

(If you put top-side switching on the CC segments you need one PNP per segment, and it would work. But why do that when you can do with 4 NPNs and bottom-side switching?)

(note I've flipped the CC unit veritcally so the LEDs point downwards)

If you look at the common-cathode and common-anode diagrams from your datasheet, the usual way to drive them is as follows:

Yes, the segments pins (14, 16 ...) are anode and thus are connected to positive. BUT because you can run a whole segment from an MPU pin, you tend to do it in groups.

Minimal means: an MCU and a single LED (via current limit resistor). Then add another LED (and resistor) Then add a pushbutton switch. Then work on 7-segments and transistor drive. Then RTC chip.

Just to repeat: my advice is get something minimal working, and grow it from there.

You have to read datasheets: this is from your linked one:

"would it make sense then to have mcu gpio pins connected to the base of a pnp resistor, with the emitter connected to the power, and the collector to the load to drive the segment pins of the display?" No, that wouldn't work. You chose a TDCR1060 display, which is common cathode. The usual way to drive this is with NPN transistors on the digit cathodes.

With your level of experience, my honest advice is you need to get yourself something minimal working. Embedded programming is quite different in many ways to java/python. Going through all the build steps to getting bare-metal MCU running is an achievement, even flashing a single LED.

DIP PIC plus breadboard is good way to proceed. I'm not familiar with the Pickit, but looks perfectly reasonable. You'll find you want to use some kind of in-circuit programming pretty soon, as unplugging chips from breadboards gets tedious, especially 40-pin ones (sure you don't want a smaller one?) You also have to decide your software programming environment: assembler, C, whatever, which is a whole other set of questions: do you have experience with anything like that?

Start with the bare minimum and work up from there. Until you can get a bare minimum working, all the desired features are just aspiration. Get something concrete lit up and you'll progress much better. The barest of minimums still has a lot to do: working with a bare MCU and getting even a single LED to blink is difficult at first. (How do you program it?)

@RIPJAWZ3200 you're going to have to divide your project down into pieces. I suggest starting with MCU + display (resistors for 1mA, no transistors) with standard power supply, perhaps just counting and making sure you can read the switches and so on. It is much easier to go from there onwards, and you add RTC, the power management and so on (consider, for example, LCD instead of LED). If you do it all at once it's very difficult.

@RIPJAWZ3200 Hope all that's helpful. Ask away if you want clarifications

Your question here: "Would changing the time or setting an alarm effectively be "programming" the microchip even if it is done through buttons connected to the mcu?" ... in normal terminology, "programming the MCU" almost always means changing its flash memory to give it a new sequence of instructions. So I'd say "wouldn't normally include changing the contents of RAM (current time) nor RAM/potentially EEPROM (alarm time).

Voltage: you must decide on your voltage range. If powered off batteries, the voltage will drop over time. Got to check the datasheets for all the devices, especially the RTC (will falling voltage mess with accuracy?)

Batteries: D batteries are perhaps as much as 20 Ah, which would give 200 hours run time which is only 8 days (some assumptions made in these calculationswhich are just for rough sense check). So: either a) use a plug-in power source, at least to start, or b) think hard about power requirements.

Regarding "is this circuit enough to get started with" ... it's close but not yet. You have to resolve a few things, especially about power. You're wanting to run off a battery ... this can be tricky if you have an LED display taking 20 mA/segment * typical 20 segments / 4 multiplex = 100 mA. Plus CPU etc.

Regarding your second diagram, honestly I thought it was worse! Here's a suggestion: redraw my model one in your CAD system (including editing the symbol for the 7seg display) ... and then modify it to your tastes (eg, the comments which thought "parallel zigzags" are sometimes better than folding as I recommended). Until you get good control of the CAD tools, you're working against yourself. I see you're using KiCAD, it's a perfectly good choice.

Hi @RIPJAWZ3200 glad you made it here

@RIPJAWZ3200 comments aren't best for this, come to the chat

Thanks ... no, I opened it and looked at the spreadsheet too, but honestly wasn't quite clear. And just to clarify, it's VE you're seeking?