>a) I already have one of those (used for depth control) but the response
>time is connected to how long it takes for the sub to settle at a depth (the
>sensor is not inside of the ballast tank).

If you are using pumped air, the air pressure in the tank will accurately relate
to final depth.  Assuming your tank floods rather freely.  If slow flooding (due
to constricted porting) it will not, as the air will be at a lower pressure as
the tank floods.

>b) I'm not sure that it is accurate enough to repeatedly take on a fixed
>amount of water in the tank (I seem to be able to measure down to about
>5/10mm of water with it).

A colored floating ball in a vertical tube parallel to the main tank will give
you depth indication.  Getting that visual indicator into the micro will be a
challenge...

>I had thought about using the depth sensor for this (and maybe moving the
>sensor into the tank to do it). But I'm concerned with the effects on the
>closed loop of the lag that you get between taking on water and the sub
>actually moving to a depth. Maybe if I move the sensor I won't have this
>problem (since it would be inside of the tank?). But I guess I would still
>have the problem of sensitivity/accuracy.

On Skip's control, I had better than 1/4" resolution of depth.  It was "dumbed
down" for the mass market.

>Just a quick note here. The way I like to run my subs is not to spend ages
>ballasting them so that taking on a full tank of water takes them to a
>neutral buoyancy (which is what most people do). Instead what I would like
>to do is to automate this part so that I can just do a simple pond side
>set-up of taking on water until the sub is neutral then store the value
>associated with this state in eeprom. Then I just command the sub to take on
>the same amount of water when I dive. I currently do this open loop (timing
>based) but this tends to be a bit hit and miss as the batteries go down...

I've only been out with Skip once, but there was no neutral setting.  He set the
boat up so that with a dry tank it had a scale waterline (lead ballast) then we
just drove it.  The APC and ADC worked to get the depth and pitch right, and it
just held itself there.  I think on this boat, a flooded tank would have allowed
it to continue diving a long time, and it was flying itself up instead.  We
stuck it to the bottom a few times (in weeds) and blew the tank to recover it.

Pretty cool, but not enough fun to get me to forsake race boats and airplanes!

Andy

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.