Understanding real life Zener Solar Regulator performance

[KevWal]

Hi

I have an array of small solar panels - 6 x 0.5v, 0.12w panels, connected in series, and then connected to a 120 ohm load resistor for testing.  I get between 0v and 4v across the load resister depending on the brightness of the sun - more commonly 2v to 3v.

If I want to limit the maximum voltage across my load to 2v, I believe I can use a 2v zener in parallel with the load, with the solid bar of the zener connected to the + side of the load.  Below 2v the zener should take no current, above 2v the zener should take whatever current is required to pull the voltage down to 2v (until the zener burns out).

I have implemented this, with a volt meter across the load, and a mA meter in series with just the zener.

Below 2v it works fine, I get 0mA through the zener.

Above 2v the zener takes around 50mA of load, which pulls the solar panel down from say 3v to 2.5v, but no more.  If I put 2 or 3 x 2v zeners in parallel the situation improves slightly, taking 60ma or 70ma through the zeners and reducing the voltage by another 0.1v, but still not down to 2v.

It seems that the zener has a internal resistance that is stopping it working fully?

Am I doing something wrong?  Whilst not text book, is this expected behaviour in the real world?

Thanks very much
Kevin

[Siwastaja]

You are expecting the zener current-vs.-voltage curve to be a step function: zero current below 2.0V, infinite current (as much as the supply can give to keep it exactly at 2.0V) above 2.0V.

In reality, unfortunately, the zeners are far from step functions. It's more like an exponential function. Look at the datasheet, the curve is printed there!

For that matter, any proper textbook should really explain this.

Additionally, what the "typical" datasheet curve doesn't show you, is:
1) Temperature variation: the voltage point where it conducts a certain current x is different at different temperatures!
2) unit-to-unit variation: similary, the voltage point where it conducts a certtain current, is different when you take the next diode of the batch; and even more different when you order a next batch.

For this reason, simple zener regulator is mostly useless for solar panel shunting, to prevent overcharging a battery, for example.

With a lot of margin and carefully selected, it may work but it would still end up discharging the battery each time there is too much shade.

The only way around this, unfortunately, is a bit more complex active circuit; for example a voltage reference IC like TL431 which can actually sink quite a lot of current (100mA).

[KevWal]

A diagram to help...

[KevWal]

Thank you Siwastaja, yes I see this now, for example in the attached, for a 3.6v Zener, the difference between 0ma and 20ma flowing is nearly 2 volts.  It seems that the % difference in voltage between not conducting and fully conducting is significant for small value zeners - nearly 50% at 2v, where as down to 2 or 3% by the time you get to 30volts.

[TimFox]

Lower voltage diodes, such as 3.3 or 3.9 V, are actual "Zener" diodes, and have relatively "soft" turn-on curves.
Higher voltage devices, such as 8.2 or 10 V, are "avalanche" diodes, and turn on much more sharply.
See https://en.wikipedia.org/wiki/Zener_diode  for a discussion of the diode behavior.

[David Hess]

It seems that the zener has a internal resistance that is stopping it working fully?

That is right, and low voltage zener diodes have a "softer" knee in their current versus voltage curve than higher voltage parts.

Am I doing something wrong?  Whilst not text book, is this expected behaviour in the real world?

The alternative to a shunt zener is a shunt voltage regulator which can have an arbitrary sharp cutoff although implementing one which operates at only 2 volts is somewhat of a challenge.  A TLVH431 shunt reference can operate down to 1.25 volts and sink 80 milliamps so might be suitable.  An LT1635 could be used to make a power shunt regulator.

[KevWal]

Thanks, In my application I actually need a cut off about 3.5v, so that I don't damage 3.3v components, as in really full sun the panels produce up to 4v.

I just used the 2v zeners for testing, as it is very hard to get really full sun on an average day in the UK

The rated voltage seems to be pretty near the start of the knee, ie a 3v zener starts conducting at 2.9v say, but isn't fully conducting till over 4v.  So I think I can work with something like that...

Thanks very much all
Cheers
Kev

[BrianHG]

If you know what you are doing, and a series 1.5 ohm resistance from the solar panel to you device being powered is acceptable, IE you device only draws 100-200ma, maybe you can make an actual near 0v dropout regulator using a Lateral N-Channel Depletion-Mode MOSFET and a normal silicon diode or 2 between Gate and GND with a 10 megaohm pullup.  It would be wired in a 'Source Follower' configuration where the drain is connected to the solar panel + and the output would be the source which should be slightly higher in voltage than the voltage at the gate input instead of a normal N-Channel enhancement mode mosfet where the source output would normally have a voltage drop due to the turn on Vgs.  To change the output voltage, wire the gate to the GND with normal silicon diodes like 1N914 and place a 10meg resistor at the gate tied to the solar panel V+.  The more diodes, the higher the output voltage.  No diodes, and the output voltage will be the minimum possible for that mosfet.  (Dont need a pull-up in this case...)

Remember, this output will be regulated to a maximum without shorting your solar panels making it safe to add more panels in parallel or stronger sunlight without burning up any zeners.  Vout will equal Vin up until the voltage where the regulation kicks in.  Normal silicon diodes with that 10meg pullup will offer 0.5v step increases as you add more in series.  Old fashioned red and green LEDs also offer a larger ~1.5v to ~2.5v step in a single device, though with the 10meg pullup, the LED would never illuminate unless the mosfet blows.

Example device:
https://www.digikey.com/product-detail/en/microchip-technology/LND01K1-G/LND01K1-GCT-ND/4918738
https://uk.rs-online.com/web/p/mosfets/9125259/?cm_mmc=aff-_-uk-_-findchip-_-9125259&utm_medium=buyNow&utm_source=supplyFrame
If you need a lower series resistance, these guys can be placed on top of each other in parallel.

Only bother with this test if you are already making an order from any supplier with stock.

[Damianos]

In these low voltage levels, the burden voltage of the mA-meter has an important effect. Try again without this meter. The result will be a little better (I guess).