resonance oscillation in full bridge

Hello, i Have a full-bridge to drive a dc motor. the problem i have is that during a Current level change, i can notice oscillation with frequency in range of 50-100 HZ in the load current (switching frequency is 40KHz).
i guess the reason is the resonance between the Load inductance (12mH) and the bulk capacitor(3mF)
which makes the resonance frequency equal to 1/2pi*sqrt(L.c).
i don't want to use resistors because it makes powerless. also in series with capacitor doesn't suit because it makes the time constant high which is not suitable for other loads connected to the same power supply.

is there any way except having extra resistors to prevent or damp these oscillations? i have read about lossless snubbers but i don't know exactly how can implement them in my full-bridge circuit.
i would appreciate any new ideas. thanks.

here is a schematics of what i have :

the Dc motor is used to adjust a valve . oscillations occur every time i want to change the valve position (by means of changing the dc current level feeding to the motor), the magnitude is sth about 15% of the steady state current and the osculations last about 0.2s to die.
i am using a buck converter to generate a 24V DC bus. the bulk capacitor(3mF) is connected to the output of this converter.

im using a DRV8402 as a full-bridge PWM driver for my load. what it does is turning on and off the transistors diagonally with frequency of 40KHZ. i didn't intend to have a freewheeling path because i liked to have 4 Quadrant full-bridge (the extra energy of the inductance goes back to the 24V bus)
what do you mean by a full featured PWM controller?
is that possible to suppress the unwanted oscillations using a snubber circuit? cause it appears to be only in a short time during the position change of the valve.

In the shown schematic, there won't be any oscillations because the DC bus is driven by a voltage source.

So the first question is, under which real conditions do you notice the said oscillations and with which magnitude?
How are you supplying the DC bus?

Secondly I think, it's primarily a problem of too simple control algorithm. A full featured PWM controller can be expected to correct the pwm duty cycle according to the actual bus voltage, so the motor voltage won't change despite of bus voltage variations. Thus the resonant circuit is virtually cut.

- - - Updated - - -

So we can say, the problem is insufficient or too slow load regulation of the buck converter.

I don't believe that a snubber is required here.

Regarding "full featured pwm controller", I meaned that the pwm duty cycle should be varied to keep the output voltage or current constant instead of setting a constant duty cycle.

one reason that i belve the problem is the LC resonance is that i tried to ad some inductance in series with motor and i could see that the oscillation frequency changed consequently (according to the mentioned formula f=1/2pi*(L.c).

in case of slow regulation of the buck converter, what can i do to solve that problem?

You are not sensing the current in the motor coils…….and that is why your motor inductance is freely resonanting with the bus capacitor…..you should have like a peak current sense in the motor coil current, then that takes a pole out of the LC circuit and it wont resonante…just the same as a current mode full bridge smps output inductor doesn’t resonate with the output cap…because the current on the inductor is controlled…..you should do the same with your motor…….as FvM said, a full featured one will do this for you…..every 25us or so the inductor coil current will hit the current pek threshold…as such the motor inductance is tightly controlled and wont ring like heck.

I was at a company that had the ringing that you speak of, and they didn’t realise…then the gaffers came over from USA and shut us down (rightly)…all 15 in the dept lost jobs

thank you for answers.it seem pretty logical to me and i will try to test it. do you know by chance any full-bridge driver IC which can sense the current in the coil and adjust the PWM accordingly ?

i hope all of those 15 in the dept are doing well with their current jobs now.

Hak99 said:

i am using a buck converter to generate a 24V DC bus. the bulk capacitor(3mF) is connected to the output of this converter.

Click to expand...

A buck converter is not a stiff voltage source, there is no way it can sink any potential over voltage bounce reflected back from your motor.

As others have already pointed out, your motor is only sensitive to current, and provided the over voltage bounce situation is not destructively high, it is not in itself a problem.

Your best bet might be to place a Hall current sensor into the motor circuit, and use that as feedback to control the switching of the bridge.

Examine all the real impedances of your power components and compute the Q of the resonant system

DRV8402, RdsOn = 90 mΩ typ. x 2 in series. = 180 mΩ
Cap 3000 µF 25V ,ESR=20~250 mΩ, depends on size, quality, Voltage rating
Inductor 12mH ?? , ESR=40mΩ to ?Ω depending on rating
Motor Locked rotor Impedance = Rs +jX(f) =Zm(f)
Motor Equivalent Inductance during problem

A step load response to the motor depends on the "%load regulation" or impedance ratio of source/load impedance whenever a step voltage or current or step load is applied.

Thus it is intuitive that the lower the source impedance at the driving frequency, the less effect the load has on the source voltage or current and thus better mechanical stability, at low frequencies, but higher series electrical Q of the driving switcher rate.

This is an inherent challenge to regulate the rate of change of motor phase with position feedback and sensing motor current to prevent oscillations when the source impedance cannot be lowered. This is why VFD's are so popular since the driving force is modulated to match the motor pole frequency with current negative feedback giving a much lower output impedance.

But in my mind a battery for a variable DC motor in parallel with the 24Vdc using PWM , I think would work 100~1000x better than a 3mF cap because of the 50kF capacity of lead acid batteries and mΩ levels of ESR.

Motor position , velocity and acceleration (current) feedback will improve the stability by controlling each with a servo target profile, rather than open loop voltage step. i.e. Accelerate to new RPM, de-accelerate when RMP error is low and lock on to RPM will lower the driver impedance by huge amounts and improve system response time.

If you know a Step Input has infinite fourier spectrum, you can expect any resonances to be excited. But if you input a controlled Ramp, the Frequency spectrum is like a low pass filter and a 2nd order ramp even more. Thus system harmonics are unlikely to be stimulated much. Make sense?

But not having seen the whole design, this is just a guess.