What is the role of this Zener diode in a LED string? (Old traffic light restoration)

Hi all,

I'm working to bring an old LED traffic light back to life and make it fully functional again. But I have a question regarding the schematics because some things have to be modified and don't want to make a mistake.

This traffic light natively runs on 48VAC. Each individual aspect (red, yellow and green) has it's own light unit mounted on a PCB; consisting of 32 correspondingly colored LEDs* in a 4 times 8 configuration. Attached is a diagram of such single light unit:
(EDIT: click here for an updated version further below!)

Each string of 8 LEDs has a Zener diode wired anti-parallel (inverse parallel) across it. What I don't understand is the role the Zener plays here.

I came up with 3 possible explanations. The Zener provides:

  • Reverse polarity protection (but why, since a rectifier is already present in the circuit?);
  • Protection when 1 LED in the string fails (open circuit; for keeping the voltage or amps the same?);
  • Surge/over-voltage protection (so the LED string is never exposed to a higher voltage than the Zener is rated for).

Are any of these right, if any at all? Thanks a lot!

- Scott

*) No RGB stuff here; just single color LEDs (red, yellow, green) for each aspect.

Probably, but it depends on the actual Zener rating.

That is weird because the reverse breakdown voltage works similarly to the forward voltage of the LEDs. We don't seem to know the forward LED voltage or the Zener voltage...

My only guess is that it helps to divert some current during the voltage peak (about 66V).

I agree with @jremington @DVDdoug . The Zener diode maybe to prevent voltage spikes and protect the LEDs because Zeners start to conduct current in reverse direction when the reverse breakdown voltage is reached.

The 2nd explanation you provided also seems appropriate to me.

I had a suggestion: You require a Capacitor in parallel with a Full bridge rectifier else the output would not be good. Maybe your rectifier has an inbuilt capacitor, but if not I recommend using One large 1000uF cap.

Regards.

Probably this.
It's a rather crude way of powering leds - but it works.

Depends on how you define 'good'. If the leds need to be flicker-free in this application, you'd have a point. They don't, apparently, so the cap is completely unnecessary and only adds costs and bulk.

Thank you all @DVDdoug @jremington @Devanshu_Acharya for your thoughts! Much appreciated. I just saw that the pic I uploaded in my first post was not the final version. Pardon me.

Here is the circuit with all the data (the Zener circuit is unchanged):

The Zener voltage is in all cases just above the total voltage an 8 LED string draws. So I guess it at least protects them from voltage spikes (explanation 3).

But how about explanation 1 and 2? If true, does the Zener protect anything and how does it achieve that?

I think you got those covered already, more or less.

There you go; there's no reverse polarity risk due to the bridge rectifier. The zener would also be a very odd choice for protecting against this.

Again, it would be a very odd choice as there's nothing in that circuit that would require the current to remain the same if one led failed. Obviously the total light output would decrease, but there would be no cascaded failure beyond that since every string of diodes has its own current limiting resistors.

So reasons 1 & 2 are eliminated due to being ineffective and inexplicable engineering choices.
Reason 3 has something going for it, especially for protecting against very brief spikes. Apparently a zener will start to conduct before a led burns itself into oblivion.

#1 makes little sense.
#2 is possible. Ask the designer what they were thinking.

It is a clever way to get better utilization of the LEDs:
Without the Zener you must size the resistors to limit the LED current in the worst case conditions (tolerance of resistors, transformer ratio, mains voltage). With the Zener you can use sightly smaller resistors and in case there is too much current at the peak of the supply voltage (it is rectified AC!) the Zener will safely shunt it.
So it is not primarily for "surge protection" but chop off peaks of the supply voltage to get more even current.

Thank you all! Makes sense indeed.

Flickering is not an issue as the rectifier does a nice job. It's a DF06M.

I've called their engineering department but sadly no technical documentation is left. Basically I have to "guess" everything. I was told this particular LED PCB was only very briefly produced and the line of traffic lights they were used in was considered end-of-life already years ago. It is literally that when any traffic light still using this PCB breaks down, it is replaced entirely. No spares, no repairs.

Another thing that surprises me from my measurements is that the 32 yellow LEDs draw more than 2.5 times as much current (58.2mA, 2.37W) when compared to red and green ones (each ~22mA, 1W). I was expecting all 3 colors to be in the same ballpark as they are all of the same 5mm clear LED type in equal amount. Do these yellow LEDs simply have a much higher current or should I possibly be worried here?

They probably worked out the required current to have all three colors light up to the desired level. Efficiency of the leds vs sensitivity of the human eye and all that.

Re: flicker: I trust the rectifier does its job right, but that does mean the leds flicker at 120Hz (or 100Hz in Europe). For some applications it's not an issue. For traffic lights I would consider it undesirable as it can be very distracting in people's peripheral vision. You don't want that at a crossroads.

I would think there is more to the circuit. I'm assuming Red, Yellow & Green are not on at the same time. Unless there is a mechanical shutter (unlikely).

Perhaps the designer was trying to balance something. Normally one would need only one series resistor, this design has two. Maybe they weren't good designers and added it for some invalid scenario.

Regarding a capacitor. It would not be appropriate here. The addition of a capacitor will result in drawing "slugs" of current through the diode bridge and the "transformer". This current would severely reduce reliability.

No, it is nonsensical.

If a LED goes open circuit, no current flows in that chain, there is nothing to be "protected". The Zener will then draw all the current from the resistor, to no benefit.

Really bad idea! The capacitor will charge to the peak voltage of the AC and increase the current/ power draw beyond the intended design.

The suggestion that the Zener actually "clips" the AC peaks to operate the LEDs at a more consistent current while not making the load on the transformer occur only on the peaks - as a capacitor would do - is the most plausible.

1 Like
  1. The different voltages on the different color LEDs is to balance the brilliance of the LEDs. Different LEDs have different efficiencies and the eye has different sensitivities to different colors. I am not allowed to use 'different' any more.

  2. An electrolytic capacitor is not required. It would be an extra item to fail and reliability is #1 with street lights. When it failed it would explode. Not a big deal. Blowing a fuse or circuit breaker, a bad thing. If the breaker is not an issue the LED string would not be as bright as others. Uniformity is a necessary requirement.

  3. The Zenner diode is to compensate for variations in line voltage. Some step down transformers for the lights may receive 110 volts, others, 120 volts, and yet others 90 volts. Again uniform brilliance is the goal.

Reliability is #1, Uniformity is #2.

Ok, got it @anon35827816 , @Paul_B , @JohnRob & @blakebr . I just wrote about a capacitor because I have seen that being used with a Bridge rectifier in Textbooks, videos etc. But everything is not applicable everywhere :slight_smile: and anyways the Rectifier does it's job.

Oh, I forgot about that the LED were in series, my bad.

I thought you might like this link, it explains why a capacitor is such a bad idea on a bridge rectifier. Folks still do it but when you have a number of power supplies in you home and each draws current only at the peak of the mains voltage it distorts your main voltage and makes the electrical company unhappy. And it forces the bridge rectifier to handle larger peak currents than is really needed by the circuit.

I think you will find most switching regulators use a very small capacitor and "regulate out" the increased ripple.

1 Like

Cost and size of a large cap is another factor.

Not so very small. Generally something like 150uF to 470uF on the primary / high voltage side depending on required output power. Generally far bigger on the secondary side. The "filtering out" isn't so much what limits the primary side capacitor due to the simple fact that it acts like a buffer for the actual SMPS to draw Joules from. Hence, there's a minimum size of this capacitor for a given output power for the SMPS to be able to maintain regulation in the first place. This can be somewhat compensated for by the filter caps on the secondary side; you often see 1000uF and beyond for low voltage PSU's in this place.
I'm sure @chrisknightley could tell you lots of relevant stuff about this. Perhaps he wants to chime in on this.

@JohnRob Are you trying to explain the same as the 2nd answer in this post? If yes, then thanks didn't knew about that...