Layout Review of H-Bridge

[sabu31]

Hi all,

I need suggestions regarding improving the layout or if its fine.
I have made a basic current source full bridge.
The input is 24V DC, 50A. The purpose is to switch at around 20kHz - 50kHz for a parallel resonant circuit. The resonant part is external to board.

 

[Easy peasy]

where did you get the idea for the series diodes up there with the top fets ?

--- Updated Nov 18, 2022 ---

I would put some serious TVS's across the fets or across each totem pole of fets - as if you have to turn off in a hurry - the energy needs to go some where ....

 

[sabu31]

Its current source inverters and not voltage source inverters

 

[FvM]

Its current source inverters and not voltage source inverters

The series inductor is shorted by diodes. No working current source scheme.

 

[sabu31]

The series inductor is shorted by diodes. No working current source

Should i use a zener diode and diode with common anode connection across the inductor.
I am attaching MIT thesis page for induction heating. Here, its using low power diodes across inductor, however, the current through indcutor will be around 20 A. Should diodes be rated same as inductor current.

Apart from that, is the layout configuration for the bridge appropriate.

 

[cupoftea]

PCB Layout of SMPS....3 principles..

1.....Go and return together....especially of the power switching current loops
2......SWitching nodes kept away from sensitive signals
3......Keep all current loops as narrow in area as possible....especially those loops carrying switching power current
4.......do not route switching (high di/dt) currents through control grounds.

A good tip also is to remember the "rule of squares"........then see the sqrs in your copper......eg a square of 1 oz copper is approx 0.5mR...from one side to the opposite....THis is how you roughly calc your copper trace/pour resistance......squares

 

[Easy peasy]

you can tell that paper fig. 3-6 is complete cods-wallop, as, diodes D1, D2 prevent mosfet turn off.

 

[BradtheRad]

Should i use a zener diode and diode with common anode connection across the inductor.

Resonant action can be obtained from a cross-biased H-bridge. Try a few arrangements testing where to tap for bias voltage and/or current. Inductors have an ability to cause oscillations as growing current is detected until a threshold is reached. Biasing networks switch back-and-forth. The inductor receives bipolar AC.

R1 is installed at a position that it generate a voltage proportional to Amperes going through it. This voltage is used to turn transistors On and Off.

To reduce spikes my simulation has snubbers, back-to-back zener diodes across the inductor. (It works in simulation although it may not be optimum for hardware designs).

Adding capacitors (in strategic locations) can alter switching frequency.
Capacitors can also change the shape of waveforms to be more sine-like. These tend to carry high current peaks thus it could be a good idea to assemble a gang of capacitors.

My simulation illustrates a concept and is not necessarily the optimum design for real-life hardware.

 

[sabu31]

Thanks for the reply . However, what I want to know is

a) What could be done in component placement/ layout for improved design.
b) The input inductor is not resonant inductor. It is for making a current source. What should be the diode/clamp set up across the diode or can we operate the system by using only the inductor(without clamp/etc).

Thanks

 

[sabu31]

Thanks for the reply . However, what I want to know is

a) What could be done in component placement/ layout for improved design.
b) The input inductor is not resonant inductor. It is for making a current source. What should be the diode/clamp set up across the diode or can we operate the system by using only the inductor(without clamp/etc).

Thanks

In continuation with the above post. How will this component positioning compare to the earlier in post 1. I am finding the routing difficult though in this configuration.