It is well known that the voltage drop across a pn junction reduces with increase in temperature. Thus, is it correct to say that a hotter diode will have a higher efficiency?
I think (open to correction).. Within limits yes ( in design however the diode is modeled as Vthreshold + id*Rd) so we can perhaps model Rd as decreasing with temp....but I think we should keep in mind type of cooling..
1. In natural cooling to some extent this takes place automatically as tjunction
= Ploss*zthermal +ambient ( it would be difficult to vary zthermal dynamically)
2. In forced cooling it may be possible to adjust Zthermal dynamically.
3. The fact that reliability and life even for electronics depends on keeping
tjunction as low as possible needs also to be considered .. In industrial
appliances this maybe more important .. the best s to choose special low loss
diodes...
Cheers
ith increased current the temperature is increased but also the voltage drop that is V=Rd*id, so to heat up tyhe diode without increasing the voltage drop you should cover it with and thermal isolator, but then if the outside temperature is increased you will have a breakdown.
You could also use a heating resistor controlled by pll connected to a PT100 in order to control the diode temperature within limits. I should recommend to use a microcontroller to switch on/off the heating resistor.
i have a question, it is said , in the previous discussions, that a hotter diode is more efficient, but what about the heat losses?? is the diode still more efficient than a cooler one?? also, don't we need to thermally isolate the hotter diode from rest of the circuit??
1.heat losses comes from the i^2r + vth*i losses in the diode .
2. the question can be understood wrt temp .. better if the forward characteristic
of the diode is seen in a catalogue containing forward characteristic with temp as
a parameter . The actual loss is then given by i*vf.
3. heat sinks are individual in some applications e.g railway rectifiers but even in
common heat sinks(insulated electrically ) it would be reasonable to assume
all diodes are needed to be operated similarly unless time multiplexing is used
to increase rating for instance..Usually the top diode in a stack may be expected
to have higher junction temperature..with bottom forced cooling ..even so for
natural cooled stacks..
hope it helps
Cheers
ok, thank you..
i have another question, with increase in temperature, voltage drop across the diode decreases, my question is , which loss accounts more, the loss due to greater voltage drop, or the heat loss due to the current flowing through the diode?? because if voltage drop is decreased, then more current will flow , thus increasing I^2*r loss, that increase heat losses .
Current is determined generally by external circuit. it is better to use
the actual characteristic e.g. Child's law to model the diode...to determine the drop across it. it is worthwhile to refer to a catalogue of the manufacturers ... with a specific circuit topology to have a clear picture..
Cheers
Dear Sujit,
besides the discussion on electrical parameters, please consider the application's requirements and the economic aspects as well.
To prevent a diode from getting hot during application - while "hot" refers to 30-40K above ambient to achieve a proper lifetime - the only chances you have are
a) use of larger diodes or
b) more aggressive cooling
both leading to higher costs for your design.
The temperature or better the temperature swing a diode in operation may achieve depends on the load profile and the desired lifetime.
A soon as you try to dynamically control the temperature of the diode by external means like heaters, the efficiency of the overall setup will be reduced as the effort will consume more energy than the diode would save.
Dear Sujit,
dear Martin,
one can see it also from the point of system optimization and facilitation. From my side, complication the power electronic system by implementing heaters that also need to cool the diodes to keep the temperature at the desired level may be not commercially feasible.
But on the other hand, one could try a purely passive approach. The diodes could be arranged at places where heat is generated anyway - close to the switches.Their thermal path could be adjusted in order to allow for higher temperature at maximum heat-sink temperature. This could be achieved by a dedictated DCB-layout or additional Rth inserted in the thermal path like additional layers or arrangement on top of the switch die.
Problems that are linked to high temperature are leakage and switching speed (and loss). Thus, high temperature operation may require a voltage derating and operation at low operating frequency.
Let me clarify what I meant. In a particular circuit, operating at fixed conditions, I have two options of heatsink size: a smaller one that permits the junction to operate say upto 110deg C or a larger one that permits the junction to operate say upto 55deg C only. Obviously the conduction losses will be smaller for a smaller overall voltage drop with the smaller heatsink. Since I am not exceeding 125deg C, there will be no voltage derating or need for higher size of diode, nor need for external heaters. However, I am not aware of any increase in switching loss at say 110deg C compared to 55deg C as Michael has said, but I want to find out. Can anyone put in further inputs or clarifications ?
Diode switching losses increase a lot with temperature.
You can find some approximate calculations in "Losses in PWM inverters using IGBTs" available in this site.
- Badges
- Science topic
- Similar topics
- Engineering
- Control Theory
- Control Systems
- Control Systems Engineering