Go Down

Topic: 2.3" 16 segment common anode LEDs, HT16K33 and ULN2803 (Read 1 time) previous topic - next topic

MitchSF

I can go back to that but doesn't that invert the signal, requiring a 74HC04 to invert back?

Grumpy_Mike

#76
Aug 27, 2019, 10:49 pm Last Edit: Aug 27, 2019, 10:50 pm by Grumpy_Mike
If it does then simply invert the signal in software.
But transistor on means LEDs get power, do you have it the other way round?

PaulRB

I feel like this thread has gone in a circle and we are back to the problems and questions from pages 1 & 2.

I don't think the signals can be inverted in software. The multiplexing is not software driven. The OP wants to continue to use the ht16k33 to drive the multiplexing, as he did previously with some smaller common cathode displays. The larger displays he wants to use now are common anode, so I think inverting drivers are needed on both the high (anode) side and the low (cathode) side. The ULN chips provide the inverting low side drivers, but designing inverting high side drivers is proving difficult.


PaulRB

What about udn2981 to drive the common anodes? Unfortunately these are not inverting, so 74hc04 will be needed to invert the COM signals from the ht16k33 to feed the inputs of the udn. Pull-ups to 5V on the 74hc04 inputs to prevent them floating when the COM pins are not pulling them low.

MitchSF

I switched to common cathode displays. Using a ULN2803 with a 2k resistor on each input to +5v, and a TBD62783 on the anodes with 2k resistors to ground on each input, everything appears to work fine.

I'm not sure if the value of the resistor is optimal. It was a guess. Is there a way to calculate the correct value?

For the segment series resistors, I believe the voltage drop across both the ULN2803 and TBD62783 together is around 3v. Two LEDs in series drops 6v, so 12 - 6 - 3 = 3v, 3/.04 = 75 ohms. Is that correct?

PaulRB

That sounds like a good estimate. You could temporarily try something higher and very safe, like 330R or 510R, and freeze the multiplexing so that one digit is permanently lit and you can measure the actual voltage drop across the series resistors.

Then you can re-calculate the optimal resistor value to achieve the 40mA when multiplexing is enabled.

MitchSF


PaulRB

Sorry, now I'm forgetting! You can't freeze the multiplexing because the ht16k33 is controlling it. You will need to disconnect that and connect one of the ULN2803 inputs to 5V and one of the TBD62783 inputs to 5V to light one segment, then measure the voltage drop across its series resistor.

MitchSF

I'll do that. Thanks.

Is 2k an appropriate value for the pullup/down? I don't see a way to calculate the value so maybe the best way is to choose the highest value that will allow the circuit to work. With no resistor it does not work at all.

PaulRB

Values of pull-up/down resistors are not usually critical. 10K or 4K7 are the most common. 2K2 is also fine, especially with higher switching frequencies or longer line lengths. Trying to find the highest value that works reliably is often not worth the effort unless you are designing a battery powered circuit.

MitchSF

Thanks for all the help. I"m just waiting for the boards now.

Go Up