Extremely high zener current in oscillator

[Zak28]

Whats a good remedy to the high current going thru both zeners? Circuit is ~3.9hz biphasic, making ac to load.

Traces:


Circuit:


Greatly appreciated.

 

[c_mitra]

Usually zeners are used in the reverse bias mode but that is not the case in the present circuit. If you are talking about D2 and D3, they will take the full current and very little will go into the load. This is as expected.

 

[Zak28]

I guess what I wanted was a better implementation to this circuit. Im pretty sure zeners can be omitted and somewhere caps will be required.

 

[Easy peasy]

have a decent look at your ckt, as soon as M1,M2 as gated on, there is no limit to the current in D2,D3 in their forward direction. It is a bit of a nonsense ...

 

[Zak28]

Limiting zener current with resistors doesn't work, cant figure this out. Id rather have it (if possible) with no diodes and only caps.

 

[KlausST]

Hi,

Your power circuit simply is wrong. It can't work this way.

--> The usual circuit for the desired function is a "full bridge".
I can't see how a capacitor could replace the diodes.

Klaus

 

[Zak28]

Apparently bridge works well with nmos (I was using pmos) however power supply sinks ~108ma for a 2k load which is too much for a battery powered circuit. What topology can I implement over present one for lower power consumption?

Power supply current:



Greatly appreciated.

 

[KlausST]

Hi,

My recommendations:
learn basics. How to drive a load with a full brige has been discussed many thousand times - here in this forum and in the internet. Every mosfet manufacturer provides application notes about this.
Don't (try to) reinvent the wheel. Gain from the pitfalls others already solved.

There are ready to buy halfbridge and fullbridge ICs,
There are ready to buy Mosfet driver ICs.
There's a good reason:
Because every self made circuit suffers from the same problem: crow_bar_currents caused by cross_conduction.

Klaus

 

[Zak28]

A pwm controller or a gate driver is useless for this circuit, my issue was missing diodes.

 

[d123]

Hi,

The 4 diodes are supposed go from 0V to the load and from the load to V+, on either side of the load connections and the MOSFET junctions. It appears in practically every H-Bridge schematic.

While it sounds valiant and confident to retort about gate driver's being useless in this circuit, it also shows your ignorance, especially after Klaus explained it to you in a previous answer. I can promise you from vast real experience - home-made gate drivers are absolute rubbish and ridiculous when proper gate drivers take up less space and cost cents and have needed high-side drive charge pump or similar or bootstrapping and maybe protection features, and stuff, you know. But you don't because it is clear you have read almost nothing about half and full bridges and nothing at all about gate drive requirements. I'm sorry to be blunt.

Maybe look in LTSpice at the source resistance. Ideal sources (no resistance) turn safe circuits into the false flag current spike rollercoaster from Hell. I've seen hair-raising kiloAmp inrush or pulsedspikes in some simulations, until I added a realistic (power) source resistance.

Good luck.

 

[Easy peasy]

d123 your comments are over the top, the OP is starting out, a little latitude, not to say politeness, goes a long way...

 

[Zak28]

...home-made gate drivers are absolute rubbish and ridiculous when proper gate drivers take up less space and cost cents and have needed high-side drive charge pump or similar or bootstrapping and maybe protection features, and stuff, you know. But you don't because it is clear you have read almost nothing about half and full bridges and nothing at all about gate drive requirements...

I infact do not know much about gate drivers and afaik they modulate gates with high frequency when my circuit demands a "flat" current train. Also I reckon the driver would not be suitable for a thru hole only board (unless its soic6-8 cause I have a breakout board for that).

...turn safe circuits into the false flag current spike rollercoaster from Hell. I've seen hair-raising kiloAmp inrush or pulsedspikes...

You just described one of LTSPICEs most annoying features I would add another thing is when it takes forever to run a simulation requiring to amend circuit. Overall the circuit Im working on consumes ~170mW which isnt too bad, I will find ways to improve it.

Thanks for your input.

 

[KlausST]

Hi,

A gate drive is not a pwm controller.
A gate driver does not modulate anything.
And the most popular gate drivers are in SO8 package.

Klaus

 

[BradtheRad]

The 555 IC itself contains a half-bridge output. It both sources and sinks current.
By adding a half-bridge, in effect you create a full H-bridge.



The load receives symmetric true AC square waves.

 

[Zak28]

The load receives symmetric true AC square waves.

Its true however Id be limited to ~18v or whatever limit a 555 has. I need ~36v. Seems like you really like falstad.

 

[FvM]

Full bridge topology is the obvious solution. But if the load impedance is really as high as 2k, circuit design becomes relaxed.

A full specification should be the starting point.

 

[KlausST]

Hi,

A clean full bridge for this load could be made with analog switches. Like two SPDT.
= very fast, very symmetric, no cross conduction problem, easy to drive, cheap, low part count

Klaus

 

[Zak28]

By analog you meant mechanical? Because for instance ADG888 cannot go to ~36v. Please hint at some part numbers.

Full bridge nmos and bjt simulations yielded VERY low DS and CE voltages and Im talking very low gate charge, low turn on(5v) fets and very high gain bjts yielding in low load current. I believe switching is my next attempt. Thank you for the generous suggestion.

 

[KlausST]

Hi,

By analog you meant mechanical?

No.

Because for instance ADG888 cannot go to ~36v.

Then choose another one.

Please hint at some part numbers.

Each analog switch manufacturer and almost every distributor has interactive selection guides. They are free to use.

Klaus

 

[Zak28]

I looked into it and the most attractive was ADG5436F however I would require 2 totalling to ~$20 so I believe a comparator is my next best bet since full bridge topology suffers from inadequate voltages. Comparator is cheaper in long run because I would not need ltc6995 and employ precision caps/resistors for oscillation.