Webcontrol next generation - feature requests

[Michael McSharry]

Are lower & higger limits the thresholds associated with each measure in the Webcontrol board?
In my mind I wanted to keep these thresholds for trigger a low or high alarm and control the HVAC with a soft setpoint ( I mean a value stored in board memory, this value that I can change from web browser)

The limits are associated with each WebControl input. If you want both a control function and an alarm function then you will need to use two inputs

To achieve this it means a computer must be running xapmcsWebControl, because the WebControl board isn't able to do that. Correct?
In my project the board must be able to run alone.

The board is able to perform proportional control without a PC connected. This should be good engough for most applications. At some remote location you can use xapmcsWebControl to adjust a setpoint and it will download the apprropriate parameters to the WebControl module to control to this new setpoint. xapmcsWebControl would only need to run when you want to change the setpoint.

One issue also is that you said here above about temperature accuracy. If 1-wire temperature measurement is not very accurate that's a problem. May be this can be solve by using another tempertaure sensor (such as PT100) on an analog input.

The temperatrue sensor supported by WebControl is plenty accurate for proportional control. You need a DS18B20 made by Maxim-ic and with a stable voltage source. That is a shortcomming of the WebControl module is its robustness for ranges of DS18B20 sensors that are available. You may need a more accurate sensor if using PD control as the derivative calculation will be affected by it. In your case it does not matter since you will want the control to be dependent only on the temperature error and not the rate of change of the error.

 

[JumpZero]

Hi

The board is able to perform proportional control without a PC connected.

!?...
proportional control needs calculation (and more with proportional-derivative)
The board with BRE (Boolean Run Engine) is not able to perform any calculation.

Did I missed something?

Thks
--
Jmp0

 

[Michael McSharry]

Hi

The board is able to perform proportional control without a PC connected.

!?...
proportional control needs calculation (and more with proportional-derivative)
The board with BRE (Boolean Run Engine) is not able to perform any calculation.

Did I missed something?

Thks
--
Jmp0

BRE is able to make a comparison within bounding limits and perform an action based upon the result. This is the extent of the calculation that is needed to do proportional control. The hysteresis element is handled with the boolean logic that uses current state to know if upper or limit should be the active limit.

 

[JumpZero]

Hello

OK I see : this is on-off control not proportional control.

http://en.wikipedia.org/wiki/Proportional_control

http://en.wikipedia.org/wiki/Control_system

--
Jmp0

 

[Michael McSharry]

No problem with this wiki author's definitions. The PID controller is the general case where error, rate of change of error and steady state error are controlled. The PD controller does not account for the steady state error. The P controller accounts only for the error between control point and current input. When the output of a proportional controller can take only only two states (ON/OFF) then you have something like a thermostat. Most commercial theremostats include a derivative component that is referred to in the lingo as anticipation. They are PD controllers even though the output has only two states. The PD contoller will be better able to control overshoot if the accuracy of the D term is sufficiently good. In real world applications this accuracy is not always easy to achive for D. A bang bang control system might be used to control something like a furnace or engine where multiple valves of different sizes are used in some combination to control the volume of fuel flow. Classic control theory does not deal real well with bang bang control because it is not a continuous system.

 

[CAI_Support]

We are sorry joined this discussion late.

In the WebControl HW rev 2.2.2, which is shipping now, we have added active pull up on the 1-wire bus, also changed the working voltage from 3V in hw rev 2.0.2 to 5V. We think those changes fixed most of the problems you had seen with temperature sensors, in addition, supporting longer 1-wire bus.

In the future coming WebControl32 board, it uses a Maxim-IC 1-wire bus mater chip to control the one-wire bus, so that it will have same 1-wire bus master like i-Button. Our goal is to provide an I/O extension board, with stand-alone control features, for all your control system software to easily talk to the devices through WebControl32. You can install a serial driver on Windows or Linux to access the 1-wire bus like those 1-wire devices are directly connected to your PC serial port - when those 1-wire devices are miles away or 1000 miles away from your controlling computer over network on the WebControl32 board.

WebControl-32 is a bigger brother of current shipping WebControl-8. WebControl-8 will be continuously developed, offered and supported. WebControl-32 is more expensive than WebControl-8.

 

[fwd03]

I have WebControl 2.0.2 boards. I removed the protection diode on the back of the board, and jumping the soldering pads, made 3v directly feed the pin 3 of the temp sensor. That works great.