Overload
From HvWiki
Overload
Exceeding the ratings of a device, intentionally or by accident.
Contents |
Why do it
You may have discovered by accident that over-rating something didn't result in its imediate destruction, or someone else has told you that you can run a gizmo at twice the rated with no ill effects. And hey! buying a smaller cheaper component and running it like its expensive big brother saves money. The component may fail earlier through mis-treatment, but if the original specification is 10 000 hours, then most hobbyists would be happy with 100 hours of extra output.
What is the effect
If done carefully, more bang for your buck, but a probable reduction in life of the component. If over-done, the blue smoke comes out
How much can you do it
It depends on the parameter being exceeded
- Voltage
- Current
- Temperature
Over Voltage
More volts across the terminals of a winding of a transformer than specified
Generally NO. DO NOT EVEN THINK about runing a step-down transformer backwards to step up the wall current, it will not work. The core has a maximum flux beyond which it saturates, measured in Volt.Seconds. Usually the transformer is rated to be running at 80/90% of this limit (110% for Microwave Oven Transformers!). Exceeding the volt.second product by 10% means a many-fold increase in magnetising current. Exceeding the volt.second product by 100% means you see the resistance of the windings with very little inductance in series, and a destructive current flows.
Specifically YES. As the Volt.Second product is limited, increasing the frequency allows a pro-rata increase in voltage. If you have the luxury of choosing your operating frequency, then a little more is better. BUT core losses increase with frequency. A 50/60Hz transformer is pretty bad above 200Hz, ferrites also get very lossy. Increased voltage may compromise the winding to core insulation.
More volts from winding to core than expected
Generally Cautiously YES. A commercial transformer will be designed for long life, which means that high voltage surges and transients have been anticipated by the manufacturer. In particular, transformers for mains connection (in the UK at least) must tolerate a 1.5kV flash test primary to core. In particular, Microwave Oven Transformers can be run with 2kV AC between primary and core, and secondary start and core ( so 4kV AC from secondary finish to core) for seriesing secondaries, in air. But that's about it, and it's likely to last longer under oil.
Specifically NSTs, DO NOT lift the grounded centre-tap and ground one pole, trying to get the full 15kV with respect to ground. The 7.5kV winding to core is too much.
More volts on the terminals of a capacitor than specified
Generally NO, but do check the surge specifications. A typical disposable photo-flash cap will be 330v working, 350v surge. A typical polypropylene will be 1000v working, 2000v surge for 2 minutes.
Voltage-reversal is a mode of operation which creates more stress on the dieletric and earlier failure than would be expected. The dieletric "remembers" some of the previous voltage across it and adds it to the new voltage. This is important for AC operation as in Tesla coils, and pulsed operation into inductive loads such as coil guns, coin shrinkers and can crushers, which can reverse charge the capacitors at the end of the pulse. Reduce the initial DC charging voltage if reversal is likely. Stick to AC specifications, not the more generous DC specs.
Rate of failure due to excess voltage is heavily temperature dependant. If you can keep your caps cool, then all will be better. A salt-water cap made from a 2litre pop bottle may take 30kV when cold and fail at 5kV when hot.
More volts across the insulation of a wire
Generally yes, up to 10kV or so. The insulation on a commercial wire is excessively thick for mechanical stress, high temperature operation, insurance against pin-holes. Especially if starting with a large diameter centre conductor to reduce the voltage gradient, a normal wire will take a surprising voltage to break it down, even if it does have 600v stamped on it. Do not rely on the insulation for safety however, only distance will acheive that over 1kV.
More volts across the terminals of a resistor
Resistors with 200v/350v specs are built with no regard to voltage gradients, conducting components in the marking ink etc. Some might take kV, some may fail at only a little above the rating, depending on the method of manufacture. Test a few from a batch and use the others accordingly. High voltage resistors do exist, and are designed to control the voltage gradients.
Over Current
More current from a transformer
For a short while YES. High current of itself does not damage copper wire. High current does not cause any problem in the core, no saturation, no excess losses. BUT, the copper gets hot. The transformer insulation will begin to break down above a certain temperature, generally above the 100°C region, far too hot to touch, so if you can still touch it, you're probably OK. Transformers generally have a continuous (isothermal) rating which depends on their surface area to dissipate heat, and a short-time (adiabatic) rating which depends on their mass to absorb heat by getting hot. If you can establish the adiabatic rating, which has a constant I2t rating for your transformer, then truly collosal currents can be drawn safely, by keeping the duration down. The temperature of the winding can be estimated by measuring the resistance. Copper increases in resistance by 0.4% per degree Celcius, so a 30% increase in resistance of any given winding represents a 75°C increase in temperature.
The practical limit for excess current drawn will be given by the voltage drop that your application can tolerate. Even a short-circuited transformer may take several seconds to reach a dangerous temperature, assuming the breakers hold.
More current from a Variac
The current path includes a brush in the output circuit. This has a much higher resistance than the copper, and is only sprung to the wiping surface. A stationary brush may take a few times the rated current, a moving brush should not be challenged with more than twice.
- More continuous current through/power from a resistor
Yes, as long as the temperature is under control. Add a heatsink or forced cooling. Remember that connections and solder joints age fast at high temperature and can fail before the the resistor. At high temperatures the resistance will change significantly. A mechanically sound joint will continue to work with a blob of liquid solder on it, whereas a joint which relies on solder for mechanical strength may pull apart if it melts.
- More pulsed current into a resistor
Like the transformer, pulsed operation relies on the thermal mass of the resistor to absorb heat, instead of dissipation. There are two types of resistor, which behave differently under pulsed conditions. Homogeneous construction, examples are water, solid carbon, free standing wire, have the entire mass of the resistor being heated and available for storing heat, so have a consistent I2t rating. Heterogeneous resistors, for example wirewound on ceramic former or metal film or ink on ceramic former, have only the conductor to absorb heat for uS pulses, so I2t will be very low. For pulses of seconds duration, the body of the resistor will absorb heat as well. Do not measure the I2t for a 5 second pulse, then extrapolate it to a 5uS pulse
Over Temperature
Generally NO. All devices degrade at a rate governed by a thermal activation energy, that is small increases in temperature above the maximum result in large reduction in lifetime. This especially holds for semi-conductors and dielectrics. If you are happy with their early death, then let them glow cherry-red. If you want to allow more margin for your over-voltage abuses, then keep them cool.
- Pulsed Overtemperature
The adiabatic heating from an overload pulse requires X number of Joules into the thermal capacity of the device to heat it to the limit temperature. This is expressed in two different ways depending on the device. Very low resistance devices - e.g. fuse, wire, SCR when on - express the rating as I2t. This is because the current is determined by the other elements in the circuit, while the element voltage is defined by its resistance and the current. Variable and higher resistance devices - e.g. FET under avalanche from a load inductor - express the rating in Joules. This is because the effective resistance is not constant, so the I2t expression is not proportional to Joules.
Examples
Some simple approximation of how some components will react to overloading.
Light bulb
- Light output is approximately proportional to Voltage3.4
- Power consumption is approximately proportional to Voltage1.6
- Lifetime is approximately inversely proportional to Voltage16
So by increasing the voltage of a lightbulb by 50% we get:
- Light output = 3.97 x
- Power consumption = 1.92 x
- Lifetime = 0.0015 x
So we quadruple the light output, double the efficiency and reduce the lifetime from thousands of hours to a couple of hours. Because of the colour shift the effective light output and efficiency will depend on the particular application.
