Home Brew Forums

Home Brew Forums (http://www.homebrewtalk.com/forum.php)
-   Electric Brewing (http://www.homebrewtalk.com/f170/)
-   -   One solution found for overshooting with Auber2352 (http://www.homebrewtalk.com/f170/one-solution-found-overshooting-auber2352-273223/)

CadKing 10-08-2011 01:43 AM

One solution found for overshooting with Auber2352
 
I had been having problems with my e-kettle always overshooting the set temperature by 2 to 4 deg. The problem I figured was that the time it took to register the heat at the temp probe, there was so much built up heat in the liquid that it would keep climbing past the set point. I tried temp probe locations both in the kettle and inline with the outlet of the pump and it would always do the same. The problem it turns out is the 5500 watt element would ramp the heat so fast the PID couldn't get control of it.

The solution I found and I hope may help others is a limit the output power of the PID to roughly 30% when I recirculated for the mash. On this PID this is set by the OUTH setting. I let the PID go full power until I reach my set temp (or very nearly) then go into the settings and force the maximum output to 30% which is more than enough to handle 10 gal.

I think the ultimate solution will be to use a DPDT switch rated for 30 amps to route either 110 or 220 to the element for an easier high/low setting. But that will have to wait until funds become more available.

Anyway, hope this little trick may help if others have the same issue. This has worked great for me and it holds the temp about .5 deg above set point the whole time.

hopsalot 10-08-2011 03:42 AM

did you rig a potentiometer?

CadKing 10-08-2011 03:52 AM

No I didn't. Would that be to lower the amperage going to the element?

Ultimately I think running 110 to the element to control the mash will be the way to go then switch to 220 for the boil. I believe that I should be pushing somewhere around 1400 watts or so at 110 which will slow the temp rise to the point that the PID can control things without overshooting.

GNBrews 10-08-2011 04:51 AM

Cad - How is the Auber changing the output of the element to 30%? Doesn't the output of the Auber just turn the element on/off full power? Do you have a PWM circuit rigged to the SSR?

jkarp 10-08-2011 12:25 PM

A little overshoot is normal in a properly configured PID. The problem with brewing implementations is, we'd prefer a critically damped system which is tough considering all the variables we throw at our setups.

The most common reason for overshoot however is integral windup. We enter significant setpoint changes that exceeds what the integral value can regulate and the error accumulates over time. This error must then be unwound, causing overshoot. The easiest way to handle this is to "reset" the PID as it gets close to the desired SV by momentarily unplugging the RTD.

CadKing 10-08-2011 03:15 PM

Quote:

GNBrews Cad - How is the Auber changing the output of the element to 30%? Doesn't the output of the Auber just turn the element on/off full power? Do you have a PWM circuit rigged to the SSR?
It doesn't change the ouput as much as limit the total output over the two second output range. The PID thinks it needs to throw full power to the element in order to get the temp up to set point. If I were at the controls, I know there is about a 5 to 10 second lag between when the power is applied and when I can see the temp change at the probe so I would shut off early and wait to see what happens before applying more heat (with the 5500 watt element a full power blast for that long really has a drastic impact on the temp of the mash). The PID just isn't that smart so by limiting its full output power to 30% even when it wants to apply 100% it gives the system the time it needs to react to the changes in heat without going crazy.

Quote:

jkarp A little overshoot is normal in a properly configured PID. The problem with brewing implementations is, we'd prefer a critically damped system which is tough considering all the variables we throw at our setups.

The most common reason for overshoot however is integral windup. We enter significant setpoint changes that exceeds what the integral value can regulate and the error accumulates over time. This error must then be unwound, causing overshoot. The easiest way to handle this is to "reset" the PID as it gets close to the desired SV by momentarily unplugging the RTD.
I think if I was able to draw the heated water directly through the tube thats reading the change in temp I would have no problem with what you describe. I tried so many possible variations to the parameters after autotune and was able to get it better but never quite solve the problem. A very frustrating task when it takes so long to test the results of a change in programming that I started looking at other options. The one thing that came to mind is that if I only didn't have such a hot element I could control this so much easier. Thats when I looked at the max output on the PID.

I think most have their tuning set pretty good but I know in searching the forums that I wasn't the only one that had some trouble in this area. Considering that it's a small thing to change this setting and it made my issues go away, I wanted to at least throw it out there as an option (not necessarily the best) for others to try if they were in the same boat.

mabrungard 10-08-2011 04:40 PM

Moving that temp sensor as close as possible to the flow off the heating element will significantly reduce any overshooting in the kettle. I use thermocouples and they are not super accurate, but I use NIST calibrated thermometers in other places through the flow stream to observe the true temp in the kettle or tun.

jkarp 10-08-2011 11:45 PM

Quote:

Originally Posted by mabrungard (Post 3370138)
Moving that temp sensor as close as possible to the flow off the heating element will significantly reduce any overshooting in the kettle.

+1. Moving my RTD to a tee on the output helped tremendously.

CadKing 10-09-2011 03:09 AM

Quote:

Originally Posted by jkarp (Post 3370825)
+1. Moving my RTD to a tee on the output helped tremendously.

I thought the very same thing so ordered a T and another RTD for measurement there but I still had the same problem. It may be that the pickup tube (Blickman 15g pot) pulls more cool wart through while the warmer wart cycles up in a column before being drawn into the pump. I would say though that I enjoy having the ability to measure from the pump while I mash and from the pot while I bring to a boil. It was worth the expense I think. I replaced the Blickman brewmometer with the RTD and it stops registering at 6.5 gal. I recirculate for a whirlpool to aid in cooling and this way I am able to see when its safe to pitch.


All times are GMT. The time now is 08:49 AM.

Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.