When you say that, are you referring to using them with no heat sink? I was planning on using a hefty heat sink an possibly a fan; would that change anything? looks like the guy in my previous link (using massive heat sink) got away with using them (well he got about 750A out of 10 paralleled before they blew). he was using a different MOSFET but same package.

No, the internal connections of the mosfet 'die' are limited by there physical size. I had a link that I can't find now (from when I first started this quest to learn electronics) that gave total ratings for different mosfet 'packages'. The only link I can find now for you is this one;
http://mcmanis.com/chuck/robotics/projects/esc2/FET-power.html

 

My motor will most likely be a 24V truck starter motor and I will likely run it at 48 to 72V (will have to see what I can get away with).

Starter motors are a poor choice as they are not intended to be used in continuous duty applications. They are very inefficient and designed entirely to produce maximum torque at near stall speeds at the lowest possible price. The life span of the brushes is dismal (maybe 50K-100K revolutions at rated current) and they draw HUGE current!!! A typical 12v starter in a small diesel pickup truck will easily draw 700 amps, and that is at about 8v across the terminals. The big trucks can require as much as 1500 amps to start them. Also, AFAIK, most big trucks these days are 12V. Not sure if they even use 24v starters anymore.

Still, I can't imagine how one of these motors would react, or even survive, if you put 2 or 3 times their nominal voltage across them.

 

The way I found out about the GTO was a couple of years ago a guy at a flea market had a box full of them. He was an electrician in a coal mine and said they were used in the mine. I remembered them when you started your EV posts.

You should really look in to finding a copy of this book; http://www.amazon.com/Rotating-Electric-Machinery-Transformer-Technology/dp/0134096401 Don't get scared when you see the price If you look at the used book price you can get one for under $4 plus shipping. I have it and it gives a LOT of information on all types of motors and their controls. It doesn't give any schematics but gives the different ways ,reasons and even formulas for how AC and DC motors and controls work.

Why do you want to run your PWM at such a high frequency. On a big motor you will spend a lot of time with the mosfets in the 'linear region' where the heat is made. I know you said to keep it inaudible but you work around VFD's at work running at ~60Hz, how much do you heat them? Industrial controls, which are real similar to EV controls if you think about it, while they use the same basic electronic principals are different by the way of there scale. They are heavier amperage wise and size wise. Kind of like the difference between the high 'E' on a guitar and the low 'E' on a bass guitar. The lower PWM frequency would if anything make it sound more like a car.

The mosfet in your post is similar to the ones I'm using in my EDM project (ste53nc50) I got them on Ebay for $7.50 each plus shipping brand new. Digi-key wants $36.75, so it pays to shop.

 

from Wiki: "GTO thyristors suffer from long switch off times, whereby after the forward current falls, there is a long tail time where residual current continues to flow until all remaining charge from the device is taken away. This restricts the maximum switching frequency to approx 1 kHz."

I think I've decided that I don't care so much about switching frequency. This decision is based on noise alone. As shortbus pointed out, industrial drives have an audible switching frequency (IIRC the last drive I read the manual for switched PWM @ 2.5khz to create a 60Hz wave) and it's not unbearably loud. If there is another reason I should care about switching frequency, I don't know it. Is there?

They also have quite different driver requirements.

Yes, and these driver requirements are somewhat shrouded in mystery to me at this point. from the same wikipedia article:

Unlike the insulated gate bipolar transistor (IGBT), the GTO thyristor requires external devices to shape the turn on and turn off currents to prevent device destruction.
During turn on, the device has a maximum dI/dt rating limiting the rise of current. This is to allow the entire bulk of the device to reach turn on before full current is reached. If this rating is exceeded, the area of the device nearest the gate contacts will overheat and melt from over current. The rate of dI/dt is usually controlled by adding a saturable reactor. Reset of the saturable reactor usually places a minimum off time requirement on GTO based circuits.
During turn off, the forward voltage of the device must be limited until the current tails off. The limit is usually around 20% of the forward blocking voltage rating. If the voltage rises too fast at turn off, not all of the device will turn off and the GTO will fail, often explosively, due to the high voltage and current focused on a small portion of the device. Substantial snubber circuits are added around the device to limit the rise of voltage at turn off. Reseting the snubber circuit usually places a minimum on time requirement on GTO based circuits.
The minimum on and off time is handled in DC motor chopper circuits by using a variable switching frequency at the lowest and highest duty cycle. This is observable in traction applications where the frequency will ramp up as the motor starts, then the frequency stays constant over most of the speed ranges, then the frequency drops back down to zero at full speed.

I don't understand all that, and the fact that I don't understand it turns me off. As I go further into the rabbit hole, every question answered presents 10 new questions. I am thinking I want to go with whats been done before and well documented on the internet. Braving uncharted (well not uncharted, but not well documented) territory with these GTO devices may be a dead end (for me anyways, being a noob). From what I've read so far, they don't seem to have take a place in the hobbyist realm, so I doubt I'd be able to find much help in forums with them. I was really excited at first, thinking "wow, one device that handles all the amps and volts I'd ever need, in one small package, no need to fuss around with paralleling anything" - but no free lunch I think.

I would stick to a Mosfet or IGBT with pos temp coefficient.

Are you saying the GTO has a negative temp coefficient? I can't seem to find that data. data sheets like this give tons more info that I don't understand, but I can't seem to find the answer there either.

For the capacitor sizing have a look HERE. Usually the manufacturer already designes and identifies the suitable caps for switching applications, so low ESR, high ripple current, high temperature etc....

Thanks for that. I will need to read up a bit more on capacitors, and I will keep this link handy when I'm ready to pick one.

Starter motors are a poor choice as they are not intended to be used in continuous duty applications. They are very inefficient and designed entirely to produce maximum torque at near stall speeds at the lowest possible price. The life span of the brushes is dismal (maybe 50K-100K revolutions at rated current) and they draw HUGE current!!! A typical 12v starter in a small diesel pickup truck will easily draw 700 amps, and that is at about 8v across the terminals. The big trucks can require as much as 1500 amps to start them. Also, AFAIK, most big trucks these days are 12V. Not sure if they even use 24v starters anymore.

Still, I can't imagine how one of these motors would react, or even survive, if you put 2 or 3 times their nominal voltage across them.

Yes I've read responses like this to DerStrom8's thread about using a starter motor to power a bicycle. I expect to find that you are absolutely right, but I plan to continue with the starter motor anyway because they are in such high supply and cheap. The end goal of this endeavor is not to have an awesome starter powered go-cart, but to learn the skills needed to make an awesome DC motor controller. I don't want to burn up a good expensive motor in the process. an interesting note, there have been awesome starter powered go carts made; how long they lasted, I don't know, but they do work.
http://www.youtube.com/watch?v=C5lp3fP1WKo

The way I found out about the GTO was a couple of years ago a guy at a flea market had a box full of them. He was an electrician in a coal mine and said they were used in the mine. I remembered them when you started your EV posts.

Just curious, did you ever use them? if so, was it a pain?

You should really look in to finding a copy of this book; http://www.amazon.com/Rotating-Electric-Machinery-Transformer-Technology/dp/0134096401 Don't get scared when you see the price If you look at the used book price you can get one for under $4 plus shipping. I have it and it gives a LOT of information on all types of motors and their controls. It doesn't give any schematics but gives the different ways ,reasons and even formulas for how AC and DC motors and controls work.

Cool beans, The 4$ is the 3rd edition. if you click on it, it shows a book that looks like from the early 80's; might have outdated info. I think I'll wait and get the newer edition.

Why do you want to run your PWM at such a high frequency. On a big motor you will spend a lot of time with the mosfets in the 'linear region' where the heat is made. I know you said to keep it inaudible but you work around VFD's at work running at ~60Hz, how much do you heat them? Industrial controls, which are real similar to EV controls if you think about it, while they use the same basic electronic principals are different by the way of there scale. They are heavier amperage wise and size wise. Kind of like the difference between the high 'E' on a guitar and the low 'E' on a bass guitar. The lower PWM frequency would if anything make it sound more like a car.

The mosfet in your post is similar to the ones I'm using in my EDM project (ste53nc50) I got them on Ebay for $7.50 each plus shipping brand new. Digi-key wants $36.75, so it pays to shop.

yeah good point, I'll start checking ebay et. al.

Thanks for all the input guys! I'm going to hit the books now and probably come back with a bunch more stupid questions.

 

Are you saying the GTO has a negative temp coefficient?

No, I have no idea what their temp coefficient is. I was saying IGBT with a pos coefficient so you could parallel them.

A lower frequency is indeed better since you will have much lower switching losses in the switching element (it will however increase losses in the motor). The audible frequency is a problem in factories where hundreds of them are being used and people are working near them like in car assembly lines. Other than that I don't know about disadvantages.

 

Third edition is what I have. I first saw the book as second edition from an EE at work. it did change a lot between 2nd and 3rd. The biggest difference in the two was that the older one didn't have as much on 'solid-state' controls, it had more on contactors and such. Is there a college near you? You could join their "friend of the library" program and get access to books that way. Costs me $20 a year and opens up different research material.

Looking more in the book I saw a way to get full torque and variable speed from a wound field DC motor, that might interest you.

You would open the motor and separate the field from the armature. Then you give full voltage to the armature and PWM voltage to the field. Since the field gets lower amperage than the armature than the field, your speed controller doesn't need to be as robust(read expensive)

I didn't buy the GTO's. When I saw them I thought they were stud mount diodes, and was thinking of building a DC converter for my welder.

 

Third edition is what I have. I first saw the book as second edition from an EE at work. it did change a lot between 2nd and 3rd. The biggest difference in the two was that the older one didn't have as much on 'solid-state' controls, it had more on contactors and such. Is there a college near you? You could join their "friend of the library" program and get access to books that way. Costs me $20 a year and opens up different research material.

good idea, never heard of that.

Looking more in the book I saw a way to get full torque and variable speed from a wound field DC motor, that might interest you.

You would open the motor and separate the field from the armature. Then you give full voltage to the armature and PWM voltage to the field. Since the field gets lower amperage than the armature than the field, your speed controller doesn't need to be as robust(read expensive).

Yeah this rings a bell. We have a handful of 70's louis allis DC motor controllers scattered around my plant. they work by this principle I believe, however instead of PWM I think they use a SCR (or something else; one might use beefy wire-wound pot?) to control the current in the field. The problem I think of first is that if you had full current in the armature and no current (or little current) in the field then you could burn up the motor pretty easily. That is not the case in my plant though, and I need to find out why. Maybe those controllers control current in the armature as well. not sure. or maybe I'm thinking about it wrong. maybe it's totally safe to put full current through the armature with no speed. I do know that those motors in my plant stay stalled out all day, no problem, and you can hear them vibrating (PWM?)

 

Looking more in the book I saw a way to get full torque and variable speed from a wound field DC motor, that might interest you.

You would open the motor and separate the field from the armature. Then you give full voltage to the armature and PWM voltage to the field. Since the field gets lower amperage than the armature than the field, your speed controller doesn't need to be as robust(read expensive)

Ok, I've been reading up on this method, known as field weakening.
here are some links:
http://www.electricmotors.machinedesign.com/guiEdits/Content/bdeee3/bdeee3_1-1.aspx - this one describes a winder, the application I've seen it used.

http://encon.fke.utm.my/courses/see_5433/dc_m.pdf - this one is real good writeup on all kinds of controls. see pg 5-13.
There doesn't seem to be any relationship between current in the armature and current in the field, so maybe current in the armature wouldn't skyrocket if the current in the field went to zero, like I though. However, the current in the armature is relative to the inductance of the armature, so if the speed were zero, inductance would be zero, essentially a short, so maybe it would burn up, as if it were stalled. I'm not sure yet. I will research more tomorrow. not that I'm expecting you to look any of this up; you just threw me bone and I'm chasing it; using this page as a diary. thanks for the bone

 

Why not use an IGBT? You drive it like a MOSFET, but the losses are much smaller and they are available in higher current ratings.

 

When people ask and get answers to motor questions on this and other site about electronics, there usually answered by 'electronics' people. Electrical(read, industrial) use different ways of doing things. I'm not discounting the help here or anywhere else, its just that industrial electronics is a different type of animal.

I'm also pretty sure that most of the stuff on the EV forums is being done at the electronic level instead of the industrial level of thinking. A mosfet or other device the size that is used in an amplifier is more sensitive to abuse than one that is the size of a soda can.

 

When people ask and get answers to motor questions on this and other site about electronics, there usually answered by 'electronics' people. Electrical(read, industrial) use different ways of doing things. I'm not discounting the help here or anywhere else, its just that industrial electronics is a different type of animal.

I'm also pretty sure that most of the stuff on the EV forums is being done at the electronic level instead of the industrial level of thinking. A mosfet or other device the size that is used in an amplifier is more sensitive to abuse than one that is the size of a soda can.

Strantor , if you want the "industrial way" then use SEMIKRON IGBT power modules and their corresponding drivers, at least that's what we used in my last two companies for AC motor drives and UPS's of a few 100kW. Prepare yourself to spend some money on this. Also it's not by coincidence that they used massive heatsinks with fans and BIG capacitors on the UPS bridges. Personally I would not think of driving 1000A with a few Mosfets in parrallel.... Until you have a working and reliable design many months/years can pass.

 

Ok, So incase you haven't seen my other post's I've decide to slow my role and start out with a more "practical" introduction to DC motor controller build. I will be using maybe 5 or 6 MOSFETs in parallel. I am thinking about buying this MOSFET, but it says

Applications
 High Efficiency Synchronous Rectification in SMPS
 Uninterruptible Power Supply
 High Speed Power Switching
 Hard Switched and High Frequency Circuits

Nothing there about motor controllers. Is there any reason why this MOSFET would not be a good choice for a motor controller?

 

I know no reason why it couldn't be used.

There is a document about motor control on the net that contains many useful informations, maybe you have a look at it: http://www.nxp.com/documents/application_note/APPCHP3.pdf

 

I was told elsewhere:

If you check out the spec for Qg you can get an idea of the time to switch (in ns) with 1 amp of gate current. More gate current faster switch time

is this the case? how accurate is that?
Here's my math:
the Qg for my transistor in post #32 is 170nC, so with one amp It would take 170nS to turn on. if I was using a FAN3121 Driver with 11.4A output, it would take 14.9nS. Then if I had 6 in parallel, it would go up to 89nS.
Ok, that sounds good, but I've read potentially conflicting information:

On page 2 of the datasheet, you should see the parameters Turn-On Delay Time, Rise Time, Turn-Off Delay Time and Fall Time. If these are all added up, it will give you the approximate minimum square wave period that could be used to switch this MOSFET: XXXns. This represents a frequency of XXMHz. Note that it would get very hot though because it would spend a lot of its time in the switching over state.

So, is this lumped "Turn-On Delay Time, Rise Time, Turn-Off Delay Time and Fall Time" in addition to the "QG; nC per nS" thumb rule?

 

strantor, did you finally apply the mosfets in parallel? And was that successful? I am in a similar situation-confused if paralleling 6-8 mosfet will work for my design!

 

strantor, did you finally apply the mosfets in parallel? And was that successful? I am in a similar situation-confused if paralleling 6-8 mosfet will work for my design!

I don't remember. it was a long time ago. I know this project didn't go anywhere. Not sure if it was the parallel mosfets giving me trouble or life got in the way.

 

I don't remember. it was a long time ago. I know this project didn't go anywhere. Not sure if it was the parallel mosfets giving me trouble or life got in the way.

If this was a modern thread, nobody would have left you hanging with a question about rise time, fall time, etc. how busy was this site back then? Too busy to answer or too little activity that your thread just drifted by like a tumbleweed in a ghost town?

 

If this was a modern thread, nobody would have left you hanging with a question about rise time, fall time, etc. how busy was this site back then? Too busy to answer or too little activity that your thread just drifted by like a tumbleweed in a ghost town?

I think it was about as busy as today but I could be wrong. I think I was on night shift at the time, so maybe my question went unanswered because it was posted at an awkward time.