Tuesday, June 19, 2018

Digital Pot


No, this is not some on-line mail order thing.  Bruce KC1FSZ has done something that I find both amazing and horrifying:  he using a "digital potentiometer"  as the AF gain control in his BITX.  Remember when pots were simple and pure, when they were just a metal wiper that ran across a crescent of carbon of increasing resistance?  Well, those happy days are apparently behind us my friends.   What now?  Will we have endless discussions about where to find the driver for the 10k pot?   Will Facebook or Amazon or Microsoft secretly seize control of the data on our AF gain, perhaps to sell us hearing aids?   But enough of my ludism.  Good going Bruce.  AGC beckons. 

Hi:

I've built a few BITX rigs from scratch now and I'm starting to move to a modular approach using fabricated PCBs to make it easier/faster to build/mix/match the stages.  My audio amplifier module is based on the LM386.  I know others have moved away from this chip, but it still works for me.

I've added one feature to my audio amp stage that opens up some good possibilities.  Instead of the traditional potentiometer between the AF preamp and the LM386, I've inserted a digital potentiometer (MCP4131-103) that is controlled by three extra pins on the microcontroller.  The digital pot lists for $0.81 on Mouser so it's no more expensive than a physical pot and it's way more versatile.

Why do this?

  • The module is quick to make if you're building a lot of rigs.
  • AF gain can be controlled via the CAT port.
  • AGC control can be done in software.  This enables all kinds of fancy AGC "attack" and "release" controls that can be seen in the docs for commercial rigs.
  • This is an enabling feature for remote ops, which is why I did this in the first place.
  • AF gain can be controlled using the same encoder (different mode) as is used to control frequency, etc. if you wanted to build a more compact rig.
The board is shown below.  This is one of my first forays into KiCAD/PCBs/etc. so it's probably not as "tight" as it could be.  Also, I'm using through-hole components so it's old-school building.  But the board only costs $7 on OSHPark and it's pretty quick to wire up.  Anyone who wants to order one from them drop me a line and I'll send you the link.

Schematic below.

73s,
Bruce KC1FSZ




1 comment:

  1. Digital pot is very useful in many cases, but is plenty of problems too to have in mind.
    The pots I used were Microchip 256 taps. Well, it looks a high resolution, or my be not?. From 256 to 128 are 6 dB and from 128 to 64 6 dB else, and so on. As higer attenuation worst resolution from 8 to 4 6 dB in four steps. Log ancient pots solved the problem but digital ones not or I did not find them.
    For AGC purpose we have many problems too. Attack time is function of uProcessor speed and programing port speed and length of sample. So, pot AGC is not very nice against classical RF detector AGC but it is still a solution better than manual control.
    AGC window decission has to be higer than 3dB up to 6dB, so not very good but aceptable.
    Combining two potentiometers we get a better resolution, final theoretical attenuation 96dB against 48dB only one pot. It is a bit SW complex but not too much.
    Noise, serial port either I2C or SPI may generate a lot of clics if pot is istalled before a large gain amplifier. Let say 50dB like LM386.

    ReplyDelete