Bootstrap hi side drive chips dont let negative voltages...why do they care?

[Flyback]

Hi,
The attached circuit is NOT OK due to the negative voltage of TS w.r.t GND.
Howcome it cant take such a neg voltage?....
(LTspice and jPeg scm attache)

2ED2182S06
https://www.infineon.com/dgdl/Infin...N.pdf?fileId=5546d4626cb27db2016cb8d7368a29e3

LTC4440
https://www.analog.com/media/en/technical-documentation/data-sheets/4440fb.pdf

 

[rjenkinsgb]

Do you ever bother to read data sheets???
Do you even understand how semiconductors work?

The LTC4440 has a maximum operating voltage rating of 80V on the power circuit being driven.

175V is more than that!

Also: Device ground is not system 0V. You are attempting to use it in a ludicrously wrong way.

Voltage "isolation" within ICs such as this is usually by reverse-biased PN junctions. Attempting to force an output to the wrong polarity will likely forward bias the isolation junctions and destroy the IC.

 

[Papabravo]

Yet another consequence of magical thinking. The entitlement of thinking things should work the way you want them to is astounding.

 

[Flyback]

Thanks, yes granted the LTC4440 is only rated to 80V....though my apologies as its the 2ED2182 (650v) that we will end up using, as its says on the schem above.
So it seems that a bootstrap driver cannot be used here.....though i suspect that we may use a digital isolator chip instead to level shift the gate drive signal...then shovel that into a hi side gate driver.
Would you agree?

Also, AYK, We cannot use a pulse transformer as we sometimes will need the FET to be held on for several hours.

I am amazed theres nothing offtheshelf for this, as driving a buck fet from such a high "Underside" is common, eg with pfet buck drivers, which need converting to nfet drive.....the hi "underside" being needed as the pfet driver needs to connect to a hi side sense resistor.

 

[rjenkinsgb]

There is nothing wrong with correctly using an appropriately rated bootstrap style driver.

eg. The high side half of such as an IR2110

https://www.infineon.com/dgdl/Infineon-IR2110-DataSheet-v01_00-EN.pdf?fileId=5546d462533600a4015355c80333167e

 

[Flyback]

There is nothing wrong with correctly using an appropriately rated bootstrap style driver.

Thankyou, indeed you have kindly implied that in proper useage they are used in the low side and the hi side fet they drive is in the hi side...ie thats the standard way to use them. As in the top schem, this modus operandi is unfortunately not available to us.
As such, we wonder about the attached configuration instead? Using a digital isolator....one that cares not that its GND1 will fall more than 100V below its GND2.
Would you back the attached, may i ask?

LTspice and jpeg of alternative modus operandi as attached

The digital isolator used would be eg ISO7810

https://www.ti.com/lit/ds/symlink/iso7810.pdf?ts=1673637471558&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FISO7810

 

[rjenkinsgb]

As in the top schem, this modus operandi is unfortunately not available to us.

Why? There is no requirement to use the low half of the driver, the two parts are separate and have individual inputs.

 

[shortbus=]

Do you ever bother to read data sheets???

And yet he wonders why there are no power supplies(I think that has been his complaint) made in his country.

 

[Flyback]

Why? There is no requirement to use the low half of the driver, the two parts are separate and have individual inputs.

Thanks, yes you are right, but if you see the application of post #1 and #6, you see that standard bootstrap drivers simply cant be used......we need to drive a hi side Buck FET from that "hi underside" rail, as shown.

 

[rjenkinsgb]

we need to drive a hi side Buck FET from that "hi underside" rail, as shown.

The upper half of an IR2110 should work fine

The only allowance you would need to make, as the load does not have an active pulldown as with a half bridge, is a supply to feed between Vs and Vb.

eg. It could be a 15V zener + cap from the Vs pin, fed from a bias resistor from the input to give a floating supply above Vs, the FET source voltage.

Or, there is a floating charge pump version on page 18 of the app note:

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

(Different driver IC, but the same principle).

Supply the VCC pin with an appropriate voltage relative to COM (Power circuit 0V), even when the lower driver is not in use.

 

[Flyback]

The upper half of an IR2110 should work fine

Thanks, thats what i thought.
Then it emerged that the "switching node" (VS) connection of the bootstrap driver (the node to which the actual fet drive bit is referenced to), cannot fall more than a few volts below the actual "system ground" (GND) connection of the bootstrap driver. ...Its even worse than that, the VS node should only go below the GND node for a few microseconds at a time (maximum).
VS falling below GND on these bootstrap drivers is one of their big time failure modes.

Top of page 2 of this explains about it......

2ED2101 datasheet:

https://www.infineon.com/dgdl/Infineon-2ED2101S06F-DataSheet-v02_03-EN.pdf?fileId=5546d462766a0c1701766abe07220167

Also, page 44 onwards of the below explains it further...

https://www.onsemi.com/pub/Collateral/TND379-D.PDF

Pg 16 of this also explains about it....

https://www.onsemi.com/pdf/datasheet/ncp5106-d.pdf

This is why bootstrap drivers cannot be used for this application.

The weird thing is that isolated gate drivers like UCC23513 dont seem to suffer this, but their datasheets dont brag about it....

UCC23513 data sheet, product information and support | TI.com

TI’s UCC23513 is a 5.7kVrms, 4A/5A single-channel opto-compatible isolated gate driver with 8 & 12V UVLO options. Find parameters, ordering and quality information

www.ti.com

....or is it that they just dont own up to it.?

 

[rjenkinsgb]

Then it emerged that the "switching node" (VS) connection of the bootstrap driver (the node to which the actual fet drive bit is referenced to), cannot fall more than a few volts below the actual "system ground" (GND) connection of the bootstrap driver

How and when can it ever do that?
The source can only go below ground by the voltage drop of D2.

Vs should always be 10 - 20V more positive than the source.

 

[Flyback]

How and when can it ever do that?

Due to PCB stray inductance......as you know, with high power SMPS, its often simply not possible to get a really tight layout.
The docs referred to above explain how it can happen.
One of the infineon chips claims to be able to handle VS going 100V below GND , but only for small (100ns) pulses.

Vs should always be 10 - 20V more positive than the source.

..Thanks, I think we have confirmed that Bootstrap drivers are no good for the example shown in #6.
Do you think a UCC23513 based solution would be OK though?

UCC23513

UCC23513 data sheet, product information and support | TI.com

TI’s UCC23513 is a 5.7kVrms, 4A/5A single-channel opto-compatible isolated gate driver with 8 & 12V UVLO options. Find parameters, ordering and quality information

www.ti.com

I am wondering how they drive fets in totem pole PFCs, as they often have a similar kind of setup to that which i show in #6 (ie fets being driven ultimately from a rail "above" them.)

 

[rjenkinsgb]

They can stand Vs being as low as -5V with no effect.

If you have that much noise and voltage drops then your problems are far bigger than the choice if IC.

 

[Flyback]

Thanks, circuit not yet built.
Bootstrap ICs are out of bounds, as you kindly helped me to understand....
A big Thankyou to RJenkinsGB for excellent info leading to the conclusion of "Bootstrappers being out of bounds" for us here. This is now seen to be correct.

...So then you get into the "OK what else can we use?" frame of mind......then you start thinking, "why on earth are bootstrap ICs so popular when they have this kind of problem?"....and "do any other isolated gate drive methods have such a problem?" (we hope not).

I am wondering about STGAP2S

https://www.st.com/resource/en/datasheet/stgap2s.pdf

....but there seems to be no dominant product in this range, and we fear future obselence...

Also looking at Infinieon "coreless technology" range, but datasheet doesnt say if they can go to 100% duty.

https://www.infineon.com/dgdl/Infineon-1EDIxxy12AH-DS-v02_00-EN.pdf?fileId=5546d46253f6505701543843c049027b

...also, if its "magnetically coupled", then why dont they give the minimum frequency of use, or at least, the maximum allowable V/us product of the windings?

 

[rjenkinsgb]

"Bootstrappers being out of bounds" for us here. This is now seen to be correct.

Only to you.

By your argument, they would not be suitable for any high-current / high-power inductive load applications - which is the main thing they are used for, in the real world: H-Bridge and three phase bridge drivers for motors etc., anything up to 50KW or more output!

(The app note also specifically include mention of buck converters.)

You are apparently not part of this world. I don't know why I bother responding, it's always a waste of effort.

You claim to know better than the IC designers and engineers that actually use these products in reliable, mass-produced, high power systems.

Now blocked.

 

[Flyback]

Thanks, no you were very helpful, Bootstrappers can only be used when V(isolated ground) is not deliberately getting driven below V(ground)
As can be seen in #6, this is not the case.

I suspect that you are possibly thinking that we could drive the NFET from the true system ground....that is indeed what a highly competent engineer , as you are, would indeed think........and that is where i am saying no we couldnt....for reasons that would take a long time to explain.......ill have a go at making it brief for any reader who is interested....So, its an LTC7860 surge buck which normally drives a 200V PFET....but we want to use it to drive an NFET instead.......as you can tell, if i try to explain more i am going to bore people away from the thread.

 

[shortbus=]

as you can tell, if i try to explain more i am going to bore people away from the thread.

For one the only reason people may not respond more to you is what rjenkinsgb said in post #16. I have never gone to school or college for this electronics stuff but so far have known more that you in almost all of your threads.

As to your not wanting to use a bootstrap for the gate drivers have you though of adding a DC-DC convertor?

 

[Flyback]

"""s to your not wanting to use a bootstrap for the gate drivers have you though of adding a DC-DC convertor?"""""
Thanks, good point,...hi side power supply plus digi isolator sounds like a plan.

 

[Flyback]

Actually, the NCD57201
https://www.onsemi.com/download/data-sheet/pdf/ncd57201-d.pdf...purports to be able to handle VS (switching node) pin going +/- 800v either side of chip "proper" ground.

I find this hard to beleive, ...if it were capable fo that, then it would have wiped all the other bootstrap drivers off the market?