When do we use isolated power supply for driving each 6 IGBTs in an inverter? Some applications use bootstrap method which is not isolated. Using 6 isolated power supplies for an inverter with some extra peripherals may increase the overall cost. So, how should we choose which method to use by considering both performance and cost? Generally, which method is used in industrial application such as variable frequency drive or UPS from 1hp to 100hp?

For the first glance following aspects are visible
- If we want to use bootstrap for high side drivers and reduce the cost. The startup circuit must support bootstrap powering. Startup procedure could be a bit more complicate in this case. Startup sequence should assume particular delays and so on. Most probably protection recovering procedure will be similar to startup procedure and a bit more complicate than general one.
- If we have really high current in inverter the bootstrap can suffer with parasitic inductors and di/dt.
- If we have really high voltage in inverter the bootstrap can suffer with parasitic capacitance and dv/dt
- If we have really high power the total cost is getting high enough to ignore potential cost issue with aux dedicated power supply for drivers. Btw, di/dt and dv/dt are getting high in the case too.
At the end it is trade of taking little additional care about potential problem or accepting extra cost.

Ilya, I wonder those motor drive manufacturer like Siemens, ABB's choices for different power level. Do you have any specific value on the application, for example voltage and current values for transition from bootstrap to isolated gate diver.

The difference are:
1. If you have something wrong at the power side using bootstap, you will have demaged microcontrollers and interfaces! If you use optocouplers or transformers there won't be this problem.
2. If you use optocoplers or transformers you don't need to protect a consumer from electric shock, it's isolated and save (I'm talking about interface)!

Hi Feyzu,
please note I am talking about voltage, current and power of the converter power stage. This is not about voltage, current and power of bootstrap circuit.

Particular figures and recommendation would be better to get from open source papers.

Only four isolated power supplies required for inverter for isolated driver. you can use single power supply for all three bottom switches.

you want save cost you can go for a bootstarp method driver with single power supply . use icoupler to isolate power side and control side. icoupler will address the isolation issues.

Figure 25 and 26 illustrate possible application. Both are with flying isolated power supply. I see no bootstrap solution.
One more note,
if you need negative voltage for gate driver and still want use bootstrap, than you need quite specific schematic. Would better estimate it, before making decision.

I think there are many limitations like dv/dt and di/dt and protection considerations to use bootstrap instead of isolated power supplies in high power converters. Use of bootstrap is limited to 10kW inverters in practical products. You can use an open-loop unregulated low cost isolated flyback converter to produce 4 output for gate drivers. The IGBT's switching performances are not very sensitive to gate driver voltage above 15V but you should employ protective 20V zeners to avoid gate circuit damage!

Very agree with Arash,
one more potential issue with bootstrap is protection schematic and current sensor.
It can happen that particular high side driver has embedded isolation and/or protection logic and/or current signal amplifier and/or sensor. And such driver IC requires dedicated power supply.

Why does bootstrap should be an alternative to isolation ? The issues discussed above are related to the lack of isolation between control and power, and I believe they do not apply to the bootsrtap technique.
The major drawback of bootstrap is simply related to the limitation in the possible PWM strategy: in practice there's an upper boundary to the on-time of the high-side switch, that depends on the discharge time of the bootstrap capacitor, and a lower boundary to the on-time of the low-side switch, that depends on the charge time of the same capacitor. Typical PWM strategies in VFDs (like sinusoidal or space-vector) normally have no problem with bootstrap. Earliest bootstrap drivers used to suffer from latchup phenomenon because of negative voltage on the GND pin caused by high current transients in the parasitic L of the layout. But newest devices (like IR21864) with some external components for protection (as suggested by app-notes) and a well-done layout will perform very well up to tens of amps and hundred of kHz (and if the driver max current is not enough to drive a large gate, you can always, just put a couple of transistors in totem pole configuration).

Isolation is another issue: high-current, high-voltage and fast transients management may take big advantage from isolation. So why not considering to use bootsrap technique with isolated commands ? That's what we normally do, when possible.

The major advantages using bootstrap are:
* no need to design and prototype a custom flyback transformer with multiple outputs

* much easier layout (no need to wire the flyback isolated outputs to the high side switches)

Instead cost reduction may be only marginal...

I agree with Davide, the isolation (from control IC to the power device gate/source) is a separate issue from whether or not bootstrap is used to generate the high-side +15 VDC or so to run the high-side drive. In this case, I think we're talking about dual gate drivers (1/2-bridge driver ICs) which almost all have very good isolation built-in to protect the control IC signals on the input side of the driver IC. Even if the power FETs/IGBTs fail, causing the 1/2-bridge gate driver to fail, it would be the output-side of the dual gate driver that fails, and due to the isolation built-into the IC, would protect the input signal side of the IC in nearly all cases. This isolation is done with either capacitive or magnetic isolation, depending on the manufacturer, but the galvanic isolation is there and generally goes up to several hundred volts if not a few thousand volts for short-lived, high transients. Also, whether one uses an opto-coupler or a small gate-drive transformer, both in general have worse limitations on min/max duty cycle at high switching frequencies than circuits using the bootstrap for the upper FET/IGBT bias supply.

IGBT driver side and signal interface, should be isolate opto / pulse transformer.
and one power supply, two out put using driver side and interface separately isolate.

It is highly reliable to have individual gate drive circuits for each of the six IGBT's with isolated power supplies. Following are the salient points:

1) Have a full bridge PWM with fixed duty cycle and a three winding transformer (one primary, two secondaries. Secondaries can be tapped to give +17V and -5V for each IGBT in one leg). Hence one power supply with isolated outputs and dedicated gate drive.

2) The same circuit can be duplicated for the other two legs.

3) This approach though has higher component count has the advantages such as
1) Same FAB for the three legs and hence manufacturing cost reduces (both for fab and assembly as the number of boards is higher and the footprint is same).
2) Higher reliability as three circuits are not interdependent and the components can be rated accordingly. Also, the replacement of parts is simple.

Bootstrap requires only one isolated power supply for all the switches, very easy to be built with off the shelf switching trafos (in flyback configuration). That's a winner.

At the very least, all 3 lower FETS should be run off the same +12 to +16VDC supply. The choice of bootstrap versus 3 additional isolated supplied for the gate drive of the upper 3 FETs should depend on the frequency of switching and the minimum lower FET pulse-width (for minimum charging time of bootstrap caps) and the maximum upper-FET pulse width (for max voltage drop on the upper FET bootstrap power supply output voltage as the bootstrap cap voltage drops more the longer the upper-FET is powered by the bootstrap cap. In all but the most demanding and very high power circuits, a bootstrap supply for the upper-FET gate drive supply is the usually the best and most reliable way, but it must be designed properly. For example, proper selection of the bootstrap isolation diode being an ultrafast and proper selection of the bootstrap cap both in value and in type. In fact, one usually must use something like a 100 nF cap in parallel with a 1uF to 10 uF cap for the bootstrap supply. And it's imperative that some sort of upper FET driver with a UVLO ckt be used in case the lower FETS are being driven below the normal minimum duty cycle (as with no-load on the the SMPS). This will protect the upper FETs during those cases. If an isolated flyback supply is used for the upper FET Gate Drive supply instead of a bootstrap supply, it must be thoroughly checked out under all conditions on line and load on flyback output, plus startup conditions, to ensure that there will always be at least +10VDC available at all times, but not more than say +16VDC, certainly not more than +18VDC transiently for driving the upper FETs. And when I say FETs I mean MOSFETs or IGBTs. The most important thing for reliability here is that the gate drive supply and the gate drive itself, give a predictable output under all operating conditions (perhaps with UVLO if needed) to protect and maintain reliability of the power FETs or IGBTs.