[time-nuts] Question about SoundCard stability?

[Bob Camp]

Hi

Here's pretty much how it works:

Integrating capacitor:

You rarely if ever see one in an oven controller that's actually acting as an integrator in the control sense. What you do see are caps to roll off the AC gain of the loop to drop noise. In other words the time constant is way to short to be useful in a control sense. Since you would need 100's or 1000's of seconds in the integrator that's not real surprising. 

Resistance: 

A normal oven controller acts as a current sink. Low temperature = pull more current. The gotcha comes when the voltage drops as the current increases. You pull more current, but don't get (much) more power. It takes longer to get back to equilibrium than normal. Once you get there you have the opposite problem on the hot side. Drop current and the power does not go down (much). The controller "hunts" with a period measured in seconds or 10's of seconds.

Insulation:

Insulation adds to the thermal gain. Pull an amp and the oven should heat up by 50 C. With X extra insulation it heats up by 100 C when you pull an amp. Twice as much gain => way more gain than the controller was designed to accommodate. 

Thermistor:

As long as the gain is correct for the thermistor location, the controller will be stable. Too much gain / to far away is a problem. There is indeed a correct gain that pairs up with any rational thermistor location. 

Been around a while myself ....

Bob


On Oct 13, 2010, at 11:36 AM, Bill Hawkins wrote:

> Group,
> 
> The primary controller issue is too little reset because the
> integrating capacitor is too small or bad. After that comes
> too much gain.
> 
> Power lead resistance would reduce the gain. Conversely, high
> voltage would increase gain.
> 
> Don't think extra insulation is a stability issue, unless it's
> between the heater and the sensor.
> 
> One other thing - dead time will destabilize the control loop,
> when there's too much delay between the heater's heat being
> sensed by the thermistor. This is usually a location problem.
> 
> Bill Hawkins
> 
> (50 years in industrial process control)
> 
> -----Original Message-----
> From: Bob Camp
> Sent: Wednesday, October 13, 2010 5:58 AM
> 
> Hi
> 
> Controller cycling is often a result of one of two things:
> 
> 1) Resistance in the power lead 
> 2) Extra insulation / dead air
> 3) Internal controller issues
> 
> There are a few other possibilities, but they are remote enough that
> you are unlikely to ever come across them. There's no advantage to
> building a controller that's cycling. It was more likely a bug than
> a feature.
> 
> Bob
> 
> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.