Are 100k Ohm pots necessary?

[MikeDubreuil]

I'm trying to find high quality single turn potentiometers and having a difficult time finding them at the 100K ohm requirement. I understand that the Robot Controller's manual indicates you should always use 100K potentiometers. But is it really necessary? What could be the disastrous result of using for instance a 5K potentiometer?

[Al Skierkiewicz]

Mike,
I have the input equivalent circuit around here someplace. I will continue to try and find it so it will help understanding the circuit. In the mean time, the 100K pot is selected to work with the input circuitry on the RC and the OI. The pot is connected to an A to D convertor through some input protection circuitry in each of the devices. Note that the wiring is different for the RC and OI analog inputs. Additionally, joysticks use the 100K pots.

[MikeDubreuil]

It was my understanding that the A to D converter simply compares the voltage on the signal line to it's 5 volt reference. What would happen if you were to use a 5K pot?

[Mike Betts]

Originally Posted by MikeDubreuil

It was my understanding that the A to D converter simply compares the voltage on the signal line to it's 5 volt reference. What would happen if you were to use a 5K pot?

Mike,

You would probably get away with it...

Let me extend your line of reasoning:

Q: What would happen if I use a 10 ohm pot?
A: I doubt if the 5V internal power supply in the RC can provide 5V/10ohm=0.5A of current.

I would think that 100K is probably a value such that a team could use 100K pots on all analog inputs simultaneously and it would work.

For a definitive answer, I would post on the IFI Technical Forum.

Regards,

[KVermilion]

If you really want "high quality" use a stepped attenuator...

Any <100k> pot should do, as long as it's controlable... if you are worried about bad tracking, noise, etc thats a fact of life. If you want a really high qaulity pot you might be able to find a pana evj or alps blue velvet in 100k varieties...

[Joe Ross]

Originally Posted by Al Skierkiewicz

Mike,
I have the input equivalent circuit around here someplace. I will continue to try and find it so it will help understanding the circuit. In the mean time, the 100K pot is selected to work with the input circuitry on the RC and the OI. The pot is connected to an A to D convertor through some input protection circuitry in each of the devices. Note that the wiring is different for the RC and OI analog inputs. Additionally, joysticks use the 100K pots.

IFI sent me the input circuitry for the RC this morning. They said they would post it, but I haven't seen it show up yet. I've attached it in the meantime.

The input circuitry for the microprocessor is on page 217 of the datasheet.

[MikeDubreuil]

In the IFI Forum it says the Robot Controller is capable of sourcing 1 Amp of current on the 5V I/O headers.

I obviously don't want to draw more current than I have to, but I have been having difficulty finding 100K precision single turn pots.

After looking at the schematic it seems impossible to draw more than 5mA of current due to the 1K Ohm resistor in series with the load.

I'll submit this as a question to the IFI Forum.

[Al Skierkiewicz]

Thanks, Joe.
Mike, if you take a look at the input circuit you will see the series resistor and capacitor. This circuit does a little low pass filter. (about 3kHz) Since this is a simplified drawing, you can guess that the A/D that follows is selected (optimized) for a particular input current/frequency in order to track and remain linear. The 100K value allows that, but a 10K would not. It is not a matter of possible damage, just not predictable results or noise immunity.

[BrianBSL]

Here is a high quality 100kohm pot we have used before:
http://www.digikey.com/scripts/DkSea...423225&Site=US

They arn't cheap, but its practically impossible to break the stops on them. That particular model is the long shaft version, they are also avilable in a short shaft version (but its far easier to shorten a shaft than lengthen a shaft, so we usually buy long shaft versions and cut them down to whatever we need).

[MikeDubreuil]

IFI has updated the 2005 Robot Controller Reference Guide. The new potentiometer requirement is:

Originally Posted by IFI RC Reference Guide; page 8

Always use 250 – 100KΩ potentiometers.

[Al Skierkiewicz]

Cool!
Thanks.

[Joe Ross]

Originally Posted by IFI RC Reference Guide; page 8

Always use 250 – 100KΩ potentiometers.

I would assume that means 250Ω - 100KΩ, since they show a 10KΩ pot below it. My first reading was 250KΩ - 100KΩ.

[Dave Flowerday]

Just wanted to point out that

Originally Posted by IFI RC Reference Guide; page 8
Always use 250 – 100KΩ potentiometers.

only applies to the Robot Controller, not to the potentiometers used on the Operator Interface (unless IFI changed the circuitry in the OI, but I doubt they have). The OI has a different circuit than the one Joe posted such that if you use anything but a 100K pot you will not get the full 0-5V range.

For those who are interested, this is done so that if there is no joystick plugged in, the analog inputs get pulled down to 0V and the OI can detect this and in turn send 127 to the robot (instead of 0). It's a safety feature so that if your joysticks get unplugged your robot doesn't start driving backwards at full speed.

[Max Lobovsky]

I'm sorry, am I an idiot or is it not possible for RC to use anything from 250 - 100k ohm potentiometers without something else in the circuit? For example, how can the RC possibly tell if its connected to a 10k ohm in the mid position, or a 5k ohm at the max position? Will you just be limited to some small section of the resolution if you use poteniometers less than 100k ohm?

[Dave Flowerday]

Originally Posted by Max Lobovsky

For example, how can the RC possibly tell if its connected to a 10k ohm in the mid position, or a 5k ohm at the max position?

A potentiometer is really just a variable voltage divider. Using Ohm's Law (voltage = current multiplied by resistance or V=IR), then at 5 volts there is (5V / 100K Ohms) or 50 uA of current flowing across the pot. Now, if you set that pot at its mid-point, you have 50K on either side of the resistor. Again, using Ohm's Law, you have V = (50uA * 50K Ohms) = 2.5 volts of a drop across that half of the potentiometer.

If you repeat this exercise for a 10K pot, you'll find that if it's set at its mid-point (5K Ohms on either side), you still have 2.5 volts dropping on either side. So in either case (10K or 100K), with the wiper set at mid-point, the RC sees 2.5 volts. The only difference is that at 10K more current is flowing. The reason they limit you to no less than a 250 Ohm pot is because any pot with a smaller resistance will draw more current than the RC is capable of providing.

Hopefully at least some of that made sense