Here an update.

Its now wired up completely. Havent yet tested with the PIC.
2 LEDs in series for each segment.

Maybe it wont be bright enough and will need higher driving voltage.

The 16F54 could be used for kindof a small clock but its not comfortable and not enough to use i2c realtime clock chip.

So I will use another PIC for that, then put all in a small plastic case, with a power supply PCB from a USB charger. Just fits into the battery compartment! Even a US plug so it goes straight into the wall outlet (needs an adapter tough).

 

I wrote some code yesterday + powered up the circuit.

The LEDs are very dim! I need higher voltage for the Anodes.
I wired the LEDs so they have common Anodes.

What is the easiest way to switch the high voltage rail?
I think I need a PNP/NPN for each anode + 2 resistors.

And I want to make a MC34063 stepup circuit, try with 6 volts and see, maybe increase. The voltage will drop through the LEDs the PIC just needs to sink the segments.

 

no replies OK.

the numbers dont look great too close to one another.
i think I should make another with diffused LEDs and different layout (more close to real 7 seg, with more spacing).

Today I'm worried about DHL, have to take photos, maybe no time to build the high voltage driving circuit.

Shipping from Hong Kong -> Ireland: SAT to MON 6AM
Clearing + domestic delivery: rest of the week.

 

Ah, you didn't post several pieces of information:
1) Where is your schematic?
2) Those LEDs look rectangular, but don't know what the color is supposed to be, nor the typical Vf @ recommended current.
It is very common for T3 and T5 water clear LEDs to be highly focused with a narrow viewing angle (15° to 25°) to spec high brightness. However, this makes such LEDs a very poor choice for something like a 7 segment display, where you would want a very wide viewing angle.
3) You don't mention or show what current limiting resistors (if any) you might be using. You need to use them, or other means of limiting the current through the LEDs.
4) Unless you are using high efficiency LEDs, you'll really need to use transistors or MOSFETs to source/sink the current to/from the LEDs, as you may need to sink or source in excess of 100mA on the common row/column, or more than 4x the capacity of the 16F52's I/O pins. There is an individual pin source/sink limit, and then there is a combined pin source/sink limit.

 

Ah, you didn't post several pieces of information:
1) Where is your schematic?
2) Those LEDs look rectangular, but don't know what the color is supposed to be, nor the typical Vf @ recommended current.
It is very common for T3 and T5 water clear LEDs to be highly focused with a narrow viewing angle (15° to 25°) to spec high brightness. However, this makes such LEDs a very poor choice for something like a 7 segment display, where you would want a very wide viewing angle.
3) You don't mention or show what current limiting resistors (if any) you might be using. You need to use them, or other means of limiting the current through the LEDs.
4) Unless you are using high efficiency LEDs, you'll really need to use transistors or MOSFETs to source/sink the current to/from the LEDs, as you may need to sink or source in excess of 100mA on the common row/column, or more than 4x the capacity of the 16F52's I/O pins. There is an individual pin source/sink limit, and then there is a combined pin source/sink limit.

thank you for your help + efforts to type the reply.

1) I thought its not needed. The LEDs are wired same as a multiplexed 7seg module, with common Anodes.

2 in series for each segment.

Anodes are on PORTB, the segments on PORTA and also on PORTB.

2) yes they are, green kinds, very bright. The problems are mostly the spacing between the numbers- its hard to distinguish them. As well the "1" is smaller, and the middle segment stands out.

Its a problem to a degree it couldnt be sold/marketed.

3) there are none. The LEDs are just very dim, I doubt they draw even a few mA all together. I found from research, a PIC16F54 can run a 2-digit 7seg (yellow) for about 3 days from a 3v button cell. Current cant be very high.

In fact I have run 4-digit modules at 5.0v for months. The PIC remains cold nothing burns out.

4) these are high efficiency. I doubt there's 100mA current or even 50mA.

I made a 5x5 matrix from SMD LEDs, powered with a MCP1640 from a single battery, and it was running for quite a long while. Actually MOSFETs were used for sinking (common cathode).

Then I made a large 768 LED matrix, all driven by PIC 16f59, 8 of them. One of them is erratic, it hangs after some hours. One time I scorched the circuit with too high voltage (made a mistake), had to replace 4 chips.

I have to get new 16F59 some time, and replace, then run really for a long time. They run at full 5 volts, since I use 6x multiplex.

By the way, powering the whole 768 LED matrix with a LM2576 module from 12V, it just heats a little. using a small 12v switching wall adapter, which also doesnt heat up. Total current seems to be some 100mA.

thats distributed over 8 PICs, also the 4 crystal osc modules need a lot of current.

5) My problem is how to switch the anodes from a higher voltage rail, i think i need 8 transistors as well 8 resistors. Cant drive PNP directly? Since I want full saturation, the base voltage will always be lower than the rail. Need to strap to the rail and short with a NPN. Or at least I think so.

6) I think the original PIC architecture was meant to be for LED calculators they were common at that time, so they somehow designed the circuits in a way they could drive LEDs directly. Later it is documentated as a current limit beyond which damage will occur. I rather think its a limit what a PIC actually will supply. Since the LEDs have a forward voltage, its never a total short to ground.

Especially at lower voltages LEDs will be dim, why add resistors? when it runs 3 days from a CR2032?

A long time test of the large matrix is something I have in mind, over a year or two at full voltage.

 

If you have a couple of spare pins on the PIC, and it has a comparator module or ADC, you might just use those to make your own boost converter.

 

If you have a couple of spare pins on the PIC, and it has a comparator module or ADC, you might just use those to make your own boost converter.

Hmm I didnt think about that. Yes the second controller does have pheripherals.

What is better ADC or comperator? How fast does the ADC have to be? I need something like a 50Khz or 100Khz square wave, then sample output how many times/second?

I could use a PIC with NCO, it has PWM too, but havent looked into that much. Slower frequency needs larger inductance (which has higher ESR).

i need 60mA or so, maybe 100, depends how bright.
Input is around 5v.

So its like I put out a square wave, then turn on/off depending on comperator result? Does it need hyeresis? Or is ADC better? I dont fully understand all of it.

but its a good idea maybe.