[epineh]

Quick question (I hope)... I was going to use the IRF540 for this bridge, mainly for the low Rds on value, but I had a look at the specs of the IRFZ44 and it has less than half the Rds value of the 540. While Vds is also half, it is still 55V and I will be using 30V for this setup. It is also half the price :banana: Has almost 20Amp higher current capability as well.

Is this a suitable FET for this app? Is there anything else I am missing or should be looking at ?

OK so that was more than one question and I am sure the answer isn't simply a yes or no heh

Russell.

[kreutz]

Quick question (I hope)... I was going to use the IRF540 for this bridge, mainly for the low Rds on value, but I had a look at the specs of the IRFZ44 and it has less than half the Rds value of the 540. While Vds is also half, it is still 55V and I will be using 30V for this setup. It is also half the price :banana: Has almost 20Amp higher current capability as well.

Is this a suitable FET for this app? Is there anything else I am missing or should be looking at ?

OK so that was more than one question and I am sure the answer isn't simply a yes or no heh

Russell.

There should be no problem using it at 30 volt for an H bridge, if you control the voltage disturbances due to the power lead inductance by means of a low ESR 0.47 to 1 uF film capacitor, rated 200 volt or higher, across +V and -V PCB power traces ( placed as near as possible to the H bridge).

[epineh]

Thanks, will order some today.

Russell.

[acondit]

I see, I also assumed Vcc was +5 volts, now I realize Vcc is also feeding the
Mosfet drivers. It is not necessary for the positive and negative supplies to be exactly paired, since you are banging from rail to rail at the output of the op-amps, you only need the positive rail to be at logic "High" in order to drive the reset pin of the S-R flip-flop. You need the negative supply only because you are using the TL082 chip. Just take into account that in the exact moment when both inputs to the comparator are the same voltage, then the output value will be the difference between both (+ and -) power supplies values, if that value (minus the diode drop) is below the Logic "Low" threshold voltage, there will be no problem.

And yes, once you adjust the voltage divider to the right values, it will work.

Kreutz,

Here is the new version of the current limiter with R10 and R11 changed to reflect the 15v vcc. I changed R10 back to 1K and R11 to 22k. So using the formula (15*1000)/(1000+22000) and get 0.6522 volts. Then dividing 0.6522 volts by .02 ohms = 32.6087 amps. And more important for my servos is using 3 (.1 ohm) resistors you get 0.6522 volts divided by 0.0333 ohms = 19.5856 amps.

So now in the interest of not having to provide a negative supply voltage, I get back to your earlier comment about single voltage supply op amps. Can you suggest a good option (single voltage, dual op-amp, 8 dip package)?

Once this is ironed out I will post the eagle brd and sch files.

Thanks,
Alan

[kreutz]

Kreutz,

Here is the new version of the current limiter with R10 and R11 changed to reflect the 15v vcc. I changed R10 back to 1K and R11 to 22k. So using the formula (15*1000)/(1000+22000) and get 0.6522 volts. Then dividing 0.6522 volts by .02 ohms = 32.6087 amps. And more important for my servos is using 3 (.1 ohm) resistors you get 0.6522 volts divided by 0.0333 ohms = 19.5856 amps.

So now in the interest of not having to provide a negative supply voltage, I get back to your earlier comment about single voltage supply op amps. Can you suggest a good option (single voltage, dual op-amp, 8 dip package)?

Once this is ironed out I will post the eagle brd and sch files.

Thanks,
Alan

I used the TLV2772AIP (dual opamp, Digikey p/n 296-7588-5-ND) with very good results. Note that you must add a protection Schottky diode between the Opamp input (connected to the 100 ohm resistor on the schematics) and ground, cathode to the opamp input. I used the SD103C-T diode.

[acondit]

Kreutz,

Thanks for all the help. Hopefully, this is a "final" schematic. I changed IC5 to the TLV2772 op amp and added schotky diode D9 to provide single voltage supply for the current limiting circuit.

Samco has a large screw connector cap located at Pads 6 and 7, while Mariss uses something like 1000uf 100v cap (user supplied) across the inputs for the G320 and G340 servo drives. For a max 80V 20 amp circuit, is it feasible to use a smaller cap (than the big can) at this point in the circuit (like the Gecko drives)?

Alan

[kreutz]

Kreutz,

Thanks for all the help. Hopefully, this is a "final" schematic. I changed IC5 to the TLV2772 op amp and added schotky diode D9 to provide single voltage supply for the current limiting circuit.

Samco has a large screw connector cap located at Pads 6 and 7, while Marisa uses something like 1000uf 100v cap (user supplied) across the inputs for the G320 and G340 servo drives. For a max 80V 20 amp circuit, is it feasible to use a smaller cap (than the big can) at this point in the circuit (like the Gecko drives)?

Alan

Now you can connect the Over-current LED's cathode to ground.


Talking about capacitors, the idea is to minimize the circuit inductance between the capacitor terminals and the H bridge, any method that provides minimum inductance is good.

Capacitor size depends on desired minimum voltage ripple and ripple current, most of the big size capacitors are rated for ripple at 100 or 120 Hz (for power supply filtering). You will want to use capacitors rated for high frequency ripple (like the ones designed for switching power supplies) near the H bridge. At the H bridge switching frequency (PWM frequency) you won't need a big capacitance to keep the high frequency ripple low (but need low ESR and low inductance capacitors), you also have to make sure your power supply capacitors are adequate to control the low frequency (2 x line frequency) ripple. Adding a 1 uF or higher value Ceramic or film capacitor in parallel to the bigger capacitors will help to filter higher frequencies.

[epineh]

The idea of using a resistor with a diode in parallel on the gate circuit is to speed up the turn off time, a better choice for the gate diode is UF4004 (50 nSec reverse recovery), and use the uF4007 on the boostrap charge position due to its higher blocking voltage.

Hey Kreutz, I know this is always done, but how does it actually speed up turn off time ? I have always been curious but never worked it out...

Cheers.

Russell.

[kreutz]

Hey Kreutz, I know this is always done, but how does it actually speed up turn off time ? I have always been curious but never worked it out...

Cheers.

Russell.

Hello Russell;

Everything is relative, the speeding up of the turn off time is relative to the turn on time. The gate charge path (mosfet turn ON) is through the resistor in parallel to the diode-series_resistor combination due to the diode blocking effect (cathode pointing to the mosfet Driver), that resistor's value is made at least double of the resistor in series with the diode. That resistance is going to limit the flow of charge to the gate, so effectively delaying the turn on time (used to control dv/dt).

When you want to turn off the Mosfet, the output of the mosfet driver goes to zero (or a negative voltage, mostly used with IGBTs) and the gate capacitance discharge path is through the least resistance branch (direct polarized diode-series_ resistor), so the discharge time constant is smaller, and so the turn off time.

It is important to notice; Diode selection is key for the good functioning of the circuit, it should be fast and the reverse recovery charge small, as well as the Vf, in order for this circuit to work as intended. The peak current on the diode will be given by the gate voltage and series equivalent resistance in the discharge circuit. The diode should be able to work reliably at that repetitive pulse current level.

[epineh]

Ah so the gate capacitance disharges back via the driver chip itself then ? I guess I figured it was all one way traffic, wasn't thinking of the driver absorbing that charge.

So I guess my next question is method or formula to calculate the gate resistor(s). I don't want to just give you what driver/FET combination I am using and ask for values, rather if you could point me in the right direction as to how to calculate this then I won't have to hassle you if I change the Driver or FET... though I am sure I will have more stupid questions as I get deeper into this

Cheers.

Russell.

[kreutz]

Ah so the gate capacitance disharges back via the driver chip itself then ? I guess I figured it was all one way traffic, wasn't thinking of the driver absorbing that charge.

So I guess my next question is method or formula to calculate the gate resistor(s). I don't want to just give you what driver/FET combination I am using and ask for values, rather if you could point me in the right direction as to how to calculate this then I won't have to hassle you if I change the Driver or FET... though I am sure I will have more stupid questions as I get deeper into this

Cheers.

Russell.

Look at this post: http://www.cnczone.com/forums/showpo...&postcount=135

The calculated values are only starting values, you need to test them on the prototype and probably change them to get the desired performance.

[acondit]

Now you can connect the Over-current LED's cathode to ground.


Talking about capacitors, the idea is to minimize the circuit inductance between the capacitor terminals and the H bridge, any method that provides minimum inductance is good.

Capacitor size depends on desired minimum voltage ripple and ripple current, most of the big size capacitors are rated for ripple at 100 or 120 Hz (for power supply filtering). You will want to use capacitors rated for high frequency ripple (like the ones designed for switching power supplies) near the H bridge. At the H bridge switching frequency (PWM frequency) you won't need a big capacitance to keep the high frequency ripple low (but need low ESR and low inductance capacitors), you also have to make sure your power supply capacitors are adequate to control the low frequency (2 x line frequency) ripple. Adding a 1 uF or higher value Ceramic or film capacitor in parallel to the bigger capacitors will help to filter higher frequencies.

Thanks for all the help. I didn't really layout the board in eagle. I use Osmond Quartz on my Mac. However, I will post the eagle .sch and .brd files in case someone finds them useful.

Alan

[kreutz]

Thanks for all the help. I didn't really layout the board in eagle. I use Osmond Quartz on my Mac. However, I will post the eagle .sch and .brd files in case someone finds them useful.

Alan

Hello;

Can you post them in pdf? I would like to take a look at them. I use Altium.

Regards,

Kreutz.

[Seb]

I hope I'm not stepping on anyone's toes... Here's the schematic as a PNG image, hopefully that will work for you, Kreutz.

EDIT: Also added a higher resolution picture in a pdf file (to overcome CNC Zone's silly restriction on PNG resolution).

[kreutz]

I hope I'm not stepping on anyone's toes... Here's the schematic as a PNG image, hopefully that will work for you, Kreutz.

EDIT: Also added a higher resolution picture in a pdf file (to overcome CNC Zone's silly restriction on PNG resolution).

Thank you!

[acondit]

Now you can connect the Over-current LED's cathode to ground.


Talking about capacitors, the idea is to minimize the circuit inductance between the capacitor terminals and the H bridge, any method that provides minimum inductance is good.

Capacitor size depends on desired minimum voltage ripple and ripple current, most of the big size capacitors are rated for ripple at 100 or 120 Hz (for power supply filtering). You will want to use capacitors rated for high frequency ripple (like the ones designed for switching power supplies) near the H bridge. At the H bridge switching frequency (PWM frequency) you won't need a big capacitance to keep the high frequency ripple low (but need low ESR and low inductance capacitors), you also have to make sure your power supply capacitors are adequate to control the low frequency (2 x line frequency) ripple. Adding a 1 uF or higher value Ceramic or film capacitor in parallel to the bigger capacitors will help to filter higher frequencies.

Kreutz,

I think I wasn't paying close enough attention when I put the capacitors C13 and C14 on. I put them from MTRA to MTRB. I believe that they should go from V+ to GND.

Alan

[kreutz]

Kreutz,

I think I wasn't paying close enough attention when I put the capacitors C13 and C14 on. I put them from MTRA to MTRB. I believe that they should go from V+ to GND.

Alan

Yes, they should go from V+ to GND.

Regards,

Kreutz.

[epineh]

Anybody have any suggestions for opto's ? I am currently building an interface board to go between my (next) Pluto-p and H-bridges. I am going for total isolation for everything, inputs outputs PWM and encoder's.

And a quick question, are the grounds tied together at the opto's ? I would imagine to get total isolation the grounds would be kept seperate, what is normally done ?

I am closing the gate after the horse has bolted, I powered up my H-Bridges but had the Pluto sending out pwm and dir instead of pwm and not pwm, took out the drive, Pluto and PC (chair) Live and learn as they say, biggest prob is that Pluto's are out of stock until December, should give me time to make some improvements.

Russell.

[neilw20]

Ok here is the latest incarnation, hope everything is where it should be...

Cheers.

Russell.

Pin outs are different on 2104 to 2184 style.
on IR2104
Pin 4 is COM.
Pin 5 is LO

[acondit]

I edited post #52 to replace the zip file with an upgraded schematic. I fixed the connections on the capacitors and revamped the connections a little (removed the -18V connection and hooked two pins to ground).

Alan