Kiril K.

's profile
image
IT Support

My background is in business administration/economics / social sciences. Over the last 3-4 years, I have drifted more and more towards engineering, guided/goaded by an urge to make useful things that work. As these usually have some sort of electronics in them, I got to rekindle my childhood fascination with circuits.


Kiril K.

's contributions
  • 07.29.2016
  • Circuit delivers constant power to a load
  • Dear Mr. Woodward, Driving the RC combination as you suggest is indeed much simpler. It is similar to configuration b) in the DI which delivers constant power. The main drawback is the uneven distribution of energy in each pulse. Depending on the load this may or may not be a problem. My particular application was a MEMS heater which under normal operation took about 60 mW. Due to the small size the power/volume ratio is about 5 times that of an incandescent light bulb filament. With setup b) and reasonable accuracy (i.e. frequency lower than 2-3 RC time constants), the peak power would be many times higher than the average power. On a micro scale there are positive feedback effects which result in uneven heating and destruction of the device. For this reason I wanted to keep peak power low. I looked further and found a way to obtain much more square-like pulses - at the cost of a more elaborate circuit. The final implementation used discrete components as the 555 timer was not switching quickly enough. For a more robust heater, your suggestion resp. configuration b) is certainly a preferred approach. Higher power can be achieved with higher driving voltage, as the driving period cannot be made much shorter than some number of RC time constants. Thank you for your comment and best regards, Kiril Karagiozov PS I am a great fan of your work. Thank you for the inspiration it has been to me.
  • 07.29.2016
  • Circuit delivers constant power to a load
  • Ok, I guess then it will not work very well. The circuit is meant to provide constant power (to a resistive load); hooking up a loudspeaker implies that one wants some sort of sound to come out. Distortion, dynamic range, efficiency etc. are relevant here and at first glance I don't see how managing the power helps. I have to admit though, I haven't looked too closely at the issues of driving loudspeakers. Kind regards, Kiril
  • 07.29.2016
  • Circuit delivers constant power to a load
  • Hello and thank you! It is an interesting question. My focus was on heaters, but I did run simulations with an inductor in series to see if it could smooth out the power delivery. The result was (as far as I remember) that I could not obtain constant power (because magnetic field energy depends on the square of the current). But as simple as it looks, there are a lot of variations, so probably something could be done. What would be the point to use it with a loudspeaker? Kind regards, Kiril
  • 09.12.2013
  • White noise source flat from 1Hz to 100kHz
  • Built and tested the circuit - works as advertised, as far as I can tell. With a TL072 and a 1N4742A zener it draws 3.5 mA. Noise voltage from the zener is about as described in the original article; my 10x and 100x outputs are not exactly that as I used 1k resistors in the feedback. I had two zeners and I used the first one that I pulled out of the little plastic bag. Looking at the noise, after sampling with an old PC (built-in line input, 192k sample rate @ 32 bit recording the 10x output of the circuit) and processing (54.6 second sample duration, Audacity highest resolution FFT with square window), the FFT is about as flat as you can hope it will be: 8Hz to 20kHz to within a 1db, 3Hz is down 2db. I got a 50 Hz (Europe here) and its first harmonic (both peaks up about 3db) as well. Sampling the 100x output (the line-in actually did not overload, strangely enough, but it is only up 15db from the 10x level) and the fft is flat from 3 Hz to 20kHz. This one worked quite well, thank you sir! PS: Just cannot help but share a little saying I have picked up somewhere ;-) Professionals built the Titanic. Amateurs built the Ark.
  • 09.06.2013
  • Soldering iron driver bridge controls temperature
  • Hey Nikola, well done! I am really glad it worked for you! Thanks for sharing, made me happy to see the LEDs blinking in a familiar pattern :-) A tidy-looking PCB as well. Good stuff.
  • 09.06.2013
  • Soldering iron driver bridge controls temperature
  • Hello Nikola, in theory you should be able to use pretty much any type of switch here. Just keep in mind that when the ST output is low, the power to the heater should be on. The P-channel MOSFET switch does exactly that and this is why I used it here. For an N-channel MOSFET switch you would need to invert the output of the ST. It takes an extra transistor to do that, but it should work. I would not recommend a BJT here. Get a P-channel MOSFET with the lowest Rds(on) you can find, or use an N-channel MOSFET with an inverter... N-channel MOSFETs tend to have slightly lower Rds(on), but then there is that additional transistor... so it is a compromise. If you are unsure how the switches etc. work you will find plenty of tutorials online. Good luck and thank you, Kiril
  • 09.06.2013
  • Soldering iron driver bridge controls temperature
  • Hello Nikola, in theory you should be able to use pretty much any type of switch here. Just keep in mind that when the ST output is low, the power to the heater should be on. The P-channel MOSFET switch does exactly that and this is why I used it here. For an N-channel MOSFET switch you would need to invert the output of the ST. It takes an extra transistor to do that, but it should work. I would not recommend a BJT here. Get a P-channel MOSFET with the lowest Rds(on) you can find, or use an N-channel MOSFET with an inverter... N-channel MOSFETs tend to have slightly lower Rds(on), but then there is that additional transistor... so it is a compromise. If you are unsure how the switches etc. work you will find plenty of tutorials online. Good luck and thank you, Kiril
  • 09.06.2013
  • Soldering iron driver bridge controls temperature
  • Thank you Deloca for your kind comment and the very careful study of the video. What you observed as hysteresis may be the result of the measurement procedure. At equilibrium the heating coil discharges just enough heat to balance the heat outflow to the surroundings so that at the point where the PTC is located the temperature stays constant. There should be no difference whether the equilibrium is reached from above or from below. However: As the iron is loosing heat over its whole area there is a temperature gradient along its length. At the tip, where I am measuring, it is coldest. The exact temperature there depends on the temperature at the PTC but also on the temperature of the surroundings, because that temperature determines the rate of heat outflow. So if you calibrate your set point at 18°C room temperature it will be off (higher) if your room is at 28°C. From a "theoretical standpoint" the explanation would be that as I was warming up the iron there developed a volume of warm, undisturbed air around it (warm room - higher settling point, 233°C at 2'30'' in the video). When cooling, in an attempt to shorten the duration of the process, I blew at the iron, disturbing that volume which did not have sufficient time / conditions to re-establish: cold room - lower settling point (231 °C around 7'20'' in the video). I also remember finding that measurement very sensitive to having a good contact between the sensor and the iron as well as meticulously avoiding any air flow around it, even after wrapping it with the glass cloth. For the video my focus was to document the stability and lack of overshoot of the circuit. Just for the record, the circuit actually would generate a single overshoot when reaching the set point, but it gets evened out by the thermal mass of the iron so it is not noticeable at the tip. Kind regards, Kiril Karagiozov