evilc66 Posted September 21, 2009 Share Posted September 21, 2009 I'm going to try and get this on the scope tonight. We will get this cleared up. Link to comment
mabviper Posted September 21, 2009 Share Posted September 21, 2009 Yeah, sorry about that. It's better to know before you buy it than to know afterwards. Generating an ac signal from dc is fairly complex. Alot of things can potentially go wrong in terms of the circuit, h-bridge, drivers, etc. The software is also pretty complex. Unless somebody makes a working open source circuit and software for this or you're knowledgeable in engineering, specifically in motor control, I would give up trying to make a diy solution for the koralias. Again totally sorry for the bad news. Link to comment
evilc66 Posted September 21, 2009 Share Posted September 21, 2009 Lets not rain on the parade until I can confirm or deny the AC rumors. Link to comment
mabviper Posted September 21, 2009 Share Posted September 21, 2009 I'm going to try and get this on the scope tonight. We will get this cleared up. Please do. I've only found the schematic pic for the controller. I'd like to see a picture of the scope reading as well. Depending on your reference, you might see some like dc-like from 0 to say 12v but the duty cycle will constantly change. I'd <3 you so much if you have a digital scope and can show a triggered time slice of around 10ms. I just want to know what scheme they're using to drive it. If I have time, I might be able to get something going but for now, more information is definately helpful. Link to comment
evilc66 Posted September 21, 2009 Share Posted September 21, 2009 You gonna love me long time then I have a USB scope so I can get you some nice screen shots of the pulse train. I played with the controller for a while after I picked it up and put it on the scope. This is why I'm not convinced that it's an AC square wave. It also doesn't make sense from a hardware cost perspective. Why go to all the trouble of sourcing a low voltage AC motor and designing relatively complex controls for it, when DC motors are dirt cheap and plentiful, and controlling them is about as easy as it gets. Link to comment
robw Posted September 21, 2009 Share Posted September 21, 2009 Please do. I've only found the schematic pic for the controller. I'd like to see a picture of the scope reading as well. Depending on your reference, you might see some like dc-like from 0 to say 12v but the duty cycle will constantly change. I'd <3 you so much if you have a digital scope and can show a triggered time slice of around 10ms. I just want to know what scheme they're using to drive it. If I have time, I might be able to get something going but for now, more information is definately helpful. Where did you find the schematic. I have been searching for info on this for weeks now with no luck, I have found some utube videos of a fellows contoller from Japan but no links of any more info. Tagging along... Link to comment
mabviper Posted September 22, 2009 Share Posted September 22, 2009 Hmm, I might have used the term schematic incorrectly. What I found was a picture of the controller showing the circuit. Anyways, last year or so, somebody from RC took a picture of their controller. That pic showed alot of info with respect to how they're driving the motors. Link to comment
jm82792 Posted September 22, 2009 Author Share Posted September 22, 2009 Well I'm a bozoo, I have no clue what is in the driver, maybe if I googled every part I could understand it. Link to comment
evilc66 Posted September 22, 2009 Share Posted September 22, 2009 Ok, a day late, but here it is This is the output at it's slowest speed that I could get without the controller erroring out. If the max speed is lower than the min, it pitches a fit and shuts the pumps down. The blue horizontal line in the middle (with the arrow on the left) is 0v. Pretty obvious that this is not an AC waveform. Screen is set up for 50uS/div, 5v/div. This shows a 20Khz output frequency with a 15v output voltage, even though the system is stated as 12v. Trailing edge rings a bit with the duty cycle this low, but tightens up as the duty cycle increases. This is max output. As you can see the duty cycle isn't quite 100%. More like 95%. Anyone need to see more? I'd say this wipes all doubt that this will be easily controllable by an Arduino. At a minimum, all you would need is a good FET (J-FET would be a good choice for the high-ish frequency) to drive the pump. An inductor would help smooth the transitions. Maybe I'll give this a shot in the near future. Link to comment
mabviper Posted September 22, 2009 Share Posted September 22, 2009 That's amazing evil ^^. Can you do an adjustment to the first plot and to 1mS/div horizontally? The first plot shows 20k frequency as a carrier to the AC signal, which should be running at 60hz. If you zoom out to 1ms, you should see something completely different. <3 you for having those plots ^^. Link to comment
evilc66 Posted September 22, 2009 Share Posted September 22, 2009 I'll see what I can do. Link to comment
KDris Posted September 23, 2009 Share Posted September 23, 2009 I'll see what I can do. I stand corrected! and am also very happy to admit so because this makes everything 100x easier. Thank you for your efforts evil. Link to comment
jm82792 Posted September 23, 2009 Author Share Posted September 23, 2009 Well you just made me extremely happy. So any schematic on how to exactly use a FET and inductor using arduino's PMW output?, the part#so that I can order the parts and buy a koralia ? http://www.arduino.cc/en/Tutorial/PWM So in my little code I've been working on, keep the duty cycle from like 95% to 10%(or have it just die?)? Sorry reading the scope isn't something I'm good at. Link to comment
evilc66 Posted September 23, 2009 Share Posted September 23, 2009 Here is the basic hookup for a FET. The three connection points are Gate (control pin), Source (incoming power), and Drain (outgoing power). To Pump | |-| DrainArduino--10K-----| Gate|-| Source | +12v Still need to measure the current draw to see what kind of FET is required, but you can easily oversize the FET to about 20A and be safe. I'd keep the pumps spinning in the low cycle. If you stop them, when they restart, they have to have the props hit the stop pin every time to get the rotation right. Beyond the noise being annoying at best, it will wear on the pump a lot quicker. You may have to play with the low speed to see just how slow they will go before they start cogging (can't keep spinning smoothly), or stop all together. Link to comment
robw Posted September 23, 2009 Share Posted September 23, 2009 Thank you very much for the scope pics! A picture is worth a thousand words. And that will put the "mystery" to rest. Now to figure out the programming and circuitry. Link to comment
mabviper Posted September 23, 2009 Share Posted September 23, 2009 Hmm, I would personally hold out on buying a Koralia motor and Arduino to try this experiement. I would still like to see a scope plot at 1ms/div or a bigger horizontal resolution to see the pwm duty cycle change w/ respect to time. Yes, right now it shows a clear DC signal. However, what I saw from the controller pics, I can personally point out to you guys the components to make an H-bridge DC to AC inverter. At this resolution all you see is the frequency of the carrier signal. I also have a Koralia controllable motor at home that I bought from somebody in these forums a while back. A simple dc pwm will not drive it. If you want to know more of what I'm talking about, I've searched out links to read. I tried websites but no luck so PDF files will have to do Just look at the pics/graphs, you don't have to read the whole thing ^^. AC pwm generation 1 Focus on how the DC duty cycle changes with respect to the slower sine wave. AC pwm generation 2 Focus on figures 4 and 8 and some of the experimental graphs they have. Hope everybody knows that I'm not trying to hirder the development of these controllers. I just want to save you guys time, money and a few white hairs. Link to comment
evilc66 Posted September 23, 2009 Share Posted September 23, 2009 I'll try and get the scope back on it today. Maybe at lunch if I'm feeling adventurous. Link to comment
evilc66 Posted September 23, 2009 Share Posted September 23, 2009 Didn't want to see this, but mabviper was right. AC carrier wave. This gets a little more tricky, but not impossible. Link to comment
Schön Posted September 23, 2009 Share Posted September 23, 2009 What about recording the wave of the full program with a soundcard, and then plays it with an mp3 player in a forever loop, and amplifying it to the desired power? I think the arduino hasnt got enough resources to play waves over 20khz in audio format. Link to comment
evilc66 Posted September 23, 2009 Share Posted September 23, 2009 That's a very interesting take on the problem. 20Khz may not be necessary to drive the motors. Hydor may have gone in that direction from a smoothness standpoint at low speed. If we can get a functional circuit working, we can try many frequencies to see what will work with the least amount of processing overhead. Link to comment
Schön Posted September 23, 2009 Share Posted September 23, 2009 The pump produces a heavy inductiv load, and it changes the waveform of the signal. Evil, did you use the pump when you made the measurement (i think yes)? There can be another source of signal errors, the efficiency of the controller, be optimist, and it is unimportant... If someone finds the original signal, maybe the arduino will be able to play it with the digital converter. Link to comment
evilc66 Posted September 23, 2009 Share Posted September 23, 2009 Scope captures were taken without the pumps connected. It's the only way I could do it without puncturing the pump harness. Link to comment
Marian Posted September 24, 2009 Share Posted September 24, 2009 ... what if we use the buzzer code for Arduino ? If I recall correctly, the routine was capable of both time and frequency ??? like "buzz(pin, freq, duration)" The interfacing is gone be tricky though. One will need a SSR PMW capable. cheers, Marian Link to comment
evilc66 Posted September 24, 2009 Share Posted September 24, 2009 ???? Why do you need an SSR? Link to comment
mabviper Posted September 24, 2009 Share Posted September 24, 2009 Hmm, I guess the sound wave could possible~ work. Basically the motor doesn't care what kind of signal you feed it. It just needs an AC signal at 60hz (speculated) with an rms of 12VAC. So if you can make an audio signal at 60hz, amplify it to the right strength, I think it could possibly work. Now, getting an amplifier to do that is the difficult part (I think, I don't have experience w/ audio amplification). If you try to compose an ac from a dc signal using an H-bridge (mosfet or igbt based), what's important is the pwm gating scheme. I know the Arduino is capable of generating pwm signals. With some coding, it can also make a sine wave at 60hz using a carrier frequency of say, 20khz. I don't have experience w/ Arduino personally but I'm sure it can be done. The main problem I see is the potential to short circuit the mosfets together simply because most Arduino chips aren't designed for pwm motor control. If I'm not making any sense, please look at the links I posted earlier. ^^ Link to comment
Recommended Posts
Archived
This topic is now archived and is closed to further replies.