isolated DC-DC convertor and mosfet driver

[shortbus=]

Hi, I know only enough to get myself in trouble. My question is about using an isolated DC-DC module instead of a boot strap circuit when using a half bridge driver to make a high side switch. I know I could use a second small mosfet on the low side of the half bridge to recharge the boot strap but why not the isolated convertor?

Any one ever done this or am I just dreaming?

 

[Papabravo]

Hi, I know only enough to get myself in trouble. My question is about using an isolated DC-DC module instead of a boot strap circuit when using a half bridge driver to make a high side switch. I know I could use a second small mosfet on the low side of the half bridge to recharge the boot strap but why not the isolated convertor?

Any one ever done this or am I just dreaming?

There is more than one way to do isolation. When you use isolation there are delay and switching time issues. Most reliable way of building something that will work is to use a gate driver.

 

[Tony Stewart]

TI's smart high-side switches use internal Nch with internal charge pumps to drive the Gate driver. They include current sense/limiting and other useful features for USB power. https://www.ti.com/lit/ds/symlink/tps2557.pdf The internal charge pump is like using the lower side of a half-bridge with the boot-strap cap and diode on Vdd.

Isolation is useful when you get into high-voltage grid-tied DCDC converters. or inverters

 

[rjenkinsgb]

My question is about using an isolated DC-DC module instead of a boot strap circuit when using a half bridge driver to make a high side switch.

Quite common in industrial servo drives.

One type I've worked on has a high frequency power output from the main high current PSU (possibly an extra winding on its electronics supply?), that is daisy chained through the driver interconnects to each servo module; they each have isolating transformers / rectifiers / smoothing to produce their own isolated power for the high side drivers.

You may be able to use isolated DC-DC converters, but as the high side power negative is connected to & switching with its half bridge output (on a common 2 x N channel half bridge), some types may not like the high voltage AC between in & out.

 

[shortbus=]

Thanks for the answers, but I'm still confused. Will answer in individual posts.

 

[shortbus=]

There is more than one way to do isolation. When you use isolation there are delay and switching time issues. Most reliable way of building something that will work is to use a gate driver.

Hi, I don't understand how there would/could be delay and switch time issues. It's my understanding that a isolated DC-DC convertor is always on when the circuit is on. I am using a gate driver, the isolated DC was to replace the boot strap circuit and simplify the circuit as well as limiting chance of shoot through.

 

[Papabravo]

Hi, I don't understand how there would/could be delay and switch time issues. It's my understanding that a isolated DC-DC convertor is always on when the circuit is on. I am using a gate driver, the isolated DC was to replace the boot strap circuit and simplify the circuit as well as limiting chance of shoot through.

Optoisolators are one way of achieving isolation. Some optos have poor propagation delay and switching times.

 

[shortbus=]

TI's smart high-side switches use internal Nch with internal charge pumps to drive the Gate driver. They include current sense/limiting and other useful features for USB power. https://www.ti.com/lit/ds/symlink/tps2557.pdf The internal charge pump is like using the lower side of a half-bridge with the boot-strap cap and diode on Vdd.

Isolation is useful when you get into high-voltage grid-tied DCDC converters. or inverters

I looked at those and other chips with on board bootstrap charge pumps. Reasons I want to avoid-
1. those chips are surface mount and have a ground pad under them that needs to be soldered to the PCB. Could probably handle the surface mount legs but have no clue as to how the bottom ground pad gets soldered.

2. the mosfets I'm using are large Isotop ,SOT-227 style and need a lot of current from the drivers to turn on fast. The drivers I have are FAN7390 ones high current out put, where the charge pump drivers are low output ~1/2 amp.

3. this is a fairly high DC voltage around 90VDC and 15A.

 

[shortbus=]

Quite common in industrial servo drives.

One type I've worked on has a high frequency power output from the main high current PSU (possibly an extra winding on its electronics supply?), that is daisy chained through the driver interconnects to each servo module; they each have isolating transformers / rectifiers / smoothing to produce their own isolated power for the high side drivers.

You may be able to use isolated DC-DC converters, but as the high side power negative is connected to & switching with its half bridge output (on a common 2 x N channel half bridge), some types may not like the high voltage AC between in & out.

Quote -

they each have isolating transformers / rectifiers / smoothing to produce their own isolated power for the high side drivers.

Isn't that what is in a isolated DC-DC convertor? Murata seems to think that's what is in theirs. But talking with or trying to talk with Murata, they didn't seem to want to share info about the question I asked here. Or maybe it was because I'm not going to buy more than 1 piece and am a hobbyist.

Quote -

You may be able to use isolated DC-DC converters, but as the high side power negative is connected to & switching with its half bridge output (on a common 2 x N channel half bridge), some types may not like the high voltage AC between in & out.

Again I'm confused. I'm only going to use the high side not the low side of the driver. And don't understand where AC comes into the circuit?


Since you work around industrial stuff, you may know what I'm trying to build. I'm designing a power supply, spark generator for a DIY sinker EDM. A project I've been working on since the 1990's, the ram and other mechanical parts have been done since then but the electronics are my down fall, like I said in the first post I know enough to be dangerous.

 

[Papabravo]

I looked at those and other chips with on board bootstrap charge pumps. Reasons I want to avoid-
1. those chips are surface mount and have a ground pad under them that needs to be soldered to the PCB. Could probably handle the surface mount legs but have no clue as to how the bottom ground pad gets soldered.

...

The answer is that you don't use a soldering iron. You use solder paste and a reflow oven. An SMD rework station is a potential alternative to a production machine.

 

[shortbus=]

The answer is that you don't use a soldering iron. You use solder paste and a reflow oven. An SMD rework station is a potential alternative to a production machine.

At my age I can't see buying those for one project. The other problem is they are under powered for my mosfets.

 

[rjenkinsgb]

Again I'm confused. I'm only going to use the high side not the low side of the driver. And don't understand where AC comes into the circuit?

Look at a typical half bridge driver diagram:

The high side MOSFET source is also the output terminal, which is switching between 0V and the output bus voltage.

That means the voltage difference between the DC-DC input and output halves is varying at the power output terminal voltage and frequency - eg. possibly a few hundred volts at 30 - 400KHz AC across the DC-DC isolation barrier.

Some DC-DC modules may get upset with that, even if they can stand far more voltage differential in a steady state.

If the H-Bridge had a P-Channel high side device, then the source and DC-DC output are just at the bus voltage, so any type with suitable isolation should work.

 

[rjenkinsgb]

See section 8 in this application note - a floating high side supply using a CMOS(?) 555!

https://www.farnell.com/datasheets/2254418.pdf

That section points out another problem with DC-DC converters; if the capacitance between input and output is too great, it can cause spurious switching.

 

[ronsimpson]

I use small 1 and 2W isolated supplies on both top and bottom side supplies.
MGJ2D121515SC is an example. In my case the MOSFETS switch very fast so I need a supply that can handle fast transients. "CMTI>200kV/uS" When the power fets switch it places stress on the isolation in the supplies. Because I am using GaN fets I need isolation that can handle fast edges.

 

[Papabravo]

At my age I can't see buying those for one project. The other problem is they are under powered for my mosfets.

In the not-too-distant future you can expect thru hole components to disappear entirely. It will simply not be economical to manufacture them. Everything will be tape and reel.

 

[ChrisP58]

Here is a 4 Amp, through hole gate driver with charge pump.
https://www.digikey.com/en/products/detail/infineon-technologies/IRS21864PBF/1300606
It is obsolete, (probably because it's through hole) but there's plenty of stock for a one off project.

 

[shortbus=]

Here is a 4 Amp, through hole gate driver with charge pump.
https://www.digikey.com/en/products/detail/infineon-technologies/IRS21864PBF/1300606
It is obsolete, (probably because it's through hole) but there's plenty of stock for a one off project.

Not wanting to disagree but the data sheet says and shows the part as needing a boot strap circuit. And the boot strap is what I'm trying to get away from. Thanks any way for the help.

 

[shortbus=]

In the not-too-distant future you can expect thru hole components to disappear entirely.

In the not to distant future I'll probably be gone too.

 

[shortbus=]

See section 8 in this application note - a floating high side supply using a CMOS(?) 555!

https://www.farnell.com/datasheets/2254418.pdf

That section points out another problem with DC-DC converters; if the capacitance between input and output is too great, it can cause spurious switching.

Hi, thank you for helping me to understand.

In the app note you linked to, Fig. 16 shows a charge pump, that's what the DC-DC module will replace. And since I'll only be using the high side, won't that eliminate any AC from the low side turning off, since it won't even be in the circuit?

 

[shortbus=]

I use small 1 and 2W isolated supplies on both top and bottom side supplies.
MGJ2D121515SC is an example. In my case the MOSFETS switch very fast so I need a supply that can handle fast transients. "CMTI>200kV/uS" When the power fets switch it places stress on the isolation in the supplies. Because I am using GaN fets I need isolation that can handle fast edges.

Hi Ron, that Murata module is the one I bought. Would it be possible for you to show me a schematic of how your using it with a gate driver? I know there are probably be a couple of other components needed. And what does the module output common go to?

I was really disappointed in Murata's information and getting help from them. I couldn't find any typical use schematics information but may have over looked them.
Even the module pin outs aren't shown even on the data sheet, or at least not on the one I did find.