Resistors and 1-digit 7-segment displays

There seems to be different approaches on forums etc as to how many resistors are required for a 1-digit 7-segment display. Those suggesting 8 resistors for a 1-digit 7-segment display do so on the basis that the "brightness" of each LED will be constant irrespective of the number of LEDs HIGH for a displayed digit.
Others suggest that one resistor connected to the common cathode (or common anode as appropriate) is sufficient to ensure that current does not exceed 40 mA.
With a 220 ohm resistor fitted to the common cathode, the product of number of LEDs HIGH, the current and the voltage across an LED was measured on a 1-digit 7-segment display for 1 to 7 LEDS HIGH to estimate power used by the LEDS. The current decreased from 12 to 2mA, as expected with increasing number of HIGH LEDs, and estimated power was relatively constant, 22 to 25mW.
So other than the variation in LED brightness depending on the number of LEDs HIGH is there another significant reason for including 8 resistors with a 1-digit 7-segment display rather than fitting one resistor to the common cathode/anode ?

Thanks for any comments.

How many 'significant reasons' will you need? A lot of newbies will try to rationalize their laziness as such.
If drastically dimmed numbers doesn't matter, by all means use a single worst case resistor accommodating 8 LEDs worth of current.
Though the math on correctly figuring that value on such a quasi parallel diode scheme is something I'd like to see.

INTP:
How many 'significant reasons' will you need? A lot of newbies will try to rationalize their laziness as such.
If drastically dimmed numbers doesn't matter, by all means use a single worst case resistor accommodating 8 LEDs worth of current.
Though the math on correctly figuring that value on such a quasi parallel diode scheme is something I'd like to see.

What? I don't remember if I tried this but in theory 8 LEDs lit should give drop of brightness to 1/8 of singe LED. Due to logaritmic nature of senses I expect "visible" drop in brightness instead if "drastically dimmed".
Why do you think math for worst case should be difficult? OP used 220R and measured safe values of current (as expected). If he don't mind the drop of brightness I don't see any reason to use more resistors. Of course if LEDs in the display were poorly matched (or even different color) it may cause problems but they are clearly not...

NDC:
So other than the variation in LED brightness depending on the number of LEDs HIGH is there another significant reason for including 8 resistors with a 1-digit 7-segment display rather than fitting one resistor to the common cathode/anode ?

Thanks for any comments.

No, that's the reason - the display looks very amateur with different brightness digits (the ratio
in brightness from 1 to 8 is a factor of 3.5 which is quite obvious to the eye. If you don't care about
appearance then you can save 7 resistors per digit.

You can all the resistors and get even brightness if you use a constant current driver.

If you get a block of 4 digits in a single package, you use 8 resistors one for each segment. Then you need to multiplex the digit that is displayed, as you get 4 common pins.

There are plenty of examples of this including on my site - in my profile. Or Google 'driving multiple 7 segment display'.

Resistors are cheap and easily wired!

Hi,
If you are multiplexing, then one resistor per digit is all that is needed, as only one segment in that digit will be ON at anytime.
If you are not multiplexing, then one resistor per segment as more than one segment will be ON at anytime.

Tom.. :slight_smile:

TomGeorge:
Hi,
If you are multiplexing, then one resistor per digit is all that is needed, as only one segment in that digit will be ON at anytime.
If you are not multiplexing, then one resistor per segment as more than one segment will be ON at anytime.

Tom.. :slight_smile:

That's not how multiplexing is typically done - you activate all the relevent segments in a digit, then move to
the next digit. Yes you could it the other way (do all the a-segments, then all the b-segments, etc), but
I don't think I've ever seen that.

@MarkT
This is a single digit. So you can multiplex the segments.

Side question (without wanting to derail the thread)
The forward voltages of the leds can be slightly different (maybe not the case for a 7-segment display?). If one does not multiplex and one uses one common resistor, what happens due to the voltage difference?

sterretje:
@MarkT
This is a single digit. So you can multiplex the segments.

Side question (without wanting to derail the thread)
The forward voltages of the leds can be slightly different (maybe not the case for a 7-segment display?). If one does not multiplex and one uses one common resistor, what happens due to the voltage difference?

Usually, the LED forward drops are extremely close due to manufacturing with the same production batch of LEDs.

Also, keep in mind (not to encourage such cheapness) that the digit "1" has 2 segments out of a possible 7 (excluding the dot) so the difference in current per segment is 2/7 not 1/8. The minus sign will be way too bright, though...

As another alternative to the more conventional solution of using one resistor per segment, the OP could also use just one resistor on the common rail and pulse width modulation to correct the brightness depending on the number active segments for the digit (or symbol) displayed. This may require a little experimentation to get the appropriate pulse width. It would even be more versatile than multiplexing at a granularity of one segment, because even the decimal point, which usually requires less current that the remaining 7 segments, could be displayed at the desired brightness (if displayed alone). If the decimal point and a digit are to be displayed simultaneously, then it becomes more complicated to ensure everything has a consistent brightness.
Personally, I would use 8 resistors.

That's a pretty cool idea there 6v6! Hats off to you.