Two MAX7219 with common Iset resistor

I've assembled two MAX7219 drivers and used just one Rset resistor for both.
During prototyping I forgot about a necessity to use separate resistors for each MAX7219.

It works perfectly. But I've got some doubts about any side effects in this schematic.
Does anyone know what is under the hood of Iset pin of MAX7219? The only thing I've found so far there is something named "Segment Current Reference" module inside the driver.

Well, it almost certainly does not work "perfectly".

Given that the ICs are closely matched, the two MAX7219s are sharing the current from the Rset resistor, so it would be as if each has a resistor of twice that value and is driving half the nominal current to the displays. But it probably is not precisely shared, so one is likely to be driving more current than the other.

What is interesting is that given the non-linear response of the eye to light intensity, you will not have noticed, and if you correct the mistake, you will notice only a very modest apparent increase in brightness of the displays.

Thank you for the quick response Paul,

Just as an additional info:
Both 7-segment displays (5 digits each) connected to the ICs are looking equally bright.
When I disconnect 7-segment display from one IC, another one doesn't change brightness visually.

I would agree that Rset value probably should be lower in my case to get same result as if I had two Rset resistors.

Whether the display itself is connected or not is entirely irrelevant. It is a matter of whether the resistor is connected.

But not lower than the minimum specified 10k value.

What is the difficulty in adding the additional resistor?

Just because it is time consuming. Nothing else.

There are 4 PCB with 2 ICs each in my project. And all of them are already assembled and interconnected with the main board. New PCB order takes about 2 month in my location. Reassembling takes a lot too, because I dont have a lot of time for this hobby, unfortunately.

So at the moment I'd prefer to look and use the final result. Probably, I redisign it in the future.

Okay, well I've seen a lot of PCB "patches" in my day. It's definitely do-able. I made one PCB in quantity 50, got a resistor in the wrong place and had to use an 1/8W through hole part to fix it because the leads helped me jump a gap. :slight_smile: But all the boards were easily fixed that way.

Because the current set input is probably low impedance, it's a bad idea to connect them. It would only take a small offset to upset the balance. The effect is to make the current working state un-dependable. Changes in temperature, supply voltage, aging, and many other things could affect it in the future.

Not to rub salt in your wounds, but did not a little voice in your head go, "hmmm, is this really cool?" when you designed it that way? Since it's not covered in the data sheet or anywhere else?

Thank you for replies, guys.

BTW, what do you think about using one capacitor for power decoupling for two MAX7219?
They are located close to each other with the capacitor just in the middle.

As always, there should be a 100 nF directly across Vcc and ground of each chip. Depending on your PCB design, this could be at one end of the MAX7219 with the two connections to Vcc and ground pins running underneath the chip or the capacitor itself may even be directly under the chip if it is a DIP (especially if socketed). It could be on the underside of the double-sided PCB.

The 10 (or 22) µF (page 10) capacitor may reasonably be shared between two ICs.

This is the recommended schematic (Fig-1) for 2xMAX7219 chips. If the chips are DIP packages, then just solder a 10 k resistor across Pin-18 and Pin-19.

max7219-16dig
Figure-1:

Does it matter on PCB layout if VCC trace first goes to 0.1mkF capacitor and then goes to pin19?
Please see the attached pics. Which one is more correct in sense of power decoupling?


Neither, but the ground connection should be the actual ground connection to the chip, pin 4 and pin 9, not some other ground connection but preferably a ground "fill".

When we talk about power trace lengths and routing, people seem to forget that this is even more important when it relates to the ground connections.

Paul,
my question is about connection between VCC, capacitor and pin 19. Are these layouts the same?
GND here is just for illustration purposes. On actual PCB GND net is connected with a ground "fill". It looks as on this picture
image
So the question is: is it important to connect VCC to the capacitor first and only then to the pin19?

Given that the capacitor connection is close to the Vcc connection on the chip, it will make no difference.

In particular, the distance between the capacitor and chip Vcc is very short compared to the connection of Vcc to anything else.

And I just had to mention that the connections to ground are if anything, more important.

Got it. Thank you!

In your post #13 the path for the ground between C2 and the ground pins 4 and 9 is very long because the ground fill is broken up with multiple tracks cutting right across it. I can't even see how the ground gets from C2 to those pins, the actual route must be outside the image. @Paul_B has emphasised the need for short tracks, unless there is a short track for ground I cannot see on that image then you need to re-think the whole ground plane.

In practice, the best way to make a ground plane on a 2 layer board, is to put a ground plane on both sides and bridge them with sufficient vias in places where it gets broken up by conductor runs.