Impedance and capacitance of RTL872xD's ADC_[0-6] pins

  1. Do the ADC_[0-6] pins use a sample-and-hold sampling circuit?
  2. What’s the recommended/maximum input impedance of the ADC_[0-6] pins?
  3. What’s the capacitance of the ADC_[0-6] pins’ sampling buffer?

I couldn’t find this information on the RTL872xD’s datasheet. Thanks in advance!

1 Like

It is a 12-bit SAR ADC, I guess sample-and-hold is included.
The electrical characteristics of ADC pins can be found in the datasheet as listed below. Hope these can help.
Input Impedance Bypass mode 4 MΩ
Resistor divider mode (1/3.3) 500 kΩ
Sampling Capacitance Bypass mode 1.9 pF
Resistor divider mode (1/3.3) 1.9 pF

1 Like

Thank you, @MoonMoon. I had an older version of the datasheet that was missing this information.

1 Like

Thanks @MoonMoon for the info~

@basteagow the latest released datasheet can be found here
https://www.amebaiot.com/?s2member_file_download=UM0401_RTL872xD_Datasheet_v2.9.pdf
(downloadable after registered a free account)

@xidameng I’m confused by some of the ADC specs in version 2.9 of the datasheet (section 7.4, “ADC Characteristics,” pp. 40-41):

  1. What are “bypass mode” and “resistor divider mode (1/3.3)”? Does this refer to an internal divider that can be toggled on/off? If so, how? I can’t find any configuration registers or eFuse bits related to this in either the datasheet (UM0401) or application notes (AN0400).
  2. According to the datasheet, input impedance is the only parameter affected by “bypass mode” (4 MΩ) vs. “resistor divider mode (1/3.3)” (500 KΩ). What about the input voltage range; is this not affected as well? If so, how?
  3. Do you guys have a block diagram/model of the circuit behind the ADC pins? The datasheet is severily lacking in this aspect and it’s unclear whether there’s a buffer and/or amplifier involved, and what the “resistor divider mode (1/3.3)” really is.
1 Like

Although I have some experience on SAR ADC design, yet I am a novice of RTL8XXX. So I am also like to know what are “bypass” and “resistor divider” modes. Maybe a block diagram can tell us about these. Or, I just can guess that, default the input is connected with a resistor divider (for some reason like protection, voltage limitation, …). You can bypass the resistor divider and connect input to subsequent buffer or sample-and-hold directly. It is just my guess.

Sadly, rather than publicly answer such basic questions about their product (which we wouldn’t have to ask if the datasheet didn’t suck in the first place), Realtek will contact you privately, pull you into an email thread with a bunch of employees, demand that you sign an NDA, jump on conference calls, and discuss your company and application at length before they’ll release any information not present in the datasheet. If you don’t comply, they’ll immediately ghost you.

This would be understandable if (a) we were asking for internal implementation details that couldn’t be described without divulging valuable IP or if (b) Realtek were trying to market this platform only to large OEMs, but neither of these is the case: (a) We’re only asking for basic information on how to externally interface with their product, which could easily be answered without giving up any trade secrets (and which other manufacturers—e.g. Nordic—include in great detail in their datasheets), and (b) Realtek is clearly trying to market Ameba to the maker community.

The problem is that Realtek wants the popularity of an open platform without actually opening anything up: They refuse to answer basic questions, their NDA process requires that you work for a medium-to-large company and go through a bunch of red tape, they hold documentation hostage, they can’t even write a decent SDK, and they purposely pull functionality (e.g. Bluetooth and USB stacks) from the standard SDK that’s publicly available. It almost feels as though they don’t want people adopting their platform… :roll_eyes:

@MoonMoon My testing indicates that there’s no internal resistor divider or, if there is, it’s not on by default. It’s possible that the datasheet is actually referring to the comparator peripheral’s resistor dividers (which can work in conjunction with the ADC in comparator-assisted mode) rather than the ADC itself. :man_shrugging:t2:

So the information we can see is that the input voltage is limited to 0-3.3V for the ADC with 4M-ohm impedance and 1.9pF capacitance. The maximum sample rate is 1M-sample/s, and will be lower if the sensor impedance is included and limited to the equation listed in the data sheet, repeated here 8.4Rout1.9pF - 0.5 us < 1/fs. That’s all. By the way, I am wondering what are you going to do with these parameters?

I know this is an old thread, but it is a very reasonable question and the answer to “what are you going to do with these parameters” - in my case is - “use it to design the appropriate interfacing for my custom board!”.
It may be worth adding that I was initially seeing very low apparent ADC input resistance, but I found that “something” was setting the PB3 pin to ‘output mode’. Having explicitly made it an input, I am now seeing the expected behaviour, and it does seem as if the input impedance is indeed in the megohm region.

2 Likes