Caution Light Panel PCB (A-10) - What resistor to use?

[KLUTCH]

Hello!

I've decided to make a caution light panel for the A-10, I want to build a PCB (because I never have before and it looks fun).

Using the worthog project's files I have the LED placement for each diode and also the holes that hold the panel standoffs.

I've seen craigs sim put and others mention the MAX7219 and luckily I have a matrix module already.

The matrix I want to build is 6x8 (6 cols/8 rows).

In testing a random 3mm green LED it's bright on it's own but dim when in series.

My question to this lovely forum, is what resistor should I use in the RSET module in the MAX7219?

Note: I know very very little about voltage/ohms etc (also being colorblind resistors are a nightmare anyway).

Here's a few shots of what I've learned from using easyeda over the past few days...

 

Any help much appreciated!

[KLUTCH]

In testing this on a breadboard, I think I made a mistake, the LEDs are in series, they are much brighter in parallel. Unfortunately I've already ordered 5 PCBs .

I've also tried a 20k resistor and a 5k resistor and the brightness all looks the same to me. Does anyone have any ideas on what I could do better?

Edited November 15, 2021 by KLUTCH

[Vinc_Vega]

If you power a LED by 5Volts, resistors between 220 and 330 Ohm are sufficient (not k-Ohms).

 

Regards, Vinc

[bojack]

Problem with the MAX7219 is that each the LED is only on for 1/8th of the time so you are never going to get the same brightness as if you drove them directly with 5V and then putting them in series makes it worse.

Do you have the brightness set all the way up in the sketch that controls the MAX7219?

What @Vinc_Vega says above is normally correct, but the MAX7219 uses a resistor (Rset) to set the current it provides per LED (which saves you from having a resistor per LED). For two green LEDs in series you are looking at 4 - 4.4V voltage drop across the two LEDS which isn't in the table for Rset values provided in the data sheet - but my guess would be that 10K should be the lowest you can go (the data sheet says 9.53kΩ for 40mA), I'm not sure that driving that voltage is even supported?

But your matrix looks sound - you may be able to drive it a different way?

[Vinc_Vega]

You are right, I didn't see the resistor for the Max7219 was questioned. I'm with you by choosing the 10k.

 

Regards, Vinc

[Rump0l3]

On 11/17/2021 at 4:04 AM, bojack said:

Problem with the MAX7219 is that each the LED is only on for 1/8th of the time so you are never going to get the same brightness as if you drove them directly with 5V and then putting them in series makes it worse.

Amen to that.  After attempting to build one of these PCBs using a MAX7219 I gave up, because at best each LED is only going to be illuminated 16.7% of the time (i.e. 1/6^th, assuming that you configure the board to use six of the available eight digits and configure the scan limit accordingly when you initialise the chip) and it was just too dim.

I ended up redesigning the board to use three TCL59282 chips (see photo).  Using a separate power supply to drive the LEDs, each "light" on the panel consists of three surface-mount LEDs connected in series.  This gives more even illumination, and also means that if anything it's too bright, but it's much easier to make it dimmer in software than to deal with hardware which isn't bright enough.  I also had to write my own Arduino code to drive the panel, but it wasn't as hard as I was expecting (having a DSO which can decode SPI helped).

Also, don't worry too much about having already ordered five of these PCBs.  It's a learning curve and mistakes are inevitable.  Sometimes they'll be salvageable by cutting traces on the PCB and/or running bodge wires, sometimes they won't be.  It took me three revisions of the PCB for my caution light panel before I was happy with the result.

If you decide you want to go down the TCL59282 route, PM me and I can give you some pointers?