Re: [ATM] Making glass blanks

["Stuart Hutchins" <stuart452@earthlink.net>]

----- Original Message ----- 
From: "Martin Cibulski" <martin.cibulski@t-online.de>
>
> Do you mean that the comparison between setpoint and measured temperature
> is done in a controller outside the computer ?
> Then I suggest a different solution.
>
> The computer reads the temperature value and compares it to a setpoint
> inside
> the software (which in turn is changed slowly for the annealing process).
> The temperature difference is used as input to a PID control algorithm.
> The PID's output (usually scaled as 0-100%) will be translated
> into a pulse width modulation which can be slow and so handled
> in software even under Windows. A 5 seconds PWM cycle with 0-5
> seconds on time in each cyle will be fast enough for temperature control.
> The on/off signal from the PWM can be written to one BIT of the printer
> port. This bit is used to switch the solid state relais.
>
> So no external temperature controller is needed because the controller
> is part of the software. Only one solid state relais must be switched
> directly from one bit on the printer port.

But...you must have a TC amplifier and some computer interface to digitize
and read the temperature, so you have extra circuits anyway.  The
"controller" I'm thinking about consists of an LM393 Voltage comparator
driving the SSR.  This is a simple bang-bang control with about 5 degrees of
temperature hysteresis built in  like a household oven.

Is the sophistication and tuning bother of a PID algorithm really needed for
annealing glass?  What precision in temperature schedule is required?

In my experience, PCs are not reliable over the time required for annealing.
I would worry about all the ways this monster could fail, power could fail,
stuck on and melt the glass, etc.

Stuart


_______________________________________________
ATM mailing list http://www.atmlist.net/