Putting the Lid on Noisy Pots

[JBrookes]

Im sure this has been discussed before, but what is a good pot (10k) too use for Picaxe ADC (18m2 or similar)?
-wirewound?
-brand?
-carbon OK?
-attach noise filters to (caps to ground.)

The history of this is I used a generic carbon pot on an ADC input, and it was a good random number generator!

Be interested in opinions...

 

[hippy]

I have only ever used generic carbon track pots and haven't had major problems. There might be a +/-1 bit change in ADC, maybe a bit more with longer wires, but nothing really major. If it really is 'random number' stuff, that suggests the way it's connected may not be quite correct or there's something else going on.

 

[eggdweather]

Did you check the quality of your PICAXE power supply - batteries or mains powered adapter (Wall Wart in the US )?

 

[westaust55]

Are you sure that you had the pot connected to the same ADC pin that you were reading with READADC?

That has caught a few folks out in the past.
Maybe posting schematic and corresponding READADC line will allow others here to cross check.



I
Have
Only ever used generic conventional carbon track pots - often with far greater values than 10 kOhms (to maximize input voltage range) without dire consequences.

 

[premelec]

Cermet and molded carbon AB pots seem to work well [AB pots got quieter with use in some tests years ago]. Have you tested the stability of your pot with an ohmeter &or with input to an oscilloscope [through a capacitor will give you noise without DC offset].

 

[szd51pilot]

In a carbon pot, I believe that the current through the track zig-zags through the carbon granules and the slider picks up small voltage variations as it moves. I suggest it may not be a good choice for your application.

Otherwise, cermet or plastic film (expensive) or wire-wound should be OK. Probably it is only worth digitizing to 8 bits if you use a single turn pot. My current project requires 10 bit resolution and uses a 10-turn wire-wound pot with an R-C filter (always a good idea) to the PICAXE pin and it works just fine.

Richard

Im sure this has been discussed before, but what is a good pot (10k) too use for Picaxe ADC (18m2 or similar)?
-wirewound?
-brand?
-carbon OK?
-attach noise filters to (caps to ground.)

The history of this is I used a generic carbon pot on an ADC input, and it was a good random number generator!

Be interested in opinions...

 

[JBrookes]

In a carbon pot, I believe that the current through the track zig-zags through the carbon granules and the slider picks up small voltage variations as it moves. I suggest it may not be a good choice for your application.

Otherwise, cermet or plastic film (expensive) or wire-wound should be OK. Probably it is only worth digitizing to 8 bits if you use a single turn pot. My current project requires 10 bit resolution and uses a 10-turn wire-wound pot with an R-C filter (always a good idea) to the PICAXE pin and it works just fine.

Richard

hmmm
The instability was when I was not turning the pot shaft - seemed to act like a microphone - I had pot in series with a thermistor...
Any ideas on RC filter values?
More experimentation needed...
Thanks
JB

 

[erco]

Try a different pot & thermistor, resolder all connections super well. If noise continues, bypass (short) thermistor and test pot's stability alone as a voltage divider.

http://www.talkbass.com/threads/can-pots-go-microphonic.843051/

 

[szd51pilot]

Hi JB,

The RC filter time constant does depend upon the application. I need to keep RF away from the ADC input and use 1k and 0.1uF disc ceramic to 0V. Your project might benefit from a larger capacitor.

Are you running your pot/thermistor from the same supply as the PICAXE and is it stabilized e.g. 5V regulator rather then a battery? If you are also using an OLED display, then I suggest running the OLED and PICAXE from separate regulators - I have found digitzation jitter when running from the same 5V regulator.

Richard

hmmm
The instability was when I was not turning the pot shaft - seemed to act like a microphone - I had pot in series with a thermistor...
Any ideas on RC filter values?
More experimentation needed...
Thanks
JB

 

[BeanieBots]

Define "good"?
You get what you pay for and pots vary in price from 50p to over £50. (YES, £50 for a pot!).
For 8-bit PICAXE use, a cheap carbon pot should be just fine. Should also be OK for 10-bit with a little care.
The problem with carbon pots is that they dry out. (crackle when you turn up the volume on an old radio)
A quick spray of cheap aftershave on a crackly pot works wonders!
Carbon actually generates noise but it should not be a problem until working at > 12-bit. (hiss on a cheap radio)

If you have a lot of random variation on your reading even when not moving the pot, then as hippy says, something is wrong.
Things to look for:-
Wrong ADC connection.
Wrong value pot. (should be 10k, 20k or above will cause erroneous readings.
Long wires. These will pick up mains hum just like a HiFi does.
Proximity to a power supply. Switchmode supplies transmit a lot of noise and transformers transmit a lot of mains hum.

Filtering can help to remove unwanted noise but it is always better to avoid letting it in than it is to let it in and then remove it.
I religiously put a 0.1uF on any ADC input. No need for a resistor because the pot itself will provide sufficient source impedance.

 

[eggdweather]

Hi JB, what value of Thermistor are you using (they are usually in preferred values ranging from 100R up to 100K at 20'C) and what is your wiring configuration? Do you have one end of the pot to Vcc (+5v) the wiper to the PICAXE ADC input and then the other side of the pot to the Thermistor then to ground? If so, you will need to consider the values in use to get the best from your ADC, for example a 1K pot and 100K thermistor (at room temp) would not be a good combination of values.

 

[JBrookes]

hmmm---Yes wiper on pot connected to ADC and also to thermistor and then on to ground. The problem went away when I used a fixed resistor in place of the pot. The R of the thermistor and pot were about the same value.
I had concluded that a cheap pot doesn't hold water for ADC, but now will try again with a new handle on subject.
To this day I don't know source of noise - must have been within pot, since fixed resistor fixed problem.
?
JB

 

[john2051]

Hi, I suppose you could consider the I2C semiconductor 'pots'. In theory they should do well with no
moving parts. I have several of these, but I've only ever used them with op amps, and not an adc.
john