H-bridge 1 phase gate driver optocouplers in series vs in parallel

[eneuro]

Hello,
Just designing another simple low power 1 phase H-bridge controler and optocouplers diodes (1.2Vf and 1.65Vf) need to be trigered from 5V MCU two pins for positive and negative output phase voltages from power rail, so I need triger optocoulers in left high side and low right side (and vice verse) at the same time, so could do it easy with one resistor in series with two optocouplers, but can do it also in parallel configuration as shown below with additional resistor.

How do you think, which method will be better for driving those optocouplers in the terms of switching timings?

I can use additional 3rd MPU pin to precharge high side gate driver bootstrap caps if needed and probably will add transistors on pin outputs, while wouldn't like to go close to MPU output pin maximum currents.

[T3sl4co1l]

Won't make a difference in timing as long as the currents are right.

[eneuro]

Yep, was thinking also how temperature changes can affect both serial vs parallel configurations without precision resistors
Anyway, I might need third MPU pin for additional PWM/break features, so parallel configuration with 4 resistors like below should do the job in simple 1 phase H-bridge where only one MPU PWM timer needed 

Small PCB MPU to 1 phase H-bridge optocouplers converter was designed to do not mess with this circuit schematics, so it is time to test this with real power rails

[Niklas]

What are the maximum reverse voltage ratings of the LEDs?

[eneuro]

It depends which optocouplers you will use
Cheap
PC817 has 6Vr,
4N25, TLP250,TLP250, TLP351f ... ?  5Vr, so we need to make some tricks if 5Vcc MPU is used 
While I like this 3 pin MPU driving can add another 1N4148 in series or one 1N4148 and transistors to drive those 3 pins MPU outputs at safe lets say 4.3V<5V, while depending on MPU and other tasks it has to do, I wouldn't like stress too much MPU pins with currents from direct optocouplers leds in parallel.
Anyway I've checked my MPU datasheet and its pins should survive even shorting both leds by assident

Made today fast prototyping using classic oldschool methods and have nice development single layer PCB with classic red LEDs which were driven from 9V battery via 2.2k resistors and it looks like this trick with 3 pins  driving those 4 H-bridge optocoupler leds works 

Need to check all posible 8 input combinations if is it possible to short one side of this H-bridge by assident, but probably not-I mean, assuming proper timing of mosfets gates 

Update: With NPWM (NOT PWM) 1 (high) not possible to short circuit any side, but with NPWM 0 (low) temporary shorts possible, but this is not desired sequence and disabling PWM (1 high in this case) allows choose (set) POS or NEG first than add NPWM if needed.


BTW: However it is interesting that in this datasheet of 5mm THT LED:
http://optoelectronics.liteon.com/upload/download/DS20-2007-0003/S_110_LTW-2S3D8.pdf
they wrote something like this and no Vr value? 
8. Reverse voltage (VR) condition is applied for IR test only. The device is not designed for reverse
operation.
5VR for 5mm THT yellow led is here:
http://optoelectronics.liteon.com/upload/download/DS-20-99-0254/S_110_LTL2H3KSK.pdf

This is not designed to drive classic leds, but optocouplers, but it looks like that in the term of this 5Vr voltage of some optocouplers input diodes it can make sense drive them in series rather than in parallel with tricky reverse voltages to switch some of them off when only 3 MPU pins are used to drive them in simple H-bridge shown above with brake and bootstrap precharging modes.

 

[eneuro]

Wrote a few lines of C code to find all possible leds outputs based on POS,NEG,NPWM inputs:

pos: 0  neg: 0  npwm: 0  pwm: 1
 d11: 0  d21: 0
 d12: 0  d22: 0

pos: 0  neg: 0  npwm: 1  pwm: 0
 d11: 0  d21: 0
 d12: 0  d22: 0

pos: 0  neg: 1  npwm: 0  pwm: 1
 d11: 0  d21: 1
 d12: 1  d22: 0

pos: 0  neg: 1  npwm: 1  pwm: 0
 d11: 0  d21: 1
 d12: 0  d22: 0

pos: 1  neg: 0  npwm: 0  pwm: 1
 d11: 1  d21: 0
 d12: 0  d22: 1

pos: 1  neg: 0  npwm: 1  pwm: 0
 d11: 1  d21: 0
 d12: 0  d22: 0

pos: 1  neg: 1  npwm: 0  pwm: 1
 d11: 0  d21: 0
 d12: 1  d22: 1

pos: 1  neg: 1  npwm: 1  pwm: 0
 d11: 0  d21: 0
 d12: 0  d22: 0

Not possible in steady 3 input conditions to short circuit any side of H-bridge, so we only have to pay attention for proper transitions between those 8 possible inputs -> led outputs and optimize for type of optoisolated gate drivers used based on possible different driver power consumption in Hi/Lo outpt state, so it can be important in the case of bootstrap caps used in high side, etc. 

[eneuro]

Anti parallel diodes can be used to bypass low reverse voltage of some optocouplers, but higher curents need to be switched on optocouplers input side 
Anyway it can be an option if we do not like put additional diodes in series with optocoupler inputs 

Wrote a few lines of code on AVR based on logic table shown above and it looks like that everything works as expected-development software version of my MPU to H-bridge optocuplers PCB slowed down to frequency easy to watch by humans shows this idea of driving H-bridge with pre charging high side gate driver bootstrap caps needed during restart, based on RC constant of gate driver bootstrap capacitor and current limiting resistor which with 100uF caps and 100 ohm resistor is about 0.03s
Bootstrap caps precharging is in 5s-20s frame , then classic H-bridge positive/negative switching with cross conducting N-channel mosfets.

Everything using only 3 MPU wires and a few lines of AVR code 
Time to connect this to real H-bridge with botstrap capacitors and see how it performs in real application, where I need to drive high inductance solenoid coil