Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

over voltage during disconnection.

Status
Not open for further replies.

ivenzar

Member level 2
Joined
May 28, 2015
Messages
46
Helped
3
Reputation
6
Reaction score
3
Trophy points
8
Activity points
361
Buck is connected to a load thorugh long cables. When load gnd pin suddenly disconnects when converter is working, large voltage spikes are produced due to paristic wire inductance. Currents of 260 Amps and 10V flows through the coils before load gnd pin disconnection.

Someone knows a way of minimize that large voltage spike? (it is not possible to slow down the disconnection process)
 

It's not just wire inductance, there's also that the whole
loop is probably "wound up" with I*R offset and when I
goes to zero the converter is still pumping at a high duty
cycle, which takes time to bleed off.

More shunt C will soak that up, faster control loop, an
overvoltage clamp external to the converter loop are
all options to explore. Suggest you get closer to the
real action to determine how much if cable L*dI/dt
and how much is the converter control loop (like do
abrupt load-dump testing with and without the cable,
so you know what is the converter's own fault).
 
It's not just wire inductance, there's also that the whole
loop is probably "wound up" with I*R offset and when I
goes to zero the converter is still pumping at a high duty
cycle, which takes time to bleed off.

More shunt C will soak that up, faster control loop, an
overvoltage clamp external to the converter loop are
all options to explore. Suggest you get closer to the
real action to determine how much if cable L*dI/dt
and how much is the converter control loop (like do
abrupt load-dump testing with and without the cable,
so you know what is the converter's own fault).


1.- Assuming, control is fast enough. Voltage spike comes from the wires. Best solution to protect the converter against sudden load disconnection?.

How much C shunt should I include? 1/2cv^2=1/2li^2? which voltage should I use in previous equation?
Is there another solution that not requires adding extra capacitance? -> converter response results slower. Voltage spikes are kVolts order.

2.- If problem is mainly caused by slow control and I cannot increase it (disconections is much faster)... what can i do?

Thanks!
 
Last edited:

Is your converter a type that requires some minimum load? Suppose you were to permanently keep a light resistive load attached, so that the converter does not become unstable when you detach your primary load?
 
Is your converter a type that requires some minimum load? Suppose you were to permanently keep a light resistive load attached, so that the converter does not become unstable when you detach your primary load?

Long wire that connects load gnd and converter gnd may be suddenly disconnected on each side. So no resistive load can be included, efficiency also must be taken into account.
If wire is disconnected at converter side, we think a spark will occur damaging the converter due to high energized inductive load. If the wire is disconmected at load side, converter will face high voltage transient.
Long wires with high currents :(
 

1.- Assuming, control is fast enough. Voltage spike comes from the wires. Best solution to protect the converter against sudden load disconnection?.

Seems like a thing that shouldn't be based on assumptions.

Voltage spike may come from the output inductor and di/dt
forced by the load. If the voltage exceeds VIN then it's an
inductive flyback. Whether this is the wires or the power
inductor wants checking. You ought to be able to figure
the value of each.

For getting a Cshunt value, try a simple SPICE simulation
with an inductor - capacitor - ipulse parallel circuit, the
ipulse initial value at your 260A, stepping to 0A in (say)
100nS, the inductor having the realistic values from the
converter and the capacitance initially using the converter
output filter values. You'll see a flyback voltage that the
C soaks up somewhat; more C should lower this and you
can work the series until you find happiness.

You may want to look at the disconnect scenario and see
if you can make it so hot always disconnects first, if
ground "comes apart" then maybe some of what you see
relates to measurement reference point. If you are worried
most about converter self-protection then make all of the
measurements relative to converter local ground plane.
Where the remote end swings to, may not matter as much
(and what is the load end's "perceived transient", relative
to its local ground, now distinct from converter ground)?
 

I don't think the stored energy in the wire inductance can harm the converter. Remember that the wiring inductance will basically act to slow the di/dt in the output, which if anything should be more gentle to the converter output than an instantaneous drop to zero. So long as the impedance at each connector is low, the voltage spike will appear on the wire side of the spark. So the spark may be a safety hazard for whoever is handling the wires, and it may damage the contacts in the connectors.

Some consideration may be given to the effect of the disconnect on common mode voltages, not just differential voltages. But this requires much more detailed info on how everything is connected.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top