Homebrewed Electricity > Controls
Using Mosfets
JW:
This topic came up recently-
--- Quote ---Quote from: Warpspeed
I had all kinds of problems with exploding mosfets, until I realised what was actually happening, and the cause of my problems were absolutely fascinating.
--- End quote ---
Over the years I have fried several thousands worth of mosfets. I eventually used this on each mostfet output. https://m.littelfuse.com/~/media/electronics/datasheets/tvs_diodes/littelfuse_tvs_diode_5kp_datasheet.pdf UNI directional
When the coil voltage collapses it would burn out the FET. without the transient diode.
Im interested how you protected your Mosfet, and what caused it.
MagnetJuice:
JW, that link didn’t work for me.
I tweaked the search and came up with this:
https://m.littelfuse.com/~/media/electronics/datasheets/tvs_diodes/littelfuse_tvs_diode_5kp_datasheet.pdf.pdf
Is that the one you meant?
Ed
JW:
Yes that's the one the one I'm using, it comes from- NTE its- 5KP100A 1444
Thanks for correcting that link.
These back EMF problems are are hard to manage with Mosfets such as when to use a coil that the field collapses. It took a great deal of selection to locate that diode.
Warpspeed:
--- Quote from: JW on April 21, 2022, 05:09:33 AM ---This topic came up recently-
Over the years I have fried several thousands worth of mosfets. I eventually used this on each mostfet output. https://m.littelfuse.com/~/media/electronics/datasheets/tvs_diodes/littelfuse_tvs_diode_5kp_datasheet.pdf UNI directional
When the coil voltage collapses it would burn out the FET. without the transient diode.
Im interested how you protected your Mosfet, and what caused it.
--- End quote ---
Two main causes of mosfet death, over voltage, and over current. Ignoring for the moment fake counterfeit parts that may not have anything like the "official" maximum rated specification for the part its supposed to be.
Overvoltage usually comes from the unrestrained release of stored magnetic energy in an inductive part of the circuit.
Current builds up in an inductor, and if the current abruptly stops, the collapsing magnetic field generates an essentially unlimited peak back emf voltage. Only two things can limit the final amplitude the voltage spike can reach, either something breaks down, dissipating all the stored energy, or circuit capacitance absorbs all of the inductive energy, and the circuit will resonantly ring.
The instantly released energy can be quite high, so we need to differentiate between a single isolated over voltage event, and something that happens repetitively as a normal part of circuit design and functioning.
If its a single event type of problem, like "Bubba" suddenly disconnecting the load under full power, or a nearby lightning strike for example, a relatively small device can effectively absorb massive power for microseconds or milliseconds, but it must start out stone cold to be able to do so.
Transient voltage supression devices are ideal. They can literally sink hundreds of amps and be no larger than a one watt resistor. The largest I have ever used myself, are the 1.5kW rated 1.5KExx series, where xx is the breakdown voltage. These have become pretty much the industry standard device, and breakdown voltages from 6.8 volts up to 400 volts are readily available. The 5kW rated devices in your link are exactly the same but larger.
The big limitation with all these, are that they are relatively small and can only absorb one big hit at a time. If they are receiving continuous small hit voltage spikes, they get rather hot and can burn up or unsolder themselves rather quickly.
For continuous over voltage protection, some mosfets are avalanche rated. These are specially designed to go into non destructive conduction when the rated maximum mosfet voltage is reached. The big advantage is that the mosfet's heatsink can help a very great deal with getting rid of all the extra heat generated.
This needs to be used with some caution, as the extra heat and power generated by soaking up spikes can burn up the mosfet. Avalanche ratings need to be carefully worked out to do it safely, but if the voltage spikes do not contain a lot of total energy, its one more available solution.
A third method for soaking up continuous not too aggressive voltage spikes is a snubber circuit. This usually consists of a diode, a capacitor and a resistor. If there is for example a high positive voltage spike, the diode can be arranged to divert that into a capacitor. If the capacitor is made sufficiently large, it will absorb most of the energy, without the voltage rising too much. The resistor is arranged to continuously discharge the capacitor. This works well for continuous low energy spikes that are fairly wide apart in time, so the resistor has ample time to fully discharge the capacitor before the next spike.
Over current will also blow mosfets. The heat generated literally melts the silicon, and the device goes dead shorted.
Three factors involved. How high the current, heat flow away from the silicon junction, and time. Its a melting process, so time definitely comes into it.
You can buy very fast acting semiconductor fuses, but they usually cost far more than the device being protected, so are very rarely used. There are applications though where redundancy might require a failed mosfet to be isolated so the equipment can still work, but that would be pretty unusual.
The main factor in mosfet reliability is simply never exceeding the manufacturers maximum ratings.
That is all very greatly misunderstood by many people. The way ratings are presented creates a very false impression of what a mosfet can actually safely do.
A silly example to illustrate.
A car may be able to do 150 miles per hour.
The same car might be able to get 30 miles to the gallon.
So someone thinks he can drive at 150 mph and get 30 mpg, and wonders why its not happening.....
Likewise mosfet maximum ratings need to be thoroughly understood, along with the safe rating curves and especially the de rating curves for safe thermal junction operating temperature.
Another example like the car one above.
Maximum rated voltage 400
Maximum rated current 15 amps
Maximum rated power dissipation 200 watts.
Maximum junction temperature 160 degrees.
So someone hooks it up to 380v and runs it at 10 amps maximum and wonders why it shorted out.
The manufacturers data sheet contains a very great deal of information, if its properly understood, and correctly applied, the device (if its not counterfeit) will almost certainly work very reliably without any nasty surprises.
The absolute maximum ratings are individual never to be reached values, and can never be used in combination.
If you want to push one particular rating to the limit, that is o/k, but you may have to vastly reduce a lot of other things to do it safely.
Mary B:
add in duty cycle...
that 380 volts at 10 amps may be tolerable for 10ms pulse use with a 90ms off time...
Navigation
[0] Message Index
[#] Next page
Go to full version