Дипломная работа - Проектирование 10 кВт преобразователя

Подождите немного. Документ загружается.

2.5. Snubber circuits

The main cause of implementing snubber circuits is to make sure the switching power

components remain within their safe-operating-area (SOA) both at turn-on and turn-off.

This is very important to ensure the longevity of the devices as well as reducing the

amounts of EMI emitted from the high power components. When introducing a snubber it

is not uncommon to be able to reduce switching losses in the switching devices, in

particular the turn-on switching loss. However, in general the losses removed from the

transistor are transferred to the snubber circuits instead while these components are

usually less sensitive to voltage spikes and are able to perform satisfactory in a wider

temperature span. As there are numerous variants of snubber circuits for different

purposes only the ones estimated to prove suitable will be presented below along with

their corresponding advantages and disadvantages. The simulations carried out are done

using Cadence Pspice Student Edition, the circuits can be viewed in Appendix E.

Although the simulations will not be used very much for estimating good snubber circuits

to obtain reduced power loss as much as ensuring longevity to the switching devices.

2.5.1. Increased SOA

Due to the non avoidable stray inductance in the DC-bus loop in addition to the internal

stray inductances in most of the components causes along with the very fast current

switching of modern semiconductors a large voltage over-shoot when switched off. To

some extent there is an over-shoot during turn-on as well but not nearly as large. The

faster switching and larger stray inductance the larger voltage over-shoot. The easiest and

least complex method to manage this over-shoot is to compensate for the stray inductance

in the DC-bus loop by connecting a small, fast and low inductance capacitor as close to

the IGBT terminals as possible as shown in figure 20 (left). Actually increasing the gate

resistor to slow down the switching process would work as well but switching losses

would rise according to it.

Figure 20 : Decoupling and restricted decoupling capacitor

There are capacitors manufactured for this purpose with very low self inductance and

designated connections to properly fit the concerned IGBT module. However when using

discrete components this gets harder as the components are separated which also causes

higher stray inductances. As the separation is one of the causes of choosing discrete

components the inductance is a large trade-off. The difference when using only the

decoupling capacitor can be seen in the figure 23.

Figure 21 : Switching device voltage without snubber circuit

As seen in figure 22 the transient at turn-off is easily large enough to cause failure in the

switching device, it is nearly twice as large as the DC-bus voltage. By attaching the

decoupling capacitor the over-shoot decreases to about 123% of DC-bus voltage but

instead an oscillating voltage will occur over the transistor as seen in figure 23. This

oscillating voltage will inflict a large current through the capacitor which often is

designed for small currents, with low self inductance as a benefit. The capacitor value can

be estimated with equation 2.5.1.1 [20].

()

DCpk

−

(2.5.1.1)

Where I

is the maximum switched current, L

being the DC-loop inductance and V

the

maximum allowed peak voltage over the switching device. This oscillating behavior is

seen in 23.

Figure 22 : Device voltage with decoupling capacitor

This oscillating behavior causing high RMS current through the capacitor is the major

disadvantage if the decoupling capacitor. The more inductive DC-loop and higher output

current the higher RMS current will run through the decoupling capacitor. This makes

this type of snubber circuit suitable for low current applications. The estimated power

loss due to in the capacitor can be estimated with 2.5.1.2 [20].

ILfIESRP

SswRMSdecouplingloss

××=×=

(2.5.1.2)

This would not result in a very high power loss but still it can be too high for these kinds

of capacitors as they often are very delicate. The decoupling capacitor is a good solution

when having a switching frequency of only a few kHz. The loss and oscillating voltage

can be reduced by adding a diode and a resistor to restrict the de-charging and reduce

oscillations. The circuit can be seen in figure 21 (right). Although to get a small enough

over-shoot the capacitance needs to higher as the diode and the resistor introduce further

inductance to the loop. The diode should be of fast and soft recovery type to avoid severe

oscillations following V

at turn off. The resistor can be calculated using equation

2.5.1.3.

()

swsn

××

(2.5.1.3)

Figure 23 : Device voltage with discharge restricted decoupling capacitor

However, as this circuit adds more components and therefore more stray inductance a

larger capacitor will be needed, although the power lost will mainly be dissipated in the

resistor and the diode as explained in equation 2.5.1.4.

(

)

swDCpksndecouplingrestricedloss

fVVCP ×−××=

(2.5.1.4)

The RCD Clamp-Snubber seen in figure 24 (right) does have a very favorable effect on

both the Turn-On and Turn-Off transients but have limited possibilities to reduce the

switching loss in the transistor. However, the loss dissipated in the snubber circuits is

much less than in the RCD Charge-Discharge Snubber (seen in figure 24, left) which is

calculated as in equation 2.5.1.5 [20].

swpksneDischeChRCDloss

fVCP ×××=

arg_arg__

(2.5.1.5)

Although the RCD Charge-Discharge Snubber inflicts some power loss it can help to

speed up both the Turn-On and in particular the Turn-Off processes directly reducing

power lost in the IGBT and diode. But it can be difficult finding component capable of

sustaining the internal loss while maintaining a low self inductance [20].

Figure 24 : RCD Charge-Discharge Snubber and RCD Clamp-Snubber

In figure 26 and 27 a switch has been simulated and plotted illustrating the difference

between the RCD Charge-Discharge Snubber and the RCD Clamp Snubber. As it can bee

seen they Turn-ON rather similar and cause about the same amount of over-shoot but the

RCD Charge-Discharge Snubber has a much quicker Turn-Off. As the Turn-Off with its

tail is the main part of the switching loss the total loss can be vastly reduced using a RCD

Charge-Discharge Snubber even without reducing the gate resistor. Reducing the resistor

would also reduce the losses but would do so by causing the over-shoot to increase once

again.

Figure 25 : RCD Charge-Discharge Snubber

Figure 26 : RCD Clamp-Snubber

2.5.2. Reducing losses

As the snubbers circuits help to attenuate the voltage transients the switching process can

be faster causing reduced switching losses by reducing the gate resistor. However,

precaution has to be taken to ensure the driver can withstand this higher peak current.

Although not actually reducing the losses as much as moving them is done with some

snubber as they consume then unwanted energy and dissipates it within its own

components. If this is a diode or a resistor they can withstand a lot of power loss at a high

temperature implying little need for excessive cooling.

Type Advantages Disadvantages

Low snubber losses

Directly and favorably effects Turn-

Off and Turn-On on voltage stress

Decoupling

Capacitor

Special solutions of casings for

modules

Produces voltage and current

oscillations in the DC-bus,

forcing usage of capacitor with

high RMS current limit

Only a good choice in low current regions

Low snubber loss

Directly reduces the Turn-Off voltage

over-shoots but also has as a favorable

effect on Turn-On voltage transients.

Additional circuitry increase

snubber inductance, making

protection less effective.

Discharge

restricted

decoupling

Capacitor

Much quieter switching as diode

blocks of oscillations.

Snappy diode could produce high

recovery voltage spikes and

dv/dts across IGBT pair.

Good in medium current ranges

RCD Very high snubber losses

Reduces Turn-Off voltage over-shoots Requires more components

Could substantially reduce Turn-Off

loss in switching device

Charge-

Discharge

snubber

circuit

NO oscillations on the DC-bus Difficult components selection

Good for high current and low DC-bus voltage applications

Low snubber losses

Directly reduces the Turn-Off voltage

over-shoots but also has as a favorable

effect on Turn-On voltage transients.

RCD

Clamp

Snubber

NO oscillations on the DC-bus Requires more components

Good for medium to high current applications

Table 6 : Snubber characteristics

A quick glance gives the impression that the RCD Charge-Discharge Snubber would be a

good choice for this inverter as the chosen transistor already has a very small gate resistor

as default. In this way loss could be reduced without adventuring a fatal over-shoot.

2.6. Principles of cooling

2.6.1. Thermal resistance

A common design problem is to remove heat enough to guarantee satisfactory operation.

Then the thermal conductance between to adjacent mediums should be as high as possible

to in the end dump as much heat as possible in a large buffer, usually the surrounding air.

The thermal resistance is as usual the inverse of the conductance and is as a normal

resistance additive when connected in series respectively. An equation describing the

temperature where the power is lost in comparison to the ambient temperature in respect

to thermal resistance can be derived from Newton’s law of cooling and known relations.

heat

TTT

TAk

θθ

−

=→

−=Δ

→

Δ××

Δ 1

(2.6.1.1)

The different components can be fragmented into many if there are several layers of

different mediums between the source and the buffer. This makes it rather simple to

estimate heat sink needs when most manufacturers state the estimated thermal resistance

of the device. However, when having different cooling topologies a slight different

approach has to be taken.

2.6.2. Heat transfer

When using a cooling bed with circulating air or fuel as cooling medium the case is

somewhat different, a lower limit of coolant weight at a certain temperature has to be

derived. The allowed rise in temperature is dependant on the following devices in the

chain and how high temperature they can withstand. A higher flow of coolant will cause

it to warm up less in each device while a higher pressure would suppress this behavior as

more coolant can absorb more heat. The energy delivered to the cooling bed can be

calculated using equation 23 with the respective thermal resistance. This energy is to be

removed by the coolant and depending on the specific heat capacity and the throughput of

the coolant the cooling bed can be held at a reasonable temperature. Assuming it is

known how many degrees the whole chain devices can raise the temperature and how

many devices there are; an average per device of allowed temperature raise can be

derived. This along with the total power loss needed to be removed and the nature of the

coolant the needed throughput can be calculated as in equation 2.6.1.2:

tmP

mTCmQ

Δ×

=Δ×→=

→

Δ×

=→Δ××=Δ

(2.6.2.1)

Where C

is the specific heat capacity, ΔQ is the energy differential and ΔT is the

temperature differential. Suppose a 2°C raise in temperature is allowed and 200 watts

needs to be removed from the cooling bed by air. Using the formula above gives a

minimum flow of 100g/s. From consulting with the engineer Lars Austrin at SAAB AB it

was given that the onboard cooler is slightly less efficient than a heat sink with forced

convection at sea level.

2.7. Chassis design

The main purpose of the chassis usually is to provide a safe and protected environment

for the usually sensitive equipment. The challenges are many and not always obvious, for

instance the most common challenges are:

•

Mechanical shock

•

Electrical hazard both to components and to environment

•

Dust and other small particles such as sand

•

Humidity and liquids

•

Electromagnetic environment

•

Cosmic radiation

•

Temperature

The mechanical shock is everything from a bump to the slightest vibration, both possible

to cause stress resulting in either short-circuit or interrupted conduction. As it can be

assumed that everything in an aircraft is very fixed the main issue is still vibration, the

remedy can be to mount everything elastically or just the chassis it self. Easily understood

it is more practical to only mount the chassis on suspensions but as it usually is much

heavier than smaller components these can suffer from a vibration of higher frequency

than the one filtered by the chassis. It is also of highly importance that the chassis does

not submit to fatigue especially when the temperature can vary in a wide span and

chemicals introduced in the environment possibly causing some material to decrease in

tensile strength.

Different materials can be used with their own advantages and disadvantages such as

polymers and metals, especially alloys of either aluminum or steel.

Polymers are usually very light and can be constructed in any imaginable form and shape,

however, it is very likely to submit to fatigue either from unfavorable temperature, strong

light or chemicals such as acids or different hydrocarbons. Metals and usually aluminum

alloys are heavier and less moldable but are much more resistant to external influences

and are much stronger. Within a great temperature span it is not affected, strong light can

also be neglected but some chemicals can cause corrosion, at least for untreated non-

precious metals. A big advantage with a metallic chassis is it being thermal conductive all

over. More characteristics for both metals and polymers will be explained further on.

Electrical hazard to components can be avoided by both protecting it from physical touch

with an enclosure and providing a fixed support for the internal conducting parts making

them firmly separated from each other. A chassis made of polymer is usually a very good

insulator which makes it very unlikely to be exposed to electric hazard by touching the

chassis in the case of an internal malfunction. In a fully conducting chassis a fatal

accident to devices or humans can occur in the case of an error making the chassis

electrifying. It also has some practical advantages due to its insulating nature when it

comes to mounting conduction parts on the chassis making individual insulators

unnecessary.

In the case of forced convection using a fan, dust and other particles can collect in the fan

or in a filter reducing its capacity and even eventually causing a failure, not to mention

the documented limited lifetime of a fan further increasing risk of failure.

Even in an external heat sink massive amount of dust can prevent natural airflow and

radiation also reducing the capacity, even if the airflow is high enough this problem is

unlikely to arise in either solution; however it is worth a thought. If a fully sealed

enclosure with no external heat sink can be used this problem is eliminated. This sealed

enclosure would further help reducing the amount of water and moist inside the chassis

which in some cases can prove fatal to circuitry and even make metals to corrode

possibly inflicting problems such as connection faults and reduced tensile strength.

A very common problem is when high power circuitry is mixed with logics. Due to

inductive and capacitive load high energy electromagnetic waves at radiofrequency is

radiated causing disturbances and noise in logic. In some extreme cases it can even

destroy sensitive devices. The noise can be generated by the device itself or generated by

other devices in the environment, in either way the susceptibility and emitted noise

should be kept to a minimal. Except from considering the cabling a shielding cage can be

made out of the chassis, it will work as the renowned Faraday cage and will keep noise

generated inside in and noise generated on the outside out. A metallic chassis

automatically inherits this benefit but a chassis made out of polymer has to have a film or

fine-meshed grating on either the inside or outside.

At high altitudes the failure rate in power-devices increases rapidly due to cosmic

radiation. Except from reducing the voltage over the device or the junction temperature

not much can be made to reducing the influence of cosmic radiation except forming an

enclosure out of a thick heavy material such as lead or concrete. However, as this task is

to optimize the power/weight-ratio this is a highly unsuitable solution.

2.7.1. Cooling scenarios

Besides the issues stated above the volume is of course critical but the main factor for

choice is how to implement cooling depending on which environment and which cooling

types that can be inherited from the surroundings.

In the case of an avionic application such as a fighter aircraft there are mainly three

possible solutions for cooling and each have to be evaluated individually:

• Exterior heat sink integrated in chassis relying on a sufficient natural airflow

around the device

•

Interior heat sink relying on either a chassis mounted fan or airflow from external

system

•

Internal fluid cooling using coolant from external system

2.7.1.1. Exterior heatsink

To use a passively cooled heat sink to remove the heat is by far the heaviest and most

space consuming solution but it is also completely independent of an external system, has

no mechanical parts that wear out and cannot suffer from leakage or condense which is

possible in fluid cooling. In an aircraft such as JAS 39 GRIPEN the temperature within

the hull can easily reach 70°C and considering operating at 10000-15000 MASL

the

convection contribution becomes less as the air becomes less dense. For instance from a

properly designed heat sink relying on natural convection at sea-level about 70% of the

heat is transferred by natural convection and 30% by radiation but at about 15000 MASL

the ratio is about the opposite. This further increase the size required for the heat sink

since the radiation part is not increasing, it is only the convection part that is decreasing.

This inverted ratio increases the overall thermal resistance to about 233% of the original

since the convection part of the heat transfer reduces to about 18% of the original while

the radiation part is constant [15].

In power equipment it is very common to make heat sinks form up the long side walls

with high loss components internally mounted on them. Then the short side is made of

sheet-metal making a suitable place for mounting indicators, connectors, switches or even

a fan. In the case of a fan an inlet or an outlet has to be made on the opposite side to

permit airflow, usually done by perforating the sheet.

The power-components should be spread evenly over the heat sink to reach the best

performance in respect to cooling. As the heat sinks usually are very sturdy, the other

sides can be made out of thinner material, reducing the weight and this principle is

illustrated in the sketch given in figure 28. A more detailed sketch with components fitted

can be viewed in appendix F.

Meters Above Sea Level