Voltage comparator?

lairdt · 2007-10-21 00:10

Anybody have a very simple voltage comparator circuit?

I need to know if an input goes below 0.40 volts (3.3v source) and I don't have much board space left.

Thanks

viskr · 2007-10-21 00:26

If you really need an accurate comparison to 0.4V then an LPV7215 or similar which is available in an SC-70 package, that and 2 resistors and you're there.

If going below 0.6V is close enough, then any NPN transistor with that level tied to the base thru a 10K resistor and a 10K pullup on the collector as an output (emitter to ground)

Mike Green · 2007-10-21 00:29

Basically, you need a voltage comparator like the LM393. This comes in an 8-pin DIP (and other packages) and will run off 3.3V. You'll need a pullup (like 3.3K) and a voltage reference which could be just a simple voltage divider running off the 3.3V supply. Whenever the input voltage goes above the reference voltage, the comparator will turn on and the output will go to zero. There are lots of other comparators, but this one is cheap and widely available.

Unfortunately, the voltage is too low for the Propeller to directly detect it.

lairdt · 2007-10-21 01:13

Thanks·for both replies.· I'll give it a try with the 393's... don't have any right now, but I have some 339's that I can test with until Mouser arrives. [noparse]:)[/noparse]

Was hoping there was a simpler way to get results other than adding another chip, oh well.

deSilva · 2007-10-21 10:24

I am impressed that they are assumed to work from 0V on from a single supply voltage.... So 3V3 allows an operating range from 0V to 1V8..
www.ortodoxism.ro/datasheets/stmicroelectronics/2164.pdf The Texas Instruments data sheet gives also many applications...

Ariba · 2007-10-21 21:39

How about this circuit:

               &#61600; +3.3V
               &#9474;
               &#61628; 33k
               &#9474;
 Inp &#61609;&#9472;&#9472;&#9472;&#9472;&#9472;&#61629;&#61630;&#9472;&#9472;&#9472;&#9531;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#9472;&#61627; Prop Pin
     &#61609;&#9472;&#9488;   18k
       &#9524;

The threshold voltage of a PropPin is 1.42 Volt (measured). If the Voltage at the Input goes under 0.4V the Voltage at the resistor divider goes under 1.42V. You can make one of the resistors adjustable if you need a very exact threshold.

Andy

lairdt · 2007-10-21 22:09

Ariba - that's almost what I ended up doing, running a different bias setup.

Now I'm having a stupid moment getting my 5v relay to energize... it *was* working, but something changed.

viskr · 2007-10-21 23:46

While the resistor divider to the driving pin may work, its a very poor practice. The threshold of any CMOS input is going to vary with temperature and time. In fact spec for inputs are defined as requiring 1.1V as a low input, 2.1V as a high input, anything in between can turn out to be anything.

If the driver is driving at 0.4 V, then the input to the prop would be at the threshold, which means any power supply noise might kick it into the opposite state. Ie, your input becomes a noise sensor and quite unreliable.

deSilva · 2007-10-22 01:05

@viskr: Your criticism is basically correct for general CMOS inputs. However we have learned that the Propeller behaves - a little bit - differently. It has a very reliable trigger level, and many of our pet applications rely on this...

I am eager to hear what issues you have experienced!

viskr · 2007-10-22 02:29

No specific experience, but I think you are working with a small sample size, unless the input is designed as a comparator vs a reference or has hysterisis, you will probably find a part next year that won't work as expected.

Parallax could probably answer this or the fab house as to what specific characteristics the input has. All the parts from a run will have similar characteristics, but from run to run that threshold will shift.

Beau Schwabe · 2007-10-22 04:42

viskr,

If you are making general comparisons between different CMOS chips, then I would say yes, that the input threshold would vary widely. But if you are only looking at one CMOS family (i.e. the Propeller) then yes, you will see some process variations typically in the 10's of millivolt range from one Propeller Chip to the next. That said, Chip Gracey has spent many long hours with specific attention to how the I/O's behave, so that any process variation, temperature variation, etc. has a minimal impact on the desired result.

Edit

VT can vary due to·changes in oxide thickness,·substrate, polysilicon and implant impurity levels and·surface charge. Accurate control of VT is very important. In the past, it varied up to 50% while today it is controlled within 25-50 mV.

The main cause of variations in process transconductance are due·to changes in oxide thickness.· Variations in W and L are caused by the lithographic process.· These variations are not correlated because W is determined in the field oxide step while L is defined in the poly and src/drain diffusion steps.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

Post Edited (Beau Schwabe (Parallax)) : 10/22/2007 6:04:56 AM GMT

Voltage comparator?

Comments