Tronic's DIY dual Mosfet H-bridge (DSMhb)

Re: Tronic's DIY dual Mosfet H-bridge_

Hi Erich,

Yes, the IRF3205 mosfet you found looks nice quick fix alternate for the currently used IRFZ44 mosfet. However, as R-eng said, the activation voltage on the gates is a little low and takes some part into the faster heating of the mosfets even if you place the largest one you can find. I personally find the IRF3207 more efficient in our case and might try an h-bridge using them!

I've seen the need for more efficient h-bridge that can be used with much more larger motors than simple wiper motors and I suppose it's time to move on to new design. The robotpower open source h-bridge looks like nice solution. Their complete OSMC board looks nice but its rather expensive for just one motor.

I'll try to redesign one that is cheaper but first I'd like to do some tests with another h-bridge that Roland ( from http://www.simprojects.nl ) sent me and waiting any day now. This h-bridge uses only N-channel mosfets and dedicated driver that sends proper voltage to mosfet gates. I have also a real 24v DC servo motor that is match to it (thanks to Monast) and will allow me to do some tests under load!

I wish I could source much more material to be able to carry as much tests on my own is possible, but even if others sharing any feedback on different equipment of motor sizes is something after all!

:thbup:


Best Regards, Thanos

Re: Tronic's DIY dual Mosfet H-bridge_

I decided to do some tests in measuring actual amperage reaching the motors on my ICE-sim. These motors are 12v, 1.5 hp, and can draw 110A at full load. I am running them off 125 Ah 12v batteries, but will finalize on 24v with two in series. In my earlier posts I have said that the fuses from the battery have blown consistently with 20A limits. And I have burnt out both mosfets and the schottky diodes. So, before increasing the fuse size or damaging any more parts, I purchased a Fluke current clamp as used by electricians to carry out the tests.

My simulator's motions are very demanding from a load standpoint, especially in longitudinal/accel mode. It has worked exceedingly well at 12v for hours on end with the standard DSMhb and no weight onboard. With 100 lbs, it was blowing fuses. Here are my results when the motors are driven directly with NO DSMhb:

12v single battery drive, no load.
-------------------------------------------------
longitudinal/accel = 22-4A avg; 29A peak
longitudinal/brake = 7-8A avg; 9A peak
lateral/left = 8A avg; 9A peak
lateral/right = 12A avg; 14A peak
repetitive longitudinal = 38A peak (rapid accel>brake repetitive movements)

24v dual battery drive, 250 lb load.
-------------------------------------------------
longitudinal/accel = 65A avg; 82A peak
longitudinal/brake = 24A avg; 30A peak
repetitive longitudinal = 95A peak

WOW! Notice I am reaching 95A peaks when the ICE-sim has 250 lbs load in it. And this is probably not the true instantaneous spikes maximum which I am guessing the Fluke cannot resolve.

Even with 50A fuses I would have blown them and probably damaged the DSMhb again. The mosfets I use are rated to 80A. This is why I will attempt to parallel at least 2-mosfets per quadrant, to give me 160A peak potential, or rather, reduce to 50A the current drawn from each mosfet. But this is still a small overhead margin.

In the end I believe I will need the power of 4-mosfets paralleled per quadrant like the OSMC, either as a modified DSMhb with p/n-chn mosfets, or figure out how to interface the DSMhb mosfet drivers to use only the n-chn IRF3207 mosfets.

I've got some work to do!

p.s. Regarding the forces generated. When I changed direction from a full accel rotation (ICE-sim nose fully up) to brake mode, the entire simulator (265 lbs) plus 250 lb load, jumped backwards 2-4 inches!

R-eng

Re: Tronic's DIY dual Mosfet H-bridge_

I'd like to see some of these tests on video if possible!!!!! :happy: Especially the jump that your simulator does when changing direction!! :clap:


Regards, Thanos

Re: Tronic's DIY dual Mosfet H-bridge_

What about using the IRF1405 , 55V; Current, Id Cont:169A;

Re: Tronic's DIY dual Mosfet H-bridge_

N-chn mosfets are not the problem here. I could use a 429A, 24v one if needed. The problem is I have not found a suitable high powered P-chn mosfet. IRF has only the IRF4905 and it is limited to -75A, -55v. The STM I use is rated -80A, -55v, not that the couple of amps makes any real difference, but then I have matched p/n pairs from STM.

That is why using only N-chn h-bridge may be the best solution.

p.s. Rds(on) is very important in this application...IRFB3207 is sweet.

R-eng

Re: Tronic's DIY dual Mosfet H-bridge_

Driving a high current DC Motor using an H-bridge

Shoot Through

It is important to ensure that the two mosfets on the same half-bridge are never on at the same time. This condition called “shoot through”, shorts the power supply and will likely damage the mosfets. The IR2109 half bridge driver chip contains logic to prevent this condition from occurring. It outputs the PWM signal to the high mosfet gate while sending an inverted version of the signal to the low mosfet gate. To be certain that shoot through does not occur during the transient, it also adds a 540ns dead time where both signals are low.

http://hades.mech.northwestern.edu/wiki ... n_H-bridge

Thought this might be worth a read.

Senetor.

Here,s a Rehash of the H Bridge.
Haven't checked properly for errors yet. Mosfets moved closer together so can use one large heat sink, back to back aluminium angle with slots cut in it to try to stabilize the heat from Mosfets. Mosfets insulated with silicon insulating washers, Sockets removed so legs are soldered directly into board ( no dry joints between leg and socket to create extra heat on Mosfet legs)
Mosfets centred to middle of board. Socket for fan gets power from AMC board which is mounted below.
AMC board with onboard regulators for Controller, USB Cooling Fan and LCD Display. FT232RL chip onboard, mounted under board on copper.

Senetor.

H-bridge secrets part 1

http://www.modularcircuits.com/h-bridge_secrets1.htm

Senetor.

The most efficient MOSFET designs use N-channel MOSFETs on both the high side and low side because they typically have a third of the ON resistance of P-channel MOSFETs. This requires a more complex design since charge pump circuits must be used to drive the gates of the high side MOSFETs. However, many integrated circuit MOSFET drivers include a charge pump within the device.

Another method for driving MOSFET-bridges is the use of a special transformer known as a GDT (Gate Drive Transformer), that gives the isolated outputs for driving the upper FETs gates. The transformer core is usually a ferrite toroid, with 1:1 or 4:9 winding ratio. However, this method can only be used with high frequency signals. The design of the transformer is also very important, as the leakage inductance should be minimized, or cross conduction may occur. The outputs of the transformer also need to be usually clamped by zener diodes, because high current spikes could destroy the MOSFET gates.

A common variation of this circuit uses just the two transistors on one side of the load, similar to a class AB amplifier. Such a configuration is called a half bridge. The half bridge is used in some switched-mode power supplies that use synchronous rectifiers and in switching amplifiers.

http://en.wikipedia.org/wiki/H-bridge

Senetor.

Hi Senetor,

The layouts you made look nice! Would be nice to share them here, with others, after you build them and test for no errors first. One thing that concerns me is the use of SMD parts, like the FT232RL. Its very difficult for the average DIY builder to buy proper tools for SMD soldering.

For USB solution with AMC I would prefer the use of ready FT232 adapter interface as the one that Sparkfun makes and we just make the host board with sockets that their breakout boards fits in as I already did for my DIY USB RS485 converter:



Here is a link to their cool breakout board! http://www.sparkfun.com/commerce/produc ... cts_id=718
I already have several of them to use on my future interfaces...


Also I guess you have commercial ISP programmer cause you have removed the ponyprog ISP circuit... :brows:


Regards, Thanos

I just started working together with R-eng in the most efficient DIY h-bridge design up to date that will use only N-channel mosfets (IRF3207) and special gate driver for the signals, the HIP4081AIP !!!

It will be a conbination of the OSMC (robotpower) and my DSMhb to be able to leave the control inputs as they currently are on the AMC1.5 side. Unfortunately (or fortunately enough) it will be meant to be used with atleast 24V motors and very large continuous current rate, more than 50A. I guess that 12v motors can be still used but I'm not sure yet. The cost of it might be a little bigger than the less power DSMhb, since we will be using probably sixteen IRF3207 mosfets for each motor with option to use only eight IRF3207 for middle size motors. Also I guess the size will be twice of that of the DSMhb, so probably each h-bridge will have its own PCB.

Wish us good luck on the sourcing of the parts!!!!


Best Regards, Thanos

I think the trick too using LESS Mosfets is a larger common heatsink, like what is used in large audio amplifiers. The heatsink literally sucks the heat from the mosfet devices to keep there heat output stabilised. The heatsink has to be able to remove move heat than the mosfet can create, otherwise it can overheat momentarily and Kapoof it's all over red rover. It's a bit like the cooling system in a car, if the radiator can not handle the heat output of the engine, the engine overheats and dies.
I'm no electronics engineer, just a hobbiest with enough knowledge of electronics to be dangerous, just like the rest of us here on this forum.
So by using a mosfet gate controller so that the gate is either open or closed to reduce heat buildup and using bigger and better heatsinking to remove and stabilise the heat produced by the mosfets is going in the right direction to getting a design that will not fail.

Senetor.

I was browsing the data sheet of the HIP4081A and noticed that it cannot handle shoot-through, which kills mosfets. Here is the first chapter quoted from the data sheet.

The HIP4081A is a high frequency, medium voltage Full
Bridge N-Channel FET driver IC, available in 20 lead plastic
SOIC and DIP packages. The HIP4081A can drive every
possible switch combination except those which would
cause a shoot-through condition. The HIP4081A can switch
at frequencies up to 1MHz and is well suited to driving Voice
Coil Motors, high-frequency switching power amplifiers, and
power supplies.

Senetor.

Would this driver be suitable
L6206 DMOS DUAL FULL BRIDGE DRIVER

Re : DIY builder to buy proper tools for SMD soldering.
No special tools are required for a lot of SMD chips. Take a look at this video guide on YouTube. There are also many other guides there too.
http://au.youtube.com/watch?v=e5qYG95bbz8

Senetor.

Here's another SMD Tutorial. http://au.youtube.com/watch?v=V_cDV92IuWY

Senetor.

And another http://au.youtube.com/watch?v=4V7tBPsECjc

It's all in the flux and using a big enough angled tip on the soldering iron to keep the solder hot enough so it will flow properly.

http://au.youtube.com/watch?v=t06malVew40

Senetor.

@ Senetor

If your cut-and-paste is accurate, then perhaps you are not understanding the sentence you are quoting.

The HIP4081A can drive every possible switch combination except those which would cause a shoot-through condition.

Read this carefully. It says that it will provide every switch combo, but NOT (except) those which would cause a shoot-through.

Excerpts from the datasheet also say...

The startup sequence turns on both low side outputs, ALO and BLO, so that the bootstrap capacitors for both sides of the H-bridge can be fully charged. During this time the AHO and BHO gate outputs are held low continuously to insure that no shoot-through can occur during the nominal 400ns boot-strap refresh period. At the end of the boot strap refresh period the outputs respond normally to the state of the input control signals...

To provide accurate dead-time control for shoot-through avoidance and duty-cycle maximization, two resistors tied to pins HDEL and LDEL provide precise delay matching of upper and lower propagation delays, which are typically only 55ns...

the lower inputs ALI and BLI dominate the upper inputs. That is, when one of the lower inputs is high, it doesn’t matter what the level of the upper input is, because the lower will turn on and the upper will remain off.

I believe others before us at the Robot Power commmunity have thoroughly vetted this device and did not employ its use in the OSMC if it could result in an uncontolled damaging condition without extreme desire to do so...Not limited to robotic combat the OSMC has been used for many other industrial, and research tasks at such intstitutions as MIT, Ford Motor Company, and Los Alamos National Labs.

Furthermore, I believe most everyone here understands the need to control thermal gains with heatsinking. Especially on a power device which can generate enough current to arc weld with!

With respect to SMD soldering, I agree it does not take special equipment or talent. I have done much of it myself with little effort and just as much success. Although I do recommend using a magnifier lamp fixture to more clearly see what one is doing, and oh yes, very steady hands!

But for most people, using through-hole devices for a diy project, represents a far easier task, less prone to mistakes. We have the luxury to add a little extra size to our projects, which are already fairly small.

R-eng

@ Senetor

Your suggestion of using the L6206 DMOS Dual Full Bridge Driver proves to be unsuitable to our stated needs. This is not only a driver for an H-bridge, but it also has a full (no breakout) H-bridge included within the device.

2.8A DC of peak output current is not what we need to drive even the small wiper motors used by some. This is in the second line of description within the datasheet. The block diagram is even more explicit of this devices true intentions.

R-eng