Display MoreThanks for your comments.
Yes, I will use PWM with IGBT very soon. My IGBT driver board could be finished this week.
Here I can change duty cylce and frequencies very well.
Also I plan to add manometer this week and refuel it. Old fuel will be stored for possible examination.
Tommorow from approx. 5:00-6:00 UTC next test with current reactor will be started.
I left the chat room after 9 PM EDT
If your mains are varying a lot, that might make it worth a change in power strategy.
Have the mains feed a switching power supply that puts out regulated DC at a high enough voltage to run your resistive heaters. Switching supplies have gotten quite inexpensive for a few hundred watts.
Then inject whatever electromagnetic noise you want to try, either into its own coil winding or, perhaps easier, isolate the heater coil from your regulated DC using an inductor/choke and capacitively couple your noise signal to the heater coil.
Or, you could transformer couple the noise to the circuit--have a high-current-capable secondary in series with the power supply and feed your noise into the primary. If that varying load current makes your supply unhappy, add a big capacitor across the supply output.
Then you could experiment separately with noise waveforms and temperature.
Actually, another approach could be easier--since you need to computer control the supply output anyway, in order to change the temperature, you could use a fixed-voltage switching supply just to isolate yourself from the power mains, and both vary the average DC level to control heating and also vary the pulse characteristics to produce the noise characteristics you like. There are a few very fast GaN power transistors available now that could make some pretty sharp signal edges.
But who knows what kind of electromagnetic noise is needed or helpful...
To compute the COP of your reactor, you need the input current and the variable input tension, then you must record the variable input tension display on the right of your reactor. Of course with a sensor input like for current. Then, in your graph, you will can replace the current by the input power.
simply multiplying U and I is not enough.
if frequency is stable and waveforms stable you can calibrate and estimate the phase shift... but with triac, and similar switching regulator it is meat for the skeptics.
best solution is indeed using DC, or at worst PFM (pulse frequency modulation) with low duty cycle...
effect of HF is either null, or controlled.
of course the best if good powermeter.
note that with good filters, an old powermeter (a billing grid energymeter?) can measure real energy as long as bandwidth is narrow to few harmonics. today it is collector!
Unfortunately we can't do nothing with the voltage. This variable voltage is coming to the building from supplier.
Hi me356. Your setup and experimental runs have been impressive. As far as the voltage problem you have, it sounds like you need a current source. That is, a power supply that provides a constant current as opposed to a constant voltage. Another way to think of it is to consider a PID controller driving its output based on current (amperage) measurement and comparing it with a current setpoint (instead of measuring temperature and comparing it with temperature setpoint).
I put the data to the Google Fusion Table. It is a good tool to analyze data.
----------
New charts added.
I've updated the spreadsheet with the run of 2015-05-12:
https://docs.google.com/spread…p4x37M/edit#gid=717997870
Below, the temperature of the calibration (750°C), there is clear NO excess heat.
Above 750°C, we extrapolate the calibration curve. The powers consumed were similar up to 900°C. Above 900°C, the power consumed was greater with fuelled than the extrapolation from calibration.
No conclusion can be done except that a calibration up to 1200°C is needed.
I didn't analyse the cooling phase here but it can give hint of the behavior as analyzed by Ecco below
How did you calculate power? Voltage was supposed to be more or less constant in the ~120V range, but yesterday it was lower than usual, apparently.
I think the last column in the data provided by me356 is supposed to be the PID duty cycle setting.
Good point Ecco, Me356 didn't say to me if I guessed right the name of the columns. If it is PID duty cycle then we can calculate the average voltage = duty_cycle_% / 100 * main AC power. Me, can sort it out so I can update the the spreadsheet?
Thanks for very nice plot!
Basically there is no duty cycle in a correct meaning. Output of the PID is just value for the triac board. This value is sent by PWM and translated back from PWM to next processor that is driving triac.
So basically last column in my data is PWM value that describe how much the triac will be open.
0% is <13, 100% is approx at 88 and more.
What kind of PID board do you use ?
It is implemented in Arduino.
It would be nice to use PID to control the current and second PID to control the temperature.
This will allow much more stable behavior.
Sorry, I wanted to say the TRIAC board of course. I know that the Arduino has PWM output.
It is implemented in Arduino.
It would be nice to use PID to control the current and second PID to control the temperature.
This will allow much more stable behavior.
I have a circuit that I use to watch the voltage drop across a shunt to provide feedback to a modular DC power supply voltage control that keeps it operating in a constant current mode. Would that be of any use to you to try to adapt to your needs?
me356,
I hope your next test will be with nickel like Parkhomov. I got my Raspberry Pi today. It took only one week from China. Tomorrow I hope to get time to boot it up, and install NOOBE 1.4 and see what I can make it do. What operating system do you use on your Raspberry?
Svein
... I got my Raspberry Pi today. ...
Is it the newest number two model? https://www.raspberrypi.org/products/raspberry-pi-2-model-b/
Pictures and comparison at http://www.mikronauts.com/rasp…erry-pi-2-model-b-review/
Me, I've through all the pages of this thread and still I don't find the answer of the Triac model board you are using. So which one it is ?
