Zener and Comparator Problem

Hi all

I am working on a project and come up against a weid problem (which I have now engineered around)

My circuit runs off a 12 volt battery and I wanted some means of a battery low indicator (battery low goes to a uhf transmitter)

Ah ha, Easy I thought.

The circuit attached below is the part of my circuit that played up.

When I wired this up on strip board it seemed to go fine. When the pot divider network outputed below the zener reference the comparator outputed false and the 10k pullup resistor caused an output true and vise versa.
The only odd thing I noticed is that close to setpoint of the zener the output fluttered, the closer to setpoint the faster the flutter until at setpoint the output went solid (I had a led on the output)

So I set the pot up so the output was only just off and I disconnected the circuit and connected a 9v battery expecting it to output as a low battery..... but no go.

The pot votage divider network did its job but the zener voltage dropped off ( the 6 volt zener was doing 2.5 volts)

Can anybody suggest what might cause these sorts of weird problems. I have little experience with comparators and every time I use them there seems to be problems.

In the end I engineered aroung the problem by using a regulted 5volt rail in another part of the circuit as a reference and the circuit worked fine.

Thanks

Bruce

What kind of comparitor component? Some op amps have problems if their inputs get too close to either power supply voltage.

Hi

I am using half of a 393N.

The zener is 6 volts and the the pot divider network is set in this vicinity and the 393's supply is 12v so operating miles off both 0v and vcc.

I dont think the problem is with the comparator setup itself but with the zener or its interaction with the comparator.

The problem I described in my original post went away when I used a regulated 5volt supply as the reference instead of the zener.
Originally I thought that the output 'fluttering' at around about vref was due to no hysteresis but interestingly this has gone away with the reg 5 volt vfef.

Thanks

Bruce

Hi Bruce,

Zener diodes require a certain range of current flowing through them to operate in their linear region. If the current is quite low, the voltage across the Zener can vary quite a bit from what's specified.

You have a 6v Zener in series with a 10k resistor, and a 12v battery. If the Zener is a 1N4627, it's actually rated 6.2v @ 250uA (microamperes). So, when your battery is fully charged (around 12.8v @25°C), the current through the resistor (and Zener) is (12.8v-6.2v) / 10k Ohms, or 6.6/10000 = 660uA; plenty for regulation. However, if the battery voltage drops to your test case, the 9v battery (which probably measures more like 8.5v under load), then you had (8.5-6.2)/10k = 2.3/10k = 230uA; not quite sufficient for achieving regulation, particularly if the comparator input is not high impedance.

If you're using a 1N5233, it's test current is 20mA! At 660uA, you wouldn't even be getting near it's rated range.

The other problem you're experiencing (fluttering; oscillations) is typical of a comparator that does not have hysteresis. Providing some feedback via a resistor from the output to the noninverting input will provide hysteresis.

But to simplify that, swap places with the Zener and the 10k resistor.
Then swap your connections for the inverting and noninverting inputs.
Then use a feedback resistor from the output to the noninverting input.

If you use a lower voltage Zener for a reference, it will maintain a more constant reference voltage over the range of input voltages.

Thanks for that Sgt Wookie.

Why would the lack of hysteresis be more apparent with a Zener than with a regulated supply(7805) Vref?

The Zener I was using was a QEOA02-06 (thats what it said on the packet anyway) supposedly 500mw 6 volts. (cant seem to find any data sheets for it though)

Load for the circuits in question is minimal (I measured 30mA max)

Thanks everyone for your help. As I said in my first post I have engineered around the problems but I am curious to find out about what has caused the problems so I dont encounter them again in the future.

Time for me to go away and read up some more on op amp impedance and minimum current for regulation on zeners

Thanks everybody for your help

Don't go away TOO quickly!

I just got done throwing together a schematic of the mods I was suggesting - see the attached.

R1 is really too large a value for the 3.3v Zener I used; as it is, the most current the Zener will see is around 1mA. As the battery voltage drops (simulated by V1 on the left) the Zener voltage also drops; but in this simulation only slightly. A "real" Zener of that type would drop off more quickly. There are "low level" Zeners available, like the 1N4684 3.3v; it'll maintain regulation within 95% even at 10uA to 100uA, and can handle up to 80mA current.

Note R4; it's for hysteresis. If you wanted a real definite on/off switch, you might drop it down to 330k, or even lower.

How do you know the minimum current for a zener then? something you get off a datasheet?

Maketronic said:

How do you know the minimum current for a Zener then? something you get off a datasheet?

Click to expand...

Yes, you should look at the datasheet.

If you don't have a datasheet (and can't find one), then you'll need to test it to find out where it's "knee" starts.

See the attached simulation of supplying current through a 4.7v Zener diode. Note that when the current (pink line) is very low, the voltage across the Zener diode (Vd1, green) is also very low. Notice that as current increases, the voltage drop across the Zener rapidly increases up to a point, then levels off. The curve almost resembles the knee of someone sitting in a chair. If you keep the minimum current above the "knee", you'll be operating in the Zener's most linear region.

Looking at your above simulation it looks likely that I have been operating my zener well below the 'knee' (total current draw is only 20 - 30 mA) and thus in the non linear region.

I will have to set up a couple of meters and do some experiments and have a play.

Thanks for all your help, hopefully I will be able to help out other forum members one day.

Regards

Bruce