HomeMade Motor Controllers

I know that the Victor series motor controller is complex, but i wanted to venture into the unknown, and make something i haven’t made yet(that being hard with how much i have done) and i was wondering what kinda of transistors were used in the Victors to obtain such high current potential, and if anyone knew a function control flow chart for it: IE does it amplify the pwm signal then use it or just use it straight from the source

The transistors used in it, are the FETs or MOSFETs, i was thinking one of those since they do not hold their state for long and need to be refreshed in a cycle (such as in DRAM)?

for the transistors to withstand the curent are they in series?

and for my own should i use a heat sink and a fan or just a fan, beacause if possible i would like to get the TO-220 package so it has some built in surface area for power diapation

and my PCB how do i make a PCB to stand up to about 50 amps or more?



a great place to learn, actually is to read a lot about combat robots. this guy eric, who participates in the northeast robot combat events is very saavy in electronics. heck, he even BUILT his own brushless motor and made the can of it as his spinning weapon. if you would like to browse websites, etc, here are a few.

www.nerc.us (theres a link to delphi forums, thats where you may find loads of indo)

these are a few, hope you find something intresting!

The circuit inside the Victor speed controllers is called a H-Bridge. You can Google H-Bridge and find a multitude of schematics and explanations.

If you’re really industrious, I posted a link a while ago that listed the exact type of FETs used in the Victor speed controller. Realize, however, the Victor uses 9 FETs, while most H-Bridge implementations use 4. That’s why the Victors can handle such a high current. However, it gets very dangerous to handle such a high current.

What are you building this on? Most breadboards use wire in the range of 22awg which is rated for less then an amp.

The Victor does amplify and condition the incoming PWM signal because the incoming signal is pulses between 1 and 2ms every 20ms. The FETs need pulses ranging from 0 to 100% duty cycle. You’ll probably find out more about that from other sites about h-bridges and motor controllers.

Normally I’d suggest buying an IC with the H-Bridge integrated, but I have a feeling you’ll blow off that suggestion.

You might want to check out this place http://www.robot-power.com/ and their OSMC project. Also, they have some parts (like the board) in their web store. I’m betting since it is Open Source, if you e-mail them they’ll be able to answer any questions you have about it.

what is the main advantage of using FETs

what is the difference between FETs and MOSFETs

For info on FETS, IRC a manufacturer, has allot of info on their web site.

Check this site out for a speed controller primer.

There are single chip speed controller chips that will work for small motors and stepper motors. They can all so drive FETS. The following link shows the hardware and software control of one of these chips.

Research to understand speed controls is OK, But as far as making one. Why reinvent the wheel when hobby RC car controllers are quit cheap. Learning how to use speed controllers and stepper drivers in control strategies may be a better place to focus your energies.

A MOSFET is a kind of FET. I’m not well versed in transistor theory, so I’m not sure what you gain my using other FETs over MOSFETS. But the reason MOSFETs are used as opposed to normal bipolar transistors is that you can parallel them. Let’s say I have a MOSFET and a bipolar that can handle 10A each. For my device to handle 50A, I would need to parallel 5 of them. Paralleling the MOSFETs is fine becuase when they heat up, their resistance increases. Let’s say that MOSFET 1 starts to heat up. Its resistance increases and it carries less current. The other 4 would pick up the slack. As 1 starts to cool down, it will carry more current and another would get a break. Of course, it isn’t that linear and it happens very fast, but I think it gives you the idea. Bipolars, on the other hand, will all heat up together. When one of them finally explodes, the others are left with distributing the full 50A load and another will blow even faster than the first one. They will all cascade until you have nothing left but a smoking pile of transistors.

so a MOSFET is like the same thing but only, i guess you would say safer?

what kind of chips or what different chips can I use to interpret the pwm signal for forward and reverse and the correct duty cycles and all that?

Not exactly, MOSFETs and bipolars each have their uses. They work in slightly differnet ways that would suite different applications. In some applications, you would need to use bipolars because MOSFETs won’t work and vice versa. For a high powered motor controller, I think the MOSFETs would be best.

It would depend on what your final application would be. I don’t remember the manufaturer/part# off the top of my head, but there’s a chip to control a MOSFET h-bridge (there’s some subtle timing/voltage issues that need to be taken care of). That specific chip doesn’t do any PWM though. For that, you’d probably have to use a PIC (possibly in conjunction with this h-bridge controller). I don’t know of any commercial ICs that handle hobby radio PWM to motor control PWM. If you can’t find an existing program for a PIC, you’d probably have to write it yourself.

All I remember is that these Metal Oxide Semiconductor field effect transistors in those victors are linked together in parallel, they are high current transistors that either has an N substrate or a P substrate. PMOS or NMOS, I also know that they require something called a MOSFET Driver that delivers short bursts of high current 1- 3 amp of 5- 22 volts to the gates of the MOSFETS, these drivers can be found either inverted or non-inverted so that a +5 could either indicate MOSFET conduct or not conduct. There is a helpful guide on the IRF website that explains how to pick your MOSFET and match it with a proper MOS Driver. Its really complicated…

examples of MOSFET DRivers are the ones from Microchip.com, TC4426 (improved version of TC426), TC426. The two listed are dual drivers, they can drive 2 different signals at one time. You can also find single drivers and Quad drivers as well. These drivers tipically require about two miliamps of current in their source. Its important to have a capacitor at their drain and source terminals. The MOSFETs we see in the Victors each can handle about 20 - 30amps? (ballpark figure…maybe more) with it hooked up in parallel, they are able to source large amounts of current.

It is important that you look at the white pages on the MOSFET specific to your project, they will tell you how much current it is able to handle under what temperatures, proper air cooling with sinks may be needed. also as mentioned above, with the rise in die temp, the resistance also increases, but this shouldnt be a problem if you are using a MOSFET driver. Hope that helps.

That’s why the Victors can handle such a high current. However, it gets very dangerous to handle such a high current.

Well in reality the only danger that you have at such low voltages is from stuff exploding. You really can’t be harmed from shock but even at 36 volts the amount of shrapnel produced by the FETs is amazing when they fail.

Not exactly, MOSFETs and bipolars each have their uses. They work in slightly differnet ways that would suite different applications. In some applications, you would need to use bipolars because MOSFETs won’t work and vice versa. For a high powered motor controller, I think the MOSFETs would be best.

Your forgetting about IGBTs which is an odd mix between an MOSFET and a BJT. Essentially, it turns on like a MOSFET but handles current like a BJT. Also Power MOSFETs are different from regular MOSFETs because there is a diode included . If you look at a circuit symbol for a power MOSFET you will see that diode include.

There is a helpful guide on the IRF website that explains how to pick your MOSFET and match it with a proper MOS Driver. Its really complicated…

It’s complicated to the point where I really wouldn’t recommend such a project if your not out of in college. If your PWM frequency is high enough you may run into transmission line problems. You start running into weird parasitic capacitances/resistances/inductances that cause the gates to ring which will result in the FETs not turning on and off cleanly. The high side drivers can actually latch on which results in the FETs staying on when they shouldn’t be. I spent the last four months with people much smarter than myself trying to do essentially what you want to do but we really had problems.

I think you might have meant 12 FETs. :slight_smile: An odd number of MOSFETs in an H-bridge doesn’t work so well. :wink:

Talk about reviving old threads!
The device used in the Victor is an International Rectifier IRL3103. What makes a MOSFET a better choice for this service is the very low “ON” resistance and the ability to switch inductive loads. (This device is actually a HEXFET, a particular processs for the silicon layers of the device) Since the series resistance is so low (12mohm per device) large currents do not cause the internal temps to rise. This device does contain an internal diode so it’s ability to be used as an amplifier is limited. These diodes are what conduct to turn lights on in the robot when it is pushed with the power off. Current generated in the motors pass through the diodes back to the robot electrical distro. The “H” bridge forms a very nice bridge rectifier when fed in reverse from the motor.
An exploding FET has little to do with applied voltage and everything to do with internal heat. A rapid release of magic smoke is what blows the case apart.
The IGBT (insulated gate, bipolar transistor) shares some charachterisitics of the MOSFET. i.e. it is a voltage driven device (very low drive current) but it does not have the low “ON” resistance needed for this application.
The output switching frequency of the Victor is a tradeoff that allows better speed control for a series of motors that share similar brush/commutator spacing specifications. As a brush DC motor rotates, the brushes switch different windings to the power source. For best control the switched input current should be optimized to this motor switching so that the motor still has good start torgue at low speeds and good speed control at high speeds. Very short pulses vs brush/commutator spacing will not allow the motor to start under load and long pulses will not give any control at high motor RPM. The 884 Victors have a designed output of 120 Hz for this reason. You will note that the ouptut frequency is not the same as the input PWM signal. The internal controller translates the 0-255 modulated PWM signal to a 94 step output signal and provides direction switching as well. Remember that an H bridge operates by turning on only two of the legs at a time for direction control and braking. There is a lot of circuitry inside an 884.
Please be aware that insulated gate and MOS devices still are subject to static induced failures and require some special handling. They are not nearly as sensitive as when first introduced but…

Wow, Talk about reviving old threads. Yea. Well I never actually had the money to play with anything. So No need to post too much info except for the fact that it may help someone else.

I am more into microcontrollers now.


I’m sure that power MOSFETs do have some differences in construction, but that diode is in all MOSFETs; it’s part of the way they’re built. This diode is called a parasitic because it exists, but isn’t there intentionally. Attached is a drawing of a MOSFET with the parasitic diode in red. Note that the diode only exists when the source and body are connected (as it usually is).



Regarding the original post…

Since you mentioned you are more into microcontrollers now, are you looking to build a full-power Victor equivalent, or maybe a smaller version of some kind of pulse-width modulated motor controller for learning and experimentation? There’s some magic going on inside a Victor which converts the 1 to 2 ms pulse into the H-bridge drive, which you can skip if you just want to build an H-bridge and drive it straight out of the microcontroller - more accurately, you can pull that bit into the micro-c and play with it. Instead of ganging up power transistors to get to 50 amp current capability, you can start out with a single transistor in each leg and a smaller motor or lower voltage so you go through less cash replacing blown-up parts.

Assuming, that is, that you’re still thinking about building this thing…

  • Steve Janesch

The main difference between your PNP/NPN transistors and P-/N-Channel (MOS)FETs is that standard transistor are “voltage” effect while Field Effect Transistos are more of “current” effect devices.

FET devices function much like variable resistors… their spreading resistance allows them to evenly share currents unlike transistors which, when paralleled, tend to put much more (or all) of the current on a single device while the others sit idle. A FET, since you can think of paralleling FETS like paralleling resistors of equal value, transmits curret equally between all of the devices in parallel.

One disadvantage of FETs is that they have strange drive characteristics… they are difficult to turn on and off. Thankfully FETs have become much more civilized and inideof most FETs many of the suppression diodes that were once external are now internal… (i suggest IR for FETs, they make the most friendly ones).

IR(InternationalRectifier) also makes some drive ICs which can handle driving your gates on and off for you, saving some circuit knowledge.

Also, using a PIC to drive your circuit helps as well. Many 16/18 series PICmicros incorporate ECCP peripherals which can directly drive (in hardware) the four lines for a full H-bridge.

If you want more detail… just let me know. Hope that answers some ?'s.


Bipolar transistors are current controlled devices while FET are voltage controlled. I think of the FET as having an insulated gate so no current can flow in a FET. Another reminder is a bipolar transistor must have current flowing through a resistor to bias the base on and cause collector current to flow. FETs do function more like a variable resistance but they are easier to parallel becuase of the very low impedances involved. Both transistors and FETs are best paralleled using a small series resistance in series with one of the output elements to balance the currents in each device. As sciguy Phil, pointed out earlier, the construction technique of power MOSFETs (HEXFETS) causes the diode to be a natural (and usable) byproduct of production.

I asked the questions 2 years ago…But thanks. Even though I have learned the answer in college by now.