Author Topic: Building an ESC  (Read 27918 times)

0 Members and 1 Guest are viewing this topic.

Offline rob77

  • Super Contributor
  • ***
  • Posts: 2085
  • Country: sk
Re: Building an ESC
« Reply #25 on: August 31, 2014, 08:40:37 am »
However I want to scale it up to the 50A. I know I can use a bunch of IRF7832s in parallel to support the 50A current however this can only be DC (from what I understand anyway). I wanted to know if there is another way which would allow true AC. An Op Amp does exactly what I am after, however the max amps is about 1A. Which means I would need a lot of them in parallel, which is not practical. Hence why I am asking this question.

what you mean by true AC and how does an op-amp fit the picture ? what you need is a 3 phase bridge and feed it with 3 phase signal (you can PWM the 3 phases to simulate "true" AC)  generated by a micro-controller or FPGA or a dedicated BLDC controller chip.

By true AC, I mean both positive and negative voltages. I have now been successful in creating a 3 phase inverter circuit and I can used IRF7832s in parallel to handle the 50 amps. However I would like to experiment with a circuit I created using op amps (it works in a simulator), which gives me perfect sine waves phase shifted 120 deg apart, but because op amps can't handle more than 1 A, it won't work without additional components (which BTW is what I am asking for).

how does op amps fit the picture? Because no current flows in or out of its inputs, there is a separation between the low current part of the circuit (powered by the A20) and the high (powered by the battery). Not only that op amps can also handle true AC which transistors can't.

But anyway, it looks like I need to ask on a different forum, because there seems to be a bad culture on this forum. A culture where you only get a responses like "actually it's nearly impossible for a newbie" and no actual answers which I can research further on to find the right solution. In programmers land, most forums never give you "your not good enough" answers (I actually don't know any that do).

don't expect different answers on a different forum ;) why ? you will realize it later on when you'll read your post few years later ;)

if i understand you correctly.... then you wish to create some kind of 3 phase sinewave VCO out of opamps and drive the BLDC with that signal ? that won't work - you have to drive the BLDC according to the position of the rotor's magnetic field (therefore the hall sensors required or the other method - measuring the induced voltage on the wire which is not driven in that very moment).

another point is the current bigger than 1A - have you ever heard about amplifiers ? op-amp + external output stage stage will do the trick (like many of the audio amplifiers with op-amp input stage). but anyways ... driving any kind of motor drawing 50A in peaks with a sine-wave coming out of a B class amplifier.... good luck with that ;)

probably you're confusing the BLDC with a 3-phase asynchronous motor (the squirrel cage one) - the variable frequency 3 phase sine-wave would be able to drive that squirell-cage motor but not a BLDC.
 

Offline ziggyfishTopic starter

  • Regular Contributor
  • *
  • Posts: 113
  • Country: au
Re: Building an ESC
« Reply #26 on: August 31, 2014, 02:01:00 pm »
don't expect different answers on a different forum ;) why ? you will realize it later on when you'll read your post few years later ;)

I look back at some of the stuff I did when I was first learning to code. So yeah, I will look back at this thread, and think maybe that rob77 was right. But for the time being, I like to do things my way :P.

if i understand you correctly.... then you wish to create some kind of 3 phase sinewave VCO out of opamps and drive the BLDC with that signal ? that won't work - you have to drive the BLDC according to the position of the rotor's magnetic field (therefore the hall sensors required or the other method - measuring the induced voltage on the wire which is not driven in that very moment).

There is actually a book that has an example of a 3 phase wave generator using op amps (from memory the book dates back to 1970s). If I am good enough I should be able to take the induced voltage on the wires and integrate it into the sinewave generator. In theory it this should be possible (changing the capacitance of the oscillators, changes the frequency of the wave) but want to play around with it, and will have some fun as well.

another point is the current bigger than 1A - have you ever heard about amplifiers ? op-amp + external output stage stage will do the trick (like many of the audio amplifiers with op-amp input stage). but anyways ... driving any kind of motor drawing 50A in peaks with a sine-wave coming out of a B class amplifier.... good luck with that ;)

Those were the words I was looking for. Now I can do some research on the different classes of amplifiers and find the best way to handle the 50A draw.

probably you're confusing the BLDC with a 3-phase asynchronous motor (the squirrel cage one) - the variable frequency 3 phase sine-wave would be able to drive that squirell-cage motor but not a BLDC.

I have seen people use 3-phase AC with BLDC. Your meant to get better torque from the motor and its meant to be more efficient. Also bigger ESC's (300A, and no, I don't need want to handle that much power) tend to always use AC, so I know its possible.

Anyway thanks for your responses.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16547
  • Country: us
  • DavidH
Re: Building an ESC
« Reply #27 on: August 31, 2014, 04:09:33 pm »
Those were the words I was looking for. Now I can do some research on the different classes of amplifiers and find the best way to handle the 50A draw.

The usual solution is Class-D which just brings you back to PWM.  The power efficiency of Class-D and other switched output designs is difficult to ignore when it directly contributes to high power density.

Linear drive is certainly feasible but you will end up with something rack sized and heavy that will heat your whole room.
 

Offline ziggyfishTopic starter

  • Regular Contributor
  • *
  • Posts: 113
  • Country: au
Re: Building an ESC
« Reply #28 on: September 01, 2014, 03:34:08 am »
Linear drive is certainly feasible but you will end up with something rack sized and heavy that will heat your whole room.

Yeah, some of the bigger ESCs do take up quite a lot of space, some are A5 size and the heat sinks on them are massive. Hopefully I can fit my design on a 5300 X 4050 mill board (same size as the A20 board).

Anyway doing some research this morning, I have found a good way to do it in theory :P. Will buy the components in a few weeks, and start playing around with it.

Thanks for all your help. When I get this thing going I will post the final schematic (got to include other things on the board as well, so don't expect it soon).
 

Offline morpheus

  • Newbie
  • Posts: 9
Re: Building an ESC
« Reply #29 on: September 01, 2014, 03:19:51 pm »
I'm working on sam me thing now, and I actually got spinning rc motor so far. I used 3x 4606 MOSFETs,  and simple MCU to drive them. Just square waves. Check YouTube about how bldc motors works, that helped me a lot. Along with reverse engineering chinese ESC from ali. Next step for me is implementing induction feedback.
 

Offline ziggyfishTopic starter

  • Regular Contributor
  • *
  • Posts: 113
  • Country: au
Re: Building an ESC
« Reply #30 on: September 02, 2014, 12:01:19 am »
I'm working on sam me thing now, and I actually got spinning rc motor so far. I used 3x 4606 MOSFETs,  and simple MCU to drive them. Just square waves. Check YouTube about how bldc motors works, that helped me a lot. Along with reverse engineering chinese ESC from ali. Next step for me is implementing induction feedback.

I've been looking on Google for ESC schematics and there are quite a few out there (seen one for a 60A ESC).

What I am now researching on, is what transistor to use. What I am confused about is the Power Dissipation figure they give you. Like for example the 4606 has a Power Dissipation of 2W. Does that mean that if Vds is 30 and Id is 6, that the heat generated by it will only be 2W?
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16547
  • Country: us
  • DavidH
Re: Building an ESC
« Reply #31 on: September 02, 2014, 12:34:10 am »
What I am now researching on, is what transistor to use. What I am confused about is the Power Dissipation figure they give you. Like for example the 4606 has a Power Dissipation of 2W. Does that mean that if Vds is 30 and Id is 6, that the heat generated by it will only be 2W?

Absolutely not.  The rated power dissipation of 2 watts means it can sustain 2 watts of continuous loss with some recommended printed circuit board pattern at 25C without melting.  That could be 2 volts Vds at 1 amp or 4 Vds at 0.5 amps or 4 Vds at 4 amps with a 12.5% duty cycle or something else.

At maximum continuous rated Ids with full enhancement, power dissipation will be something like 1 watt = 6^2 A * 0.030 Ohms but the rated pulsed drain current is 5 times higher.
 

Offline ziggyfishTopic starter

  • Regular Contributor
  • *
  • Posts: 113
  • Country: au
Re: Building an ESC
« Reply #32 on: September 02, 2014, 01:11:41 am »
What I am now researching on, is what transistor to use. What I am confused about is the Power Dissipation figure they give you. Like for example the 4606 has a Power Dissipation of 2W. Does that mean that if Vds is 30 and Id is 6, that the heat generated by it will only be 2W?

Absolutely not.  The rated power dissipation of 2 watts means it can sustain 2 watts of continuous loss with some recommended printed circuit board pattern at 25C without melting.  That could be 2 volts Vds at 1 amp or 4 Vds at 0.5 amps or 4 Vds at 4 amps with a 12.5% duty cycle or something else.

At maximum continuous rated Ids with full enhancement, power dissipation will be something like 1 watt = 6^2 A * 0.030 Ohms but the rated pulsed drain current is 5 times higher.

That makes perfect sense then. Thanks for that info.

Thanks

Brendan
 

Offline Bukurat

  • Regular Contributor
  • *
  • Posts: 65
  • Country: au
Re: Building an ESC
« Reply #33 on: September 02, 2014, 02:00:31 am »


What I am now researching on, is what transistor to use. What I am confused about is the Power Dissipation figure they give you. Like for example the 4606 has a Power Dissipation of 2W. Does that mean that if Vds is 30 and Id is 6, that the heat generated by it will only be 2W?

When quantifying anything you should include the unit of measure, Vds is 30 and Id is 6 could mean anything. :)

Looking at the data sheet for the AO4046 N Channel..

The maximums are Vds 30V, Vgs 20V, Id 6A@25C ambient temperature or 5A@70C ambient, Idm 30A, Pd 2W, Tj -55 to 150C

These are maximum ratings for each of Voltage, Current, Power Dissipation and Junction Temperature for this mosfet and you can't exceed any one of these.

Further down the data sheet are the parameters you need.
Look at Rds(on) - this will give you the Static Drain-Source on resistance, Typically 25milliohm. If we look at the maximums - If we push 6A through a 25millohm resistor ohms law give us a dissipation of 150mW, well within the 2W maximum.

The 150mW will raise the junction temperature, to find out how much at the thermal characteristics. The thermal resistance Junction to Ambient is typically 74C/W, Max 90C/W  - this means the junction will increase in temperature by  between 74C and 90C  for every Watt dissipated.
The 150mW will raise the junction by a max of 13.5C. Assuming 25C ambient, the junction is at 38.5C so it is well within the rating.

You should redo the thermal calculations when you know where and how this will be mounted.


That should point you in the general direction.







 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16547
  • Country: us
  • DavidH
Re: Building an ESC
« Reply #34 on: September 02, 2014, 03:18:43 am »
Look at Rds(on) - this will give you the Static Drain-Source on resistance, Typically 25milliohm. If we look at the maximums - If we push 6A through a 25millohm resistor ohms law give us a dissipation of 150mW, well within the 2W maximum.

P=I^2R

6 amps through 0.025 ohms is 0.15 volts.  0.15 volts times 6 amps is 0.9 watts.

It is still comfortably within the maximum power ratings.
 

Offline Bukurat

  • Regular Contributor
  • *
  • Posts: 65
  • Country: au
Re: Building an ESC
« Reply #35 on: September 02, 2014, 03:47:39 am »
Look at Rds(on) - this will give you the Static Drain-Source on resistance, Typically 25milliohm. If we look at the maximums - If we push 6A through a 25millohm resistor ohms law give us a dissipation of 150mW, well within the 2W maximum.

P=I^2R

6 amps through 0.025 ohms is 0.15 volts.  0.15 volts times 6 amps is 0.9 watts.

It is still comfortably within the maximum power ratings.

That's what happens when I try and do things my head  :-[
 

Offline ziggyfishTopic starter

  • Regular Contributor
  • *
  • Posts: 113
  • Country: au
Re: Building an ESC
« Reply #36 on: September 02, 2014, 05:21:54 am »
Look at Rds(on) - this will give you the Static Drain-Source on resistance, Typically 25milliohm. If we look at the maximums - If we push 6A through a 25millohm resistor ohms law give us a dissipation of 150mW, well within the 2W maximum.

P=I^2R

6 amps through 0.025 ohms is 0.15 volts.  0.15 volts times 6 amps is 0.9 watts.

It is still comfortably within the maximum power ratings.
Sorry, you confused the hell out of me, until I Googled the actual formula:

V = I * R
P = I * V
Therefore P = I * I * R = I^2 * R

So to confirm my understanding of it. If I pick a transistor that has 0.0021ohm Rds(on) and I have a 50A load. The power loss would be:

V = 50A * 0.0021ohm = 0.105V

Because there is a diode in a transistor, I have to add 0.7V which gives me a total of 0.805V. So:

P = 50A * 0.805V = 40.25W of power

Using a RthJC of 2.0 C/W. The heat generated by the transistor is Max 2 C/W * 40.25 W = 80.5C (or Typical 1.4 C/W *40.25W = 56.35) above Ambient. Given that the current temp in Brisbane is 29C, the actual temp will be 80.5 + 29 = 109.5C Max (85.35 Typical). And if the Max temp for the MOSFET is 150C, I am well within the specification.

And to get the thing to turn on (Vgs = 1.2V), I need a min D and S of 0.805V, and 1.2V on G.

Have I understood all of this correctly?
 

Offline Bukurat

  • Regular Contributor
  • *
  • Posts: 65
  • Country: au
Re: Building an ESC
« Reply #37 on: September 02, 2014, 09:01:32 am »
Look at Rds(on) - this will give you the Static Drain-Source on resistance, Typically 25milliohm. If we look at the maximums - If we push 6A through a 25millohm resistor ohms law give us a dissipation of 150mW, well within the 2W maximum.

P=I^2R

6 amps through 0.025 ohms is 0.15 volts.  0.15 volts times 6 amps is 0.9 watts.

It is still comfortably within the maximum power ratings.
Sorry, you confused the hell out of me, until I Googled the actual formula:

V = I * R
P = I * V
Therefore P = I * I * R = I^2 * R

So to confirm my understanding of it. If I pick a transistor that has 0.0021ohm Rds(on) and I have a 50A load. The power loss would be:

V = 50A * 0.0021ohm = 0.105V

Because there is a diode in a transistor, I have to add 0.7V which gives me a total of 0.805V. So:

P = 50A * 0.805V = 40.25W of power

Using a RthJC of 2.0 C/W. The heat generated by the transistor is Max 2 C/W * 40.25 W = 80.5C (or Typical 1.4 C/W *40.25W = 56.35) above Ambient. Given that the current temp in Brisbane is 29C, the actual temp will be 80.5 + 29 = 109.5C Max (85.35 Typical). And if the Max temp for the MOSFET is 150C, I am well within the specification.

And to get the thing to turn on (Vgs = 1.2V), I need a min D and S of 0.805V, and 1.2V on G.

Have I understood all of this correctly?

You have an unwanted 0.7V there. The diode in the MOSFET is reverse biased in normal operation and doesn't count in these calculations.

You may need more than 1.2V Vgs to saturate the MOSFET you select. Check the curves on the data sheet.
 

Offline hans

  • Super Contributor
  • ***
  • Posts: 1626
  • Country: nl
Re: Building an ESC
« Reply #38 on: September 02, 2014, 09:10:37 am »
Look at Rds(on) - this will give you the Static Drain-Source on resistance, Typically 25milliohm. If we look at the maximums - If we push 6A through a 25millohm resistor ohms law give us a dissipation of 150mW, well within the 2W maximum.

P=I^2R

6 amps through 0.025 ohms is 0.15 volts.  0.15 volts times 6 amps is 0.9 watts.

It is still comfortably within the maximum power ratings.
Sorry, you confused the hell out of me, until I Googled the actual formula:

V = I * R
P = I * V
Therefore P = I * I * R = I^2 * R

So to confirm my understanding of it. If I pick a transistor that has 0.0021ohm Rds(on) and I have a 50A load. The power loss would be:

V = 50A * 0.0021ohm = 0.105V

Technically not power, but that's nitpicking I don't want to go in to. ;)

Quote
Because there is a diode in a transistor, I have to add 0.7V which gives me a total of 0.805V. So:

P = 50A * 0.805V = 40.25W of power

The diode is probably reverse biased, so I don't see how adding 0.7V would help.
Moreover, getting a diode drop of 0.7V at 50A is a very good diode.

You need to multiply 50A x 0.105V = 5.25W loss.

Quote
Using a RthJC of 2.0 C/W. The heat generated by the transistor is Max 2 C/W * 40.25 W = 80.5C (or Typical 1.4 C/W *40.25W = 56.35) above Ambient. Given that the current temp in Brisbane is 29C, the actual temp will be 80.5 + 29 = 109.5C Max (85.35 Typical). And if the Max temp for the MOSFET is 150C, I am well within the specification.

RthJC = thermal resistance Junction-Case. That is not ambient.

You need the RthJA figure, which is the total "series" thermal resistance from junction to a typical use-case of free-standing TO-220 for example. That is usually in the order of 60 C/W.

As a general rule of thumb, assume you can dissipate 1-2W in side a single TO-220 with no heatsink, but at 2W it's really really hot.

If you install a heatsink, you will modify the RthJA figure and can't use it anymore. In that case you need to calculate your own, which is the sum of your heatsink, plus any mounting material (like heating compound or isolation pads), plus the Junction-Case specification.
So a heatsink of 10C/W with perfect contact will have a RthJA of 10+2=12C/W.
In addition when calculating , I would assume that ambient remains constant and work towards the die temperature.  For example:
Tj = Ta + (RthJC + RthHeatsink) * P = 29 + (2 + 10) * 5.25 = 83.5C

Quote
And to get the thing to turn on (Vgs = 1.2V), I need a min D and S of 0.805V, and 1.2V on G.

Have I understood all of this correctly?

The gate voltage is measured against the source voltage voltage.

For a low-side driver the gate is tied to ground. 1.2V (compared to the gate.. so compared to GND) the MOSFET will open very slightly (note: the Rds on is huge and you will blow the arse out of the MOSFET with any significant current).
For a high-side driver N-channel mosfets are usually a bit of a pain, because you need to drive a voltage above the supply voltage (e.g. VCC is 12V, you would need 13.2V for the threshold voltage). A P-channel MOSFET or MOSFET driver are quite useful for that, but you may need to take into consideration if there are any duty cycle limitations.

VDS is a maximum voltage the MOSFET can withstand when it's OFF. If you exceed it, you will kill the MOSFET. The diode that is drawn is more of an "artifact" during the production processes than actually desired, and you need to remember that it's there (i.e. it is not a bi-directional switch) but I personally would not intend on using it.
« Last Edit: September 02, 2014, 09:13:31 am by hans »
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16547
  • Country: us
  • DavidH
Re: Building an ESC
« Reply #39 on: September 02, 2014, 10:26:47 am »
Because there is a diode in a transistor, I have to add 0.7V which gives me a total of 0.805V. So:

P = 50A * 0.805V = 40.25W of power

Power MOSFETs usually have an anti-parallel diode so it will be reverse biased and not in series with the drain to source path.  Bipolar transistors do not have a diode voltage drop either between their collector and emitter.  IGBTs do though.

Quote
And to get the thing to turn on (Vgs = 1.2V), I need a min D and S of 0.805V, and 1.2V on G.

An enhancement mode power MOSFET typically requires somewhere between 5 and 10 volts between the gate and source to fully turn on.
 

Offline ziggyfishTopic starter

  • Regular Contributor
  • *
  • Posts: 113
  • Country: au
Re: Building an ESC
« Reply #40 on: September 02, 2014, 01:17:53 pm »
Technically not power, but that's nitpicking I don't want to go in to. ;)

Yeah I know, however my head was hurting at that stage in the day.

Quote
The diode is probably reverse biased, so I don't see how adding 0.7V would help.
Moreover, getting a diode drop of 0.7V at 50A is a very good diode.

Yeah, I was watching and he says to add 0.7 (starts from 12:00 ).

Quote
You need to multiply 50A x 0.105V = 5.25W loss.

I thought that was the case, it just didn't look right. Because if you take RthJC at 2 C/W then the temp only increases by 10.5C, which just doesn't seem right at all.

Quote
RthJC = thermal resistance Junction-Case. That is not ambient.

You need the RthJA figure, which is the total "series" thermal resistance from junction to a typical use-case of free-standing TO-220 for example. That is usually in the order of 60 C/W.

As a general rule of thumb, assume you can dissipate 1-2W in side a single TO-220 with no heatsink, but at 2W it's really really hot.

Surface mounted ambient is rated at 26 C/W (there is no stat for non surface mounted). So temp would be 136.5 + 29 = crazy hot. So yeah I assumed that I will need a heatsink, hence why I am worrying about this now so I don't destroy any transistors. I also have to take into account that this thing will be powering propellers which are really just massive fans, and the device will also be moving around as well. So temp will be more like 5.25W * 1.4C/W = 7.35C + heat sink. Which isn't to bad.

Quote
The gate voltage is measured against the source voltage voltage.

For a low-side driver the gate is tied to ground. 1.2V (compared to the gate.. so compared to GND) the MOSFET will open very slightly (note: the Rds on is huge and you will blow the arse out of the MOSFET with any significant current).
For a high-side driver N-channel mosfets are usually a bit of a pain, because you need to drive a voltage above the supply voltage (e.g. VCC is 12V, you would need 13.2V for the threshold voltage). A P-channel MOSFET or MOSFET driver are quite useful for that, but you may need to take into consideration if there are any duty cycle limitations.

VDS is a maximum voltage the MOSFET can withstand when it's OFF. If you exceed it, you will kill the MOSFET. The diode that is drawn is more of an "artifact" during the production processes than actually desired, and you need to remember that it's there (i.e. it is not a bi-directional switch) but I personally would not intend on using it.

That clears up a fair bit in terms of MOSFET.

Now I know enough about transistors and power dissipation, I just need to pick one that meet my specs, then start building the thing and testing it. That's where the fun will hopefully start.

Thanks

Brendan
« Last Edit: September 02, 2014, 01:25:50 pm by ziggyfish »
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16547
  • Country: us
  • DavidH
Re: Building an ESC
« Reply #41 on: September 02, 2014, 07:44:19 pm »
Now I know enough about transistors and power dissipation, I just need to pick one that meet my specs, then start building the thing and testing it. That's where the fun will hopefully start.

Based on drawer filled with blown transistors and pieces of transistors and my experience with learning how to design high power switching regulators, I recommend picking up a good face shield as well.  A hockey mask might intimidate the design into working or at least not exploding.  A catcher's mask has a certain amount of style but I would wear eye protection under it:

https://www.youtube.com/watch?feature=player_detailpage&v=0ds0wYpc1eM#t=132

Are you going to start a power transistor failure mode pool?  My vote is disappearance.
 

Offline diyaudio

  • Frequent Contributor
  • **
  • !
  • Posts: 683
  • Country: za
Re: Building an ESC
« Reply #42 on: September 02, 2014, 08:45:46 pm »
Now I know enough about transistors and power dissipation, I just need to pick one that meet my specs, then start building the thing and testing it. That's where the fun will hopefully start.

Based on drawer filled with blown transistors and pieces of transistors and my experience with learning how to design high power switching regulators, I recommend picking up a good face shield as well.  A hockey mask might intimidate the design into working or at least not exploding.  A catcher's mask has a certain amount of style but I would wear eye protection under it:

https://www.youtube.com/watch?feature=player_detailpage&v=0ds0wYpc1eM#t=132



Are you going to start a power transistor failure mode pool?  My vote is disappearance.

 :-DD
 

Offline Kremmen

  • Super Contributor
  • ***
  • Posts: 1289
  • Country: fi
Re: Building an ESC
« Reply #43 on: September 03, 2014, 04:56:03 am »

[...]

It depends on what you call fun. Over the past 2 months I have had so much fun learning about how electronics work, and how sub-circuits work. I also have been successful in building a circuit that can read the 3rd wire and power a small BLDC motor (I am a fast learner).

However I want to scale it up to the 50A.

[...]


Rolling the discussion back a bit. By your own words you are a raw rookie (see above & nothing wrong in that). As others have already pointed out, in your project this (50A) is the point where a) it gets tricky and b) far more than elementary knowledge is required. Fast learning is good, but it won't replace experience. You are trying to build a (relatively) high current PWM driven polyphase modulator - in the entire recorded history no rookie has ever accomplished this. You might be the first one but what are the chances of that...?
Please don't take offense because this is not meant as such. But frankly, you need to aim lower initially. You do have a steep and high learning curve ahead and need to recognize that your confidence springs from ignorance. You know - fools go gladly where angels fear to tread.

Suggestions:
- make an acquantance with the concept called 'hard switching'. This will reveal to you that there are loss mechanisms not computable just by looking at things like Rdson.
- once there, get to know the MOSFET gate and the parasitic elements distorting the ideal behavior. While there, study the evolution of gate charge / voltage in time during a switching event and how the parasitics impact that. That will give you tools to make intelligent device choices and guides you towards the next step.
- study how gate drivers handle the parasitics to control switching losses and spurious oscillations in specific application cases. Some heavy IGBT drivers can be unbelievably complicated for no obvious reason unless you understand - really understand - the above. Luckily 50A is easily doable with a regular MOSFET but not just like that. A good circuit won't happen by accident.
Nothing sings like a kilovolt.
Dr W. Bishop
 

Offline ziggyfishTopic starter

  • Regular Contributor
  • *
  • Posts: 113
  • Country: au
Re: Building an ESC
« Reply #44 on: September 03, 2014, 11:21:32 am »

[...]

It depends on what you call fun. Over the past 2 months I have had so much fun learning about how electronics work, and how sub-circuits work. I also have been successful in building a circuit that can read the 3rd wire and power a small BLDC motor (I am a fast learner).

However I want to scale it up to the 50A.

[...]


Rolling the discussion back a bit. By your own words you are a raw rookie (see above & nothing wrong in that). As others have already pointed out, in your project this (50A) is the point where a) it gets tricky and b) far more than elementary knowledge is required. Fast learning is good, but it won't replace experience. You are trying to build a (relatively) high current PWM driven polyphase modulator - in the entire recorded history no rookie has ever accomplished this. You might be the first one but what are the chances of that...?

I know people with PHDs in software engineering, who can't code to save their lives. And I know people who have never studied at a university that are programming for Google and some have even started multi million dollar companies.

Heck even I was programming professionally before I even finished my first year of my IT degree. Just thinking about my first programming job brings back scary memories, I was given a week to learn a 100,000 line codebase (the code was so unreliable that in the end we had to start the code from scratch and in the end we ended up with about 1,000,000 lines for the finished product) before the developer I was hired to assist decided not to turn up to work again and it took the company 6 months to realise that he wasn't coming back, and on top of that the project was 7 years late when I started (had managers, CEO and CIO breathing down my neck to get it finished). And on top of that there was other projects that just had to be completed before we could complete that project and I also had to preform maintenance on the existing system that implemented in a proprietary language that was no longer supported (and hadn't been supported for 10 years) and the system was built before I was born. And of course the proprietary language wasn't a C, Pascal or even a FORTRAN derivative. Then once I finished those projects, I was given another system to maintain, that to make a change, the change had to be done on 5 servers (dev, uat1, uat2, prod1, and prod2), with no source control. Each change was usually 6 files + 6 db stored procedures. And of course they would always tell you on a friday morning, that they wanted it done before 12pm that day.

So yeah theory isn't everything, its how you use your knowledge and knowing where to find the information that counts. These days, with the internet, you can find information very easily. And if you are stuck on a problem there are many forums out there to give you a hand ;).

I could easily download a schematic of an ESC online somewhere and be done with this ( http://static.rcgroups.net/forums/attachments/2/2/4/4/4/6/a2494477-97-SuperSimple%2050amp.jpg as an example, and isn't that difficult to understand either). However I am not going to have as much fun as doing one myself. Its always fun to come up with your own design and put your own twist on it.

Please don't take offense because this is not meant as such. But frankly, you need to aim lower initially. You do have a steep and high learning curve ahead and need to recognize that your confidence springs from ignorance. You know - fools go gladly where angels fear to tread.

None taken.
 

Offline Kremmen

  • Super Contributor
  • ***
  • Posts: 1289
  • Country: fi
Re: Building an ESC
« Reply #45 on: September 05, 2014, 04:57:08 am »
Years ago i was employed in one of the leading power electronics corps (ABB) designing control systems for all kinds of industrial heavy hardware - paper machines, coaters, continuous casting units, steel mills etc etc. A lot of that was programming various kinds of embedded and regular processing systems, but we also designed all of our own hardware, computers included. In that job i was in a good position to realize how big the difference between nice, deterministically behaving soft/firmware and the fickle, recalcitrant hi speed hi power hardware. Things that seemed to work nicely just failed spectacularly and unreasonably a moment later. People here make much of a TO220 MOSFET blowing up but just wait until you see a 3 kiloamp 2 kilovolt Semikron buck thyristor explode at full power load. That goes beyond nuisance because the shrapnel qualifies as weapon grade. Or see what happens when you stare at the commutator of a 300 kW DC motor when it overspeeds and flashes over. The mini lightning is an unforgettable sight&sound (provided you see or hear anything again, that is).
The point, if there is one, is that you _will_ see spectacular failures once you start the power part in practice. It is just not possible to theorize about the design and then spring a perfect one out from your pen. The number of uncomputable variables is just too large, starting from the parasitics of the physical layout and going from there. But don't let that stop you. The world is full of functioning designs so it is entirely doable, you just need to accept that there is a lifecycle that includes exploratory designing and prototyping before the final design.
Nothing sings like a kilovolt.
Dr W. Bishop
 

Offline ConKbot

  • Super Contributor
  • ***
  • Posts: 1380
Re: Building an ESC
« Reply #46 on: September 05, 2014, 06:23:40 am »
I wouldnt be so quick to dismiss those schematics. Youre going to end up with your 3 phase bridge, mosfet drivers, and something to control it, and a divider for sensorless feedback no matter what youre doing.  The bugger is laying out the board so that parasitics dont kill you, and the code to the micro.  Have it hang, or miscalculate your dead time and youre blowing up fets. Too much gate resistance and youre overheating and blowing up fets. Too little gate resistance with poor routing and youre overheating and blowing up FETs

So get to it and start blowing up some FETs ;)

Do you have a scope though?  It would be rather difficult to diagnose and figure out exactly why youre blowing up FETs if you dont have one.

When you find a limitation to that board that you cant fiddle with or tweak, then you move onto making your own since you know why you need, and what makes up a better one.
« Last Edit: September 05, 2014, 06:28:05 am by ConKbot »
 

Offline ziggyfishTopic starter

  • Regular Contributor
  • *
  • Posts: 113
  • Country: au
Re: Building an ESC
« Reply #47 on: September 05, 2014, 09:27:18 am »
... Things that seemed to work nicely just failed spectacularly and unreasonably a moment later. ...

Yeah, I know that story too well. I once managed to break the homepage of NineMSN, and the update we did passed through 4 QA teams. The page would work for us, but failed outside of our network, in the end it was one of the sysadmins that forgot to configure the system correctly.

I now work as a freelancer developing in house systems. I built a online ad tracking system which sounds easy but trust me it wasn't. Getting the data into the system was the easy part. But because the client wanted to report on everything down to the millisecond, with 20 data points selectable from the front end. getting the data out of the system was the challenge. Try doing a group by on a db table that had 10 billion rows in it with more coming each second. It can take hours to complete, and adding indexes just slowed down the inserts. And on top of that, lookups of ids in that table had to be really fast. We were exceeding the limits of the database engines itself. We ended up using a multi-stage storage no-sql solution.  The system broke because of to many users. We tested it with 20,000 users per second without a problem. When I looked at the logs the system (running on a 20 node cluster) was receiving 100,000 connections per second.

One of the projects I would like to try when I get some free time, is building a 100 node cluster out of A20s. Just to see if it would work.
 

Offline ziggyfishTopic starter

  • Regular Contributor
  • *
  • Posts: 113
  • Country: au
Re: Building an ESC
« Reply #48 on: September 05, 2014, 09:34:32 am »
Quote
Have it hang, or miscalculate your dead time and youre blowing up fets. Too much gate resistance and youre overheating and blowing up fets.

This is why I was am trying to find a solution that doesn't require a micro controller. I know how CPUs can stop working for no reason at all. Using PWM just seems like a hack to me, and hacks are never good. I was looking for an elegant solution.

Quote
Do you have a scope though?  It would be rather difficult to diagnose and figure out exactly why youre blowing up FETs if you dont have one.


Yes I do.

Quote
When you find a limitation to that board that you cant fiddle with or tweak, then you move onto making your own since you know why you need, and what makes up a better one.

Agreed.
 

Offline retrolefty

  • Super Contributor
  • ***
  • Posts: 1648
  • Country: us
  • measurement changes behavior
Re: Building an ESC
« Reply #49 on: September 05, 2014, 10:51:13 am »
Quote
I know how CPUs can stop working for no reason at all.

 Do you seriously believe that?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf