High-side MOSFET driver without bootstrap?

[Tilmann Reh]

Hello,

I am looking for a high-side MOSFET driver that is functionally
equivalent to the IR2117/18, but does not use a bootstrap supply
since I need a static on/off switch (and would like to avoid the
extra supply charge pump, if possible).

Does anyone know of a "self-supplying" high-side driver for
N-channel power FETs? IRF and Intersil (first stops for FET drivers)
don't seem to have anything...

 

[Robert Baer]

Hello,

I am looking for a high-side MOSFET driver that is functionally
equivalent to the IR2117/18, but does not use a bootstrap supply
since I need a static on/off switch (and would like to avoid the
extra supply charge pump, if possible).

Does anyone know of a "self-supplying" high-side driver for
N-channel power FETs? IRF and Intersil (first stops for FET drivers)
don't seem to have anything...

...and you will not find it.

 

[Daniel Haude]

On Thu, 01 Jul 2004 09:03:19 +0200,

I am looking for a high-side MOSFET driver that is functionally
equivalent to the IR2117/18, but does not use a bootstrap supply
since I need a static on/off switch (and would like to avoid the
extra supply charge pump, if possible).

Well, the (in your case, permanent) high-side gate voltage (with a
little current) has to come from somewhere, doesn't it? Let's consider
some options:

1) bootstrap capacitor (not possible for permanent on)

2) charge pump (you don't want it)

3) isolated DC/DC converter

4) Extra small power supply running from mains voltage

5) lithium battery

6) Solar cells

If you don't care about switching speed, power dissipation or top output
voltage I guess an optocoupler and a source follower could be used...

--Daniel

 

[Tilmann Reh]

Well, the (in your case, permanent) high-side gate voltage (with a
little current) has to come from somewhere, doesn't it? Let's consider
some options:

1) bootstrap capacitor (not possible for permanent on)

2) charge pump (you don't want it)

I don't want it as additional components, if possible.
By function, a charge pump is exactly what I need. I would
like to use a driver IC that has this CP integrated.

3) isolated DC/DC converter
4) Extra small power supply running from mains voltage
5) lithium battery
6) Solar cells

All not usable in this case.

If you don't care about switching speed, power dissipation or top output
voltage I guess an optocoupler and a source follower could be used...

I need high current (10A) and minimum voltage drop in ON state,
that's why I will use a high-current N-channel FET (8 mOhm ON).
No alternatives (except for mechanical relays, which I must
replace here).

 

[Tilmann Reh]

...and you will not find it.

You already searched that before?

(Seems like I'll have to build the CP by myself then. Good luck
that a 555 will do ).

 

[Jamie Morken]

Hello,

I am looking for a high-side MOSFET driver that is functionally
equivalent to the IR2117/18, but does not use a bootstrap supply
since I need a static on/off switch (and would like to avoid the
extra supply charge pump, if possible).

Does anyone know of a "self-supplying" high-side driver for
N-channel power FETs? IRF and Intersil (first stops for FET drivers)
don't seem to have anything...

I am not sure if it is what you need but you may want to check out the
HIP4081.
www.intersil.com/data/tb/tb321.pdf

cheers,
Jamie

www.intersil.com/data/tb/tb321.pdf

 

[Ruediger Kluge]

Hi,

Hello,

I am looking for a high-side MOSFET driver that is functionally
equivalent to the IR2117/18, but does not use a bootstrap supply
since I need a static on/off switch (and would like to avoid the
extra supply charge pump, if possible).

Does anyone know of a "self-supplying" high-side driver for
N-channel power FETs? IRF and Intersil (first stops for FET drivers)
don't seem to have anything...

I'm not really sure that I understand the problem right. But Perhaps
Infineon Profet is the right solution. They are used to replace relays
in car industry.

http://www.infineon.com/cmc_upload/migrated_files/document_files/Application_Notes/baba96a1.pdf
Regards
Rüdiger

 

[Winfield Hill]

Tilmann Reh wrote...

You already searched that before?

There are several good n-channel MOSFET high-side switch drivers
with internal supply generators out there. I have their datasheets
stored in my computer; now if I can just recall their part numbers!

There are also some nice p-channel high-side switch drivers.

And there are quite a nice assortment of parts that include the
switching MOSFET or BJT, which allows them to have current limit
and thermal-limit features, complete with fault-reporting outputs.

Let me ask, how big is your MOSFET, fast does your switching time
need to be, and how many times per second do you need to switch it?

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

 

[Winfield Hill]

Jamie Morken wrote...

Tilmann Reh wrote ...

I am not sure if it is what you need but you may want to check out
the HIP4081. www.intersil.com/data/tb/tb321.pdf

The '4081 is one of my favorite parts, and I've written about its
virtues several times here on s.e.d. But sadly it does require a
continuous switching signal to maintain its high-side bias supply,
just like the IR2117, etc., so it won't help Tilman in his search.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

 

[Tilmann Reh]

I'm not really sure that I understand the problem right. But Perhaps
Infineon Profet is the right solution. They are used to replace relays
in car industry.

Thanks.
I know these parts, but was not aware of the high-current types yet.
Seems like the BTS660 is the perfect choice. I'll check for cost
and availability...

 

[Tilmann Reh]

There are several good n-channel MOSFET high-side switch drivers
with internal supply generators out there. I have their datasheets
stored in my computer; now if I can just recall their part numbers!

That's a common problem... and increasing with age .

There are also some nice p-channel high-side switch drivers.

P-FETs rarely have the required RDSon (below 10 mOhm).
And are hard to get, too.

And there are quite a nice assortment of parts that include the
switching MOSFET or BJT, which allows them to have current limit
and thermal-limit features, complete with fault-reporting outputs.

Yes, but commonly not for the specs I need.

Let me ask, how big is your MOSFET, fast does your switching time
need to be, and how many times per second do you need to switch it?

I was thinking about the IRF3205, which is an 8 mOhm 55V type.
Switching time is not critical, as it will be used only to turn
on and off the supply to some power drivers every now and then
(say, each few seconds).

Ruediger Kluge pointed me to the high-current PROFETs from Infineon,
seems like they're the way to go. If they are obtainable...
(Otherwise, using the 555 as charge pump, a few transistors
for level shifting and gate driving, and the IRF3205 as switch
will give a pretty cheap and reliable solution too. It's just
more parts...)

Thanks,

 

[Winfield Hill]

Tilmann Reh wrote...

That's a common problem... and increasing with age .

And with an increasing number of data sheets, I'm up to 18395 now.

Yes, but commonly not for the specs I need.


I was thinking about the IRF3205, which is an 8 mOhm 55V type.
Switching time is not critical, as it will be used only to turn
on and off the supply to some power drivers every now and then
(say, each few seconds).

Ruediger Kluge pointed me to the high-current PROFETs from Infineon,
seems like they're the way to go. If they are obtainable...

Yes, see my assertion above. I've stocked our parts drawers with
five different PROFET BTS-series switches, obtained from DigiKey.
E.g., the BTS650 is rated 6 milliohm max, and is stocked at $6.14
each, $3.40 qty 1k. There are lower-cost parts with higher Ron.

What are you switching, and how high is the supply voltage?

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

 

[Daniel Haude]

On 1 Jul 2004 04:03:03 -0700,

The '4081 is one of my favorite parts, and I've written about its
virtues several times here on s.e.d. But sadly it does require a
continuous switching signal to maintain its high-side bias supply,
just like the IR2117, etc., so it won't help Tilman in his search.

Huh? Fifth item on the data sheet "Features" list: "On-Chip Charge-Pump
and Bootstrap Upper Bias Supply"

While the data sheet and app note don't specifically stress the point it
seems that the internal charge pump can supply sufficient voltage and
current for indefinite intervals. In fact it seems possible to run the
408x without bootstrap diode and cap of course at the price of
"unacceptably long" turn-on time.

I had thought that Tilman, for whatever reason, didn't want any kind of
charge pump (integrated or not), thus my list of intentionally idiotic
gate power supplies.

--Daniel

 

[Daniel Haude]

On Thu, 01 Jul 2004 13:41:25 +0200,

I was thinking about the IRF3205, which is an 8 mOhm 55V type.
Switching time is not critical, as it will be used only to turn
on and off the supply to some power drivers every now and then
(say, each few seconds).

Then the 4081A with almost no external parts seems a good choice. It has
a built-in charge pump and is good for up to 80V positive rail. There
may be something I'm missing about the charge pump (the data sheet says
it has one, Win Hill thinks it doesn't or isn't fit for the task -- see
elsethread).

--Daniel

 

[Tilmann Reh]

And with an increasing number of data sheets, I'm up to 18395 now.

Agreed. (I didn't count the number of DS in my about 12 m of
data books and 1 m of CDs... Is that number exact? )

Yes, see my assertion above. I've stocked our parts drawers with
five different PROFET BTS-series switches, obtained from DigiKey.
E.g., the BTS650 is rated 6 milliohm max, and is stocked at $6.14
each, $3.40 qty 1k. There are lower-cost parts with higher Ron.

What are you switching, and how high is the supply voltage?

I am switching the complete power stages of a special industrial
control unit (one of our custom specific products), rated 24V 10A.
I have to deal with supply peaks of about 50V (limited by TVS).
The BTS 660 seems to be perfect, it is available, and affordable
(about 2.40 Euro in small quantities). I already ordered samples.

(BTS 555 would be a real boomer, but it's more expensive.)

Thanks,

 

[Tilmann Reh]

Then the 4081A with almost no external parts seems a good choice. It has
a built-in charge pump and is good for up to 80V positive rail. There
may be something I'm missing about the charge pump (the data sheet says
it has one, Win Hill thinks it doesn't or isn't fit for the task -- see
elsethread).

It would probably work, but it's waste (it's a full bridge driver),
and so it costs more than the BTS 660 (which already includes
a powerful FET).

I think the ProFET is the perfect solution here.
I just wasn't aware of these high-current types (have been using
smaller ones for years).

 

[Winfield Hill]

Tilmann Reh wrote...

Agreed. (I didn't count the number of DS in my about 12 m of
data books and 1 m of CDs... Is that number exact? )

It was earlier this morning. It's the number of data-sheet files
in my computer's "Companies-Semi" directory, further organized by
company names. It has grown a bit, and now stands at 18434 files.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

 

[Klaus Vestergaard Kragelund]

Tilmann Reh wrote...

It was earlier this morning. It's the number of data-sheet files
in my computer's "Companies-Semi" directory, further organized by
company names. It has grown a bit, and now stands at 18434 files.

With such a large collection - one would think it would be nice to have a
website with all the datasheets. Now thats storage costs is ever decreasing
maybe one day some nice guy/woman is going to upload it all

I myself am only up to 1722 (607MB) by now

Cheers

Klaus

 

[Winfield Hill]

Tilmann Reh wrote...

It would probably work, but it's waste (it's a full bridge driver),
and so it costs more than the BTS 660 (which already includes
a powerful FET).

No, the "charge pump" in the '4080 series works like most other chips,
and only charges an external capacitor (through an external diode) on
each cycle of the driver IC. So if it's continuously on, the charge
drains away and then a sad end to the story. You need a part with an
internal oscillator and charge-pump switch. There are a few such parts.

Or you can use one of IR's unusual "Photovoltaic Isolator" (an IR LED
with a stack of photodiodes), like the PVI5033. This part provides a
slow 5uA gate turn-on current (which means it takes a few ms to switch
large FETs), but has a special circuit to rapidly discharge the FET's
gate capacitance for a "fast" 400us turn off.

I think the ProFET is the perfect solution here.
I just wasn't aware of these high-current types (have been using
smaller ones for years).

They're amazing, especially the elusive BTS550 with its best-in-class
220A current limit. But you should double-check the operating-voltage
specs and consider the time-duration of your over-voltage events.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

 

[R.Legg]

With such a large collection - one would think it would be nice to have a
website with all the datasheets. Now thats storage costs is ever decreasing
maybe one day some nice guy/woman is going to upload it all

I myself am only up to 1722 (607MB) by now

14K files in 3.5G here.

There's no end to this rubbish - you never have the info you need for
the next project.

That's what the mfr's websites are for.

RL