ADC input Z

[Jaguarjoe]

Hi guys, I've been gone for quite a while due to an extended bout of stage 4 esophageal cancer. I have more stitches than a baseball! Actually, the cancer itself was pretty greuling but once all the surgury was over I had no aftereffects. I can't say that at all for the chemotherapy though. It loosened my front teeth till I lost three, I had to get dentures. My lining in my mouth and on my tongue is pretty much destroyed so now I can't eat anything remotely spicy. I was given a 5% chance of survival and here I am!
FWIW, there's talk of a better way than chemo which should be described as cruel and unusual punishment, like a Catholic grammar school. Oh yeah, 12mos on a feeding tube too. I had to retire way earlier than planned. Nothing beats going from $130k to $40k instantly.

Anyhow, one simple question- what's the adc input Z on the 08m2? MCP says it's 10k but "can be higher" with a cap across the input. How much is "can be higher"? Why the cap?

Thank you very much. It's nice to get back in the groove.

 

[BeanieBots]

Sorry to hear about your woes. FWIW in my day job I work on linear accelerators used for radio therapy so I have a very small outside understanding of what you are going through.

As for your ADC issue, it's complex and I mean that in the litteral sense as apposed to L & Cs.
The ADC works by taking a sample using a standard sample & hold method within the chip. This means that effectively a capacitor is switched in out of the circuit. The initial charge on that capacitor depends on where it was connected before (another ADC input) and what voltage was at that point at the time it was connected. Hence, the load on the ADC input depends on what voltage is on other inputs and which other inputs you have read and in what order.
The DC input resistance can be ignored as it is extremely high compared just about anything you would ever want to connect.
However, to avoid problems caused by the sample & hold principle (as described above) and also to provide a low enough impedance to charge the sample & hold cap, your source impedance should be kept below 10k to ensure accurate conversion. This is the PIC datasheet spec but in reality can be as high as 50k before you start noticing the effect over and above other effects such as supply rail decoupling. Because the issue is charge related, putting a cap on the ADC input helps a lot. I have often used a 100k source with 100nF on the ADC input and observed perfect conversions. However, that is only true when only using one ADC input. If you constantly switch between ADC inputs at extreme different input voltages you will notice several bits of error even with a 100nF on each input.

If you have the time, patience and suitable test equipment, try it out. Also observe the effects of not using the enhanced download circuit which becomes particularly crucial when using high source impedances.