Re: Back EMF
Posted by
caudlet
on 2004-03-21 06:54:12 UTC
--- In CAD_CAM_EDM_DRO@yahoogroups.com, "afogassa" <afogassa@y...>
wrote:
Most zeners are rated for 1/2 to 5W. You will need to dissipate
possibly 10 times that amount for a short period. Semiconductors are
not like magnetic and passive components that can take large over
currents for a few seconds and recover. A zener is a diode that
starts conducting at the rated voltage. Once it turns on it will
pass "infinite" current (well actually limited by some factors but a
lot more than it's rated for). In circuits where the impedence
(series resistance) is high, i.e. you are not dealing with high
currents a zener can suppress transients and clamp low current
overvoltage. To use one to try and limit the over voltage that would
be present at the filter cap of your supply would mean it would have
to dump a lot of stored energy. In that case you need a "shunt
regulator" which shunts the overvoltage to ground. The design is
pretty straight forward in that you use an amplified zener which
consists of a zener in the base circuit of a large power transistor
across the offending voltage. Dump that into a low ohm current limit
resistor in series with the power transistor and bingo you have the
power to dump the overvoltage.
There is a circuit in the files section from Mariss to do just that.
Voltage transients will destroy semiconductors in milliseconds
(nanoseconds in some cases). Inductive loads are great generators of
voltage transients. Turning an inductive load on and off rapidly
generates voltage spikes that can be many times the applied voltage,
and no small amount of RF noise. The circuits have suppression
diodes built in to offer the noise/spikes a low impedence path to
ground. Make sure that you have a good ground path from the
switching source (A3977 power ground) to the circuit power ground (-
on filter cap). It only takes one bad connection or a few inches of
higher resistance run to allow a voltage transient to reach component
killing levels.
In building your own motor controller the physical layout and
attention to grounding and noise suppression are as important as the
applied voltages. Depending on the value of the power supply filter
it would take a lot of back EMF to drive 30 volts up over the 35V
spec and if you monitor the filter cap voltage with a scope you can
see the DC levels change.
Dealing with switching/power electronics is a dance between
efficiency (instant turnon and zero losses) and the suppression of
huge amounts of voltage and EMI/RFI. To do it right you need
extremely fast "snapshots" in time like from a high frequency storage
scope with the ability to trigger at a preset voltage point. A high
frequency current probe is also a useful tool.
I have the current probe and a 350MHZ scope. It would have to borrow
or rent one of the newer digital storage scopes to do a design.
I bought a set of Geckos and spent my time on the linear motion
aspects of the project.
wrote:
> HI, Allto
> I'm building a A3977 based step motor controller, it works ok up
> 30V. than it burns up, I think it's the Darn thing called "Back EMF"the
> I'm thinking about using a zener diode on the VBB input to clamp
> high voltage generated by the motor. what do you think ?I think your zener will burn up followed closely by the A3977.
>
> Thank's
> Fogassa
Most zeners are rated for 1/2 to 5W. You will need to dissipate
possibly 10 times that amount for a short period. Semiconductors are
not like magnetic and passive components that can take large over
currents for a few seconds and recover. A zener is a diode that
starts conducting at the rated voltage. Once it turns on it will
pass "infinite" current (well actually limited by some factors but a
lot more than it's rated for). In circuits where the impedence
(series resistance) is high, i.e. you are not dealing with high
currents a zener can suppress transients and clamp low current
overvoltage. To use one to try and limit the over voltage that would
be present at the filter cap of your supply would mean it would have
to dump a lot of stored energy. In that case you need a "shunt
regulator" which shunts the overvoltage to ground. The design is
pretty straight forward in that you use an amplified zener which
consists of a zener in the base circuit of a large power transistor
across the offending voltage. Dump that into a low ohm current limit
resistor in series with the power transistor and bingo you have the
power to dump the overvoltage.
There is a circuit in the files section from Mariss to do just that.
Voltage transients will destroy semiconductors in milliseconds
(nanoseconds in some cases). Inductive loads are great generators of
voltage transients. Turning an inductive load on and off rapidly
generates voltage spikes that can be many times the applied voltage,
and no small amount of RF noise. The circuits have suppression
diodes built in to offer the noise/spikes a low impedence path to
ground. Make sure that you have a good ground path from the
switching source (A3977 power ground) to the circuit power ground (-
on filter cap). It only takes one bad connection or a few inches of
higher resistance run to allow a voltage transient to reach component
killing levels.
In building your own motor controller the physical layout and
attention to grounding and noise suppression are as important as the
applied voltages. Depending on the value of the power supply filter
it would take a lot of back EMF to drive 30 volts up over the 35V
spec and if you monitor the filter cap voltage with a scope you can
see the DC levels change.
Dealing with switching/power electronics is a dance between
efficiency (instant turnon and zero losses) and the suppression of
huge amounts of voltage and EMI/RFI. To do it right you need
extremely fast "snapshots" in time like from a high frequency storage
scope with the ability to trigger at a preset voltage point. A high
frequency current probe is also a useful tool.
I have the current probe and a 350MHZ scope. It would have to borrow
or rent one of the newer digital storage scopes to do a design.
I bought a set of Geckos and spent my time on the linear motion
aspects of the project.