Wow!! Your depth of knowledge and skill set is that of someone twice your age. Assuming you are studying electronics at uni, then I think it's fair to say that the first class honours already has your name on it. Another fascinating video, thank you.
Your audio is noisy because your clock frequency is only 29kHz. You probably want to increase it up to 44.1kHz or 48kHz depending on your audio sampling rate.
You are a very knowledgeable person, your videos contain so much information, I am amazed. I only just learned about the difference between BLDC & PMSM only a few days ago, I mean the real thing, not the superficial crap of most TH-cam videos when they refer to both as only Brush less motors! Distinguishing between the Trapezoidal waveform and Pure 3 phase sinusoidal waveform to drive these motors, the sinusoidal which is 100% efficient. I guess the next step is to supply the inverter with power from the mains but going through an home-built PFC (Power Factor Correction) circuit. People would state that its "The rise of the Machines" era, but I think its "Rise of the Inverters" with EVs, AI coming afterwards!
Woohoo go inverters, love them! Yeah I was planning a PFC video but still haven’t got round to it, will probably be a video in the direct current control series.
Great stuff mate !^.^! btw, 99.99% of the time you want an E-stop button and circuit to be a normally closed circuit. that way if any part breaks or any wire comes loose then the product stops, making it so you must fix the loop so it ACTUALLY WORKs when you do actually need it
I know, this was the only red button I had lol. Also the inverter is so light the whole box would just slide away unless the E-stop was on top but then I can't make a lovely equipment stack
It’s a very standard inverter, nothing special. 3x MOSFET half bridges, each FET has an optically isolated gate drive with an isolated DCDC. They’re all controlled from a microcontroller.
For gate drive I would've swapped the position of the cap bank and the Fets. Having your gate drive paths being that long is not always a good idea. Swapping the caps and Fets positions would make your gate drive paths much shorter. Even with it come that way your high current tracks would still be able to be extremely short as well. There's a lot of emi going to be generated which can cause issues with your gates. You can also place decoupling on your main chip directly on the headers on the top of the board, that would work very well.
Oh certainly, this inverter is now more of a "what not to do" for me now. I designed it before I'd been taught any power electronics at uni so it was all just based on stuff I'd seen online or my own assumptions. Though when I was designing it aesthetics were my top priority (don't judge) and I think it looks better this way round 😅. Also if I remember correctly that heatsink hangs down the back of the board a bit so I would've needed a different one to mount it in the middle.
I'm curious did you actually try measuring the back EMF of the aeroplane motor? I've only had the opportunity to measure a handful of these hobby motors, and fully expected to get a trapezoidal waveform out of them, but so far they've all been sinusoidal! In fact I've never encountered a 3 phase brushless motor that has had a trapezoidal back EMF which surprised me. Sensorless control schemes, utilising field orientated control, are actually the pinnacle of brushless motor commutation. Allowing for a perfect alignment of the generated stator magnetic field and rotor field to create the highest torque and the highest efficiency. These can have issues at zero/low RPM, as there's no/very little back EMF to sense, but some schemes continuously inject high frequency signals into the stator coils, then measure the result, in order to accurately detect the rotor position. TIs Instaspin FOC works incredibly well at doing this. At least the older versions did so I can't imagine it's got worse. Positional rotor sensors can be very useful in mixed control systems. With sensored feedback being used at low RPM but then as soon as there's enough back EMF the system switches to sensorless. The other option being to use an open loop start up but these aren't as useful in applications that require full torque from zero RPM. Of course spinning a motor that has trap back EMF, with sinusoidal signals, isn't the best for efficiency. As far as I'm aware you'll lose a couple of % in such a case. However it will be quieter than driving it with trap signals.
I didn't no, that is quite interesting to hear, I have an "in-runner" BLDC that even appears to have distributed windings so I presume that will have fairly sinusoidal back EMF. A video about the inductive zero-speed positioning could be quite interesting🤔.
@20:20 To be safe use for each ouput a BZV55C16 voltage limiter ISA only D6 *or* place a high value resister parallel with the capacitor C12 and C14 (100k or so) else there is a possible voltage shift between +V and -V
Hi, thanks for great video. I have several questions. First, why trace between high voltage capacitors are so thick compared to before and after the capacitors. Second, I want to also build my custom three phase inverter for PMSM, can you tell me some papers or websites or some tips for studying?
Awesome project to begin with! Really love the symmetry of PCB. You must have put a lot of hours into it. One question: I am also building a frequency converter and was wondering how you got that lookup table in the software with the third harmonics (the modulated wave that looks like a saddle-like shape). Did you use some kind of calculator for the lookup table? I would really like it to implement this in my converter. Thank you in advance!
Thanks! I generated the saddle-like wave using a python script I made. I think I did a very over complicated method so as probably not the best to advise haha!
@@electrarc240 Thanks for the reply. In the meantime I have also made a Python program that generates a lookup table with a third harmonic in it. Fourier comes in handy again.
Question... you've used 3 half H-bridges to make a 3 phase output... could you just use 2 half H-bridges using these drivers to run a brushed DC motor?
@@IvandelMuntSoler As this is just for testing motors, I first test the windings with the LCR meter, and also slowly ramp up the modulation index with one of the knobs on the front panel. So the likelyhood of damage is very low
Wow!! Your depth of knowledge and skill set is that of someone twice your age. Assuming you are studying electronics at uni, then I think it's fair to say that the first class honours already has your name on it. Another fascinating video, thank you.
Thank you!!
I read several comments and realized I'm part of the ElectrArc240 fan club that's super impressed by the depth of your knowledge!
There’s a fan club 😳. Thanks!
You've got a knack for explaining things. Great video, appreciate the effort!
Your audio is noisy because your clock frequency is only 29kHz. You probably want to increase it up to 44.1kHz or 48kHz depending on your audio sampling rate.
With better FETs I definitely plan on that
I love the wonky connectors. Just slightly off from the intended position.
It's a must for any real EE.
You are a very knowledgeable person, your videos contain so much information, I am amazed. I only just learned about the difference between BLDC & PMSM only a few days ago, I mean the real thing, not the superficial crap of most TH-cam videos when they refer to both as only Brush less motors! Distinguishing between the Trapezoidal waveform and Pure 3 phase sinusoidal waveform to drive these motors, the sinusoidal which is 100% efficient. I guess the next step is to supply the inverter with power from the mains but going through an home-built PFC (Power Factor Correction) circuit. People would state that its "The rise of the Machines" era, but I think its "Rise of the Inverters" with EVs, AI coming afterwards!
Woohoo go inverters, love them! Yeah I was planning a PFC video but still haven’t got round to it, will probably be a video in the direct current control series.
Great stuff mate !^.^!
btw, 99.99% of the time you want an E-stop button and circuit to be a normally closed circuit. that way if any part breaks or any wire comes loose then the product stops, making it so you must fix the loop so it ACTUALLY WORKs when you do actually need it
I know, this was the only red button I had lol. Also the inverter is so light the whole box would just slide away unless the E-stop was on top but then I can't make a lovely equipment stack
Please continue uploading great videos
I’ll try!
I was waiting for this one... Love it.
Love to see you back 👍, another great video!
Love to see you back 👍, another great comment!
Wow excellent 👌👌
great stuff!
What a nice project. I am considering about doing something like that!!
Thanks! Couldn’t recommend more as a project.
Hi, Mr. I was wondering if it is possible you could share the schematics of that project. I would like to understand better and to do it by myself 🙌🏼
It’s a very standard inverter, nothing special. 3x MOSFET half bridges, each FET has an optically isolated gate drive with an isolated DCDC. They’re all controlled from a microcontroller.
For gate drive I would've swapped the position of the cap bank and the Fets. Having your gate drive paths being that long is not always a good idea. Swapping the caps and Fets positions would make your gate drive paths much shorter. Even with it come that way your high current tracks would still be able to be extremely short as well. There's a lot of emi going to be generated which can cause issues with your gates. You can also place decoupling on your main chip directly on the headers on the top of the board, that would work very well.
Oh certainly, this inverter is now more of a "what not to do" for me now. I designed it before I'd been taught any power electronics at uni so it was all just based on stuff I'd seen online or my own assumptions. Though when I was designing it aesthetics were my top priority (don't judge) and I think it looks better this way round 😅. Also if I remember correctly that heatsink hangs down the back of the board a bit so I would've needed a different one to mount it in the middle.
I'm curious did you actually try measuring the back EMF of the aeroplane motor? I've only had the opportunity to measure a handful of these hobby motors, and fully expected to get a trapezoidal waveform out of them, but so far they've all been sinusoidal! In fact I've never encountered a 3 phase brushless motor that has had a trapezoidal back EMF which surprised me.
Sensorless control schemes, utilising field orientated control, are actually the pinnacle of brushless motor commutation. Allowing for a perfect alignment of the generated stator magnetic field and rotor field to create the highest torque and the highest efficiency.
These can have issues at zero/low RPM, as there's no/very little back EMF to sense, but some schemes continuously inject high frequency signals into the stator coils, then measure the result, in order to accurately detect the rotor position.
TIs Instaspin FOC works incredibly well at doing this. At least the older versions did so I can't imagine it's got worse.
Positional rotor sensors can be very useful in mixed control systems. With sensored feedback being used at low RPM but then as soon as there's enough back EMF the system switches to sensorless.
The other option being to use an open loop start up but these aren't as useful in applications that require full torque from zero RPM.
Of course spinning a motor that has trap back EMF, with sinusoidal signals, isn't the best for efficiency. As far as I'm aware you'll lose a couple of % in such a case. However it will be quieter than driving it with trap signals.
I didn't no, that is quite interesting to hear, I have an "in-runner" BLDC that even appears to have distributed windings so I presume that will have fairly sinusoidal back EMF. A video about the inductive zero-speed positioning could be quite interesting🤔.
@20:20 To be safe use for each ouput a BZV55C16 voltage limiter ISA only D6 *or* place a high value resister parallel with the capacitor C12 and C14 (100k or so) else there is a possible voltage shift between +V and -V
Yes good point I suppose my Zener clamp does nothing for the balance between V+ and V-
Hi, thanks for great video. I have several questions. First, why trace between high voltage capacitors are so thick compared to before and after the capacitors. Second, I want to also build my custom three phase inverter for PMSM, can you tell me some papers or websites or some tips for studying?
"If it can drive an inductive load, it can drive a speaker"
Next up: driving an 18 inch subwoofer with an EV traction inverter 🤣
@@electrarc240 How about that speaker from back to the future?
Awesome project to begin with! Really love the symmetry of PCB. You must have put a lot of hours into it. One question: I am also building a frequency converter and was wondering how you got that lookup table in the software with the third harmonics (the modulated wave that looks like a saddle-like shape). Did you use some kind of calculator for the lookup table?
I would really like it to implement this in my converter. Thank you in advance!
Thanks! I generated the saddle-like wave using a python script I made. I think I did a very over complicated method so as probably not the best to advise haha!
@@electrarc240 Thanks for the reply. In the meantime I have also made a Python program that generates a lookup table with a third harmonic in it. Fourier comes in handy again.
Question... you've used 3 half H-bridges to make a 3 phase output... could you just use 2 half H-bridges using these drivers to run a brushed DC motor?
Yes! And only one of you don’t need to run it in reverse!
Great video.
Can you also show some insights on single phase then 3 phase Solar inverter?
They are fairly similar to traction inverters, just less power density. I will cover traction inverters a lot more soon
Formula Student, ah okay sounds familiar to me 🙂 So great 🙂
Its amazing and how much this inverter costs?
Maybe £100 in parts, thanks!
@@electrarc240 its quite expensive,but great creation ❤️
Where is the current sensing? Where is the position feedback? Why use SVM without any of those?
Did I say I’m using SVM? This is designed for open loop testing mostly though does have provisions for an angle sensor if ever needed
@@electrarc240 Yes at 22:56. There is no current limiter in the inverter then? How do you handle short circuits on the motor side?
@@IvandelMuntSoler As this is just for testing motors, I first test the windings with the LCR meter, and also slowly ramp up the modulation index with one of the knobs on the front panel. So the likelyhood of damage is very low
@@IvandelMuntSoler Ah and I see the SVM part sorry I forgot I said that. You can have SVM without FOC (open loop), which is what I have in this case
you are a very man. Can to provide the PLC code for PLC chip you use, please?
👍👍
What’s a quid?
£