The old "Replace blue LEDs with red LEDs" 7 segment question (74HC595)

[trondl]

Hei from Norway!
I'm posting it here in this section as I'm a n00b in the technical sense when it comes to LEDs and current limiting.

I have an older (2011?) Mytek USB-DAC with seven blue 7 segment displays that after the years have gotten dimmer on the most used segments (48 kHz...), making it look like "burn in".
Is this evidence of early blue LEDs wearing out faster?
The unit has seen 24/7 on highest brightness.
It is dimmable, and on the lowest setting, the LEDs are somewhat similar.

I have gotten access to the front panel parts and layout document from the company, but not the schematics.
The unit is seeing too much usage to give it downtime for poking around yet.

What I can see is that the thru-hole 7 segment LEDs are Lite-On LTS-2801AB (3.8Vf @ 20mA If, common anode).
They seem to be controlled by a 74HC595A per segment and have a 220ohm current limiter per line.
Since the Vf is 3.8, I believe the 595 is running at 5Vcc.
Could this also be a symptom of failing Shift Registers? Doesn't quite seem logic to me (pun?).
I've heard rumours that a 74HC595 isn't ideal for driving LEDs, and can wear/burn out.
Also, how does dimming work on a 595?
PWM?

LTS-2801AB is obsolete, and I'd wish to replace them with the red LTS-2801AP (2.1Vf @ 20mA If) which is still somewhat available, or an equivalent.
I have a hunch that red LEDs last longer, and I prefer them for not being so intruding on the eyes (who doesn't?)

I've tried to find other alternatives, but the physical size seems to be a limiting factor: (h)10mm x (w)7.4mm x (d)6.1mm.
Is there a "standard" size for 7 segment displays?

Now, the main question of course, can red LEDs with ~2Vf last longer than 2 seconds in this circuit?
3.3V seems to be available on the PCB for driving an IR receiver.
Run the 595 on 3.3Vcc instead?

Thanks in advance for your time and input!

Mvh
Trond

[pcprogrammer]

If possible change the 220 Ohm current limiting resistors to 470 Ohm when replacing the 7 segment displays. This will limit the current to about 6mA instead of 13mA the 220 Ohm allows. The 13mA won't kill the red ones but they might be to bright with it and it can kill the 74HC595A because with all segments lit the current will be above the max Icc of 75mA given in the datasheet.

Simple rule is to take the LED forward voltage and subtract it from the supply voltage and divide that by the current limiting resistor to get the operational current. So with a Vf of 3.8V and a supply of 5V and a 220 Ohm resistor it is 5 - 3.8 = 1.2 / 220 = 0,005454 Ampere. Will be a bit less due to the output voltage of the 74HC595A not being equal to the supply voltage.

The 74HC595A has an output enable. Controlling this with a PWM signal can dim the display. The outputs will only supply current when enabled and on average lower the current through the LED thus dimming it.

[Kleinstein]

When driven from an HC595 chips the LEDs should get way less than 20 mA. With a 5 V supply one would have the 220 ohm resistor + some 50 ohm resistance for the HC595 and about 3 V for the LED. This would be some 5-7 mA per LED. At that low current the LEDs should not wear out.  Some 7 mA also also the maximum one should use with the 595 chips to keep the current for the supply pins reasonable.

For the reduced intensity it could still be LED (a bad series), the resistor (cheap carbon type), the 595 chip and also the supply that aged.
With RED leds one may want to increase the resistors to something like 390 ohm. If the displays are bright to start with one could also use less current / larger resistors. So it could be worth testing how much current is wanted.

Dimming is likely via PWM.

There are a few common sizes for the 7 seg displays. 10 mm hight is already on the small end, many are larger.

[SeanB]

Had a similar issue with a conversion from numitron filament displays to LED, where the resistor value selected, 82R, was making the display way too bright, likely because the original displays tried were old Monsanto ones, that needed 50mA to reach rated outout. The new ones were much better, but being loath to undo the cordwood board used as adaptor plate, and not wanting the pain of replacing 160 resistors as well, and soldering the 160 wires back to them, instead opted to use a diode in series with the common cathode of all the non multiplexed displays. Tried a single 1N5401, which made it dimmer, and a second one in series made a near perfect match with the original displays, though the diodes were running rather warm when you did a display test at full brightness, seeing as that made the current peak a little over 4A. 25mA per segment, at peak brightness, within the display capacity, as at 3 diodes they were too dull, and no room to install 6A diodes either, so 2 3A diodes, stuck to the mounting frame with lacing twine and conformal coat, would have to do. Bonus was that only had to do 4 units, getting all the non failed numitron displays as spares for the others.

[kevin.gibbs]

Measure the output voltage of the 74HC595. Compare the voltages of the bright and dim segments. If the voltages are approximately the same, then the chip is good.

[trondl]

Thanks for all the input!

So, there are 7 * 8 = 56 220ohm SMD resistors.
Changing them is doable, but time consuming.

After some more research, the LED package is a 10-DIP, so it narrows the options down to basically the same LTS series because they are positioned very close to each other (no space at all horizontally?), and pinouts are apparently not standardized in any way.

My theory about the dim LEDs is the fact that the only version in the LTS series being obsolete is the blue one and the others are still somewhat active.
Seems to me it was a too early adoption of blue situation.

If the red version is considered "safe" in the current configuration with 220ohm resistors, but the 595 may kneel, the 595 could be replaced with a stronger version that I recently heard about?
But I do then risk having too much current available?
Maybe changing the resistors to 380-470 is worth it after all.
Although it is dimmable, I'd rather be on the slightly dim than too bright side by default.

[xvr]

> the 595 could be replaced with a stronger version

They could (tpic6(a/b/c)595). But they has different package, not compatible with PCB for 74HC595. And different pinout.

[Kleinstein]

With many individual resistors the change in the resistors is a lot of work. An alternative could be dropping the supply voltage with a few diodes in the supply or a change in the 5 V regulator (e.g. to 3.3 or 3.6 V).

For a little more powerfull version, there may be 74AC595, though the higher speed also make the signal integrety a possible issue.

PWM typ dimming could be used for a intensity reduction but is not reducing the peak current. So it hardly solves the drive limitations or the 595.

[David Hess]

When I dimmed LEDs driven by 74HC595s, I controlled the LED's supply voltage which was separate from the logic supply.

I doubt the 74HC595s are failing.  The limit for them is the GND and VCC pin current which should not exceed the datasheet value.

[trondl]

It seems like dropping the supply to 3.3Vcc takes me exactly down to the ~5mA the blue LEDs operate under (Vf = 2.1).
Uncertain if the red LEDs are as bright though under those conditions.

I now consider this to be the most likely solution, as there is a 3.3V "supply" available on the PCB for the IR receiver.
Will need to investigate if it is powerful enough though, as I guess at least having 6-700mA available is a minimum for all the LEDs and logic (56 segments * 0.005mA = 280mA + logic and headroom).
There is also another 8 regular LEDs + 595 that handle the dBFS readings that still need their 5V.

I have some doubts though, as I can see on the layout that it appears that the IR supply is just a 3V3 Zener diode dropping the 5V line.
How much current can a small SMD Zener usually handle?
>1A?

[xvr]

> How much current can a small SMD Zener usually handle?

Milliampers, 10mA maximum (IMHO)

[pcprogrammer]

I have some doubts though, as I can see on the layout that it appears that the IR supply is just a 3V3 Zener diode dropping the 5V line.
How much current can a small SMD Zener usually handle?
>1A?

It is not so much as to what the zener can handle, it is about the series resistor dissipating the remainder of the voltage. The zener is biased to some current and can keep the voltage in check as long as the current drawn stays within limits. To be able to use the 3V3 made this way to drive up to 300mA the resistor needs to be 5.6 Ohms and the zener must be able to dissipate about 1 Watt of energy when no current is flowing into the displays.

Most likely it will not be designed for this, so not a good idea to start using this supply for driving the displays.

A series diode or two in the common connection of the 7 segment displays, like SeanB suggested, might be an easier solution.

[kjr18]

I have a question, every led module has it's own 595? Or there are two? Maybe your display is multiplexed? That way there would be less resistors to replace

[trondl]

One 595 per LED segment module (seven in total), and an additional one for the dBFS LEDs.

[kevin.gibbs]

You can turn down the brightness and not replace the resistors. However, you will reduce the adjustment range because the indicators cannot be turned on at full brightness. At night, it may be too bright (you can't reduce the brightness any further). So, I would change the resistors. Replacing seems like a lot of work, but it doesn't take much time to unsolder 50 resistors, clean all the pads, and solder the other resistors. You'll probably take longer to change the indicators than the resistors.