Is there a preferred range for the resistors used with the inputs on a 339/393? I know how to figure the voltage divider end of the circuit, but what ohm's should be used? 1kΩ, 10's of kΩ, mΩ's ?

Or doesn't it matter as long as the end result voltage is right.

 

Is there a preferred range for the resistors used with the inputs on a 339/393? I know how to figure the voltage divider end of the circuit, but what ohm's should be used? 1kΩ, 10's of kΩ, mΩ's ?

Or doesn't it matter as long as the end result voltage is right.

Not sure, but you can start here.
http://web.engr.oregonstate.edu/~traylor/ece112/lectures/comparator_ckts.pdf

I'm going to assume "10's of K" may be in the ballpark. You don't want to load down the circuit that drives the comparator by using values too low, so high resistor values seem to make more sense.
If a value of maybe 10K seems "buggy", try 100K....the breadboard is your friend

 

IMO main limitations are:
- too much power dissipation through low value resistors, so I calculate them to be able to use the small 1/8 or 1/4W resistors.
- easy noise pickup or drift due to leakage currents if the values are very high, i.e. a few MOhms

So I usually take anything from 1k to 560k.

 

Thank you both for answering! I have learned enough about how stuff works, and can build from a schematic. But now I'm trying to design a circuit of my own, and want to have a chance that it won't let out the smoke, first time its turned on.

Is it kosher to use trimmer pots to set voltages for a comparator circuit? The project will need some tweaking before it works right, I know. I've seen trimmers in store bought things that have trimmers, then they use what looks like nail polish to lock the adjustment screw at the correct position. Bad practice or not?

 

Trimmer pots are not the most accurate things in the world, the cheap ones especially do tend to drift. You can buy some pretty good ones, even some that turn not thru 270 degrees but 15 full turns around.

Without knowing your circuit I could not rule them out, and they may just work and at worst you will still learn something

 

You could use the trimpots for initail adjustments. Once you know you won't adjust them anymore you could measure them and replace them with resistors.

 

Take a look at the input offset voltage and input bias current on the datasheet of the comparator.
Suppose you want to keep the voltage drop across the input resistor on par with the input offset voltage, which is about 1mV. If the input bias current is about 100nA then the input resistance would be about 10kΩ. That is a typical value for the input resistor.

 

It depends on how much accuracy you want vs how much power you want to use.
These comparators' inputs have bias current and offset current. Bias current should be self-explanatory. Offset current is the difference between the bias currents of the two inputs. In the LM393/339, at 25°C,
Ibias=250nA max
Ioffset=50nA max
If you connect a voltage divider to one of the inputs, the actual voltage at the input (when the differential input voltage=0) will be
Vth+Ibias*Rth
Where Vth is the unloaded divider voltage, and Rth is the equivalent resistance looking back into the divider. If it is simply made up of two resistors, R1 and R2, Then Rth=(R1*R2)/(R1+R2).
So, for example, if Rth=1MegΩ, then the voltage due to Ibias could be as much as (1Meg*250nA)=250mV above Vth. If Rth=10k, then the offset will only be 2.5mV max.
Now, if Rth on both inputs are identical, then the differential offset will be Rth*Ioffset. So, with Rth=1Meg, the differential offset will be 50mV max.

EDIT: While I was composing (and watching Closer), MrChips posted a succinct version of my treatise.

 

Thank you all! Like I said until now I have a small knowledge of how things work internally, and in a circuit, but not the ins and outs of there interconnection in a complete circuit.

You guys here are and have been a great resource. I'll be having a lot more "dumb" questions coming up. Thank you all.

 

The bus is getting shorter and shorter.