MOSFET gate left floating?
i...@easystreetrecumbents.com
In an effort to reduce the cost of parts on my slowly evolving signal
light system for bikes, I am considering replacing the solid state
relays that it uses with power MOSFETs. They seem well suited for the
purpose, and for my power requirements (up to an amp or two at 12V) they
can come quite small . . . down to a SOT-23, which is compact but can
still be soldered onto a board without a robot's help.
I am figuring on using N-channel enhancement mode FETs, so switching a
low positive voltage to the gate would let power flow through the FET.
But, I have heard that MOSFET gates do not necessarily bounce back to
"off" once the gate voltage is removed, and may "float" unless a
negative charge is put to them. Of course, this problem is easily
solved by connecting the gate to ground through a high-ohm resistor.
The problem is that with several little MOSFETs crammed together on a
board, I have to do quite a bit of extra plumbing to link them all back
to ground. Moreover, I need a separate high-ohm resistor for each
gate. This can be done by taking lots of space for separate resistors,
or by using tiny little resistor networks that I probably will need a
robot to solder on. (I ordered a few and inspected them...and given the
trouble I must take to solder a 1206 chip resistor on its two ends, I
cannot see attaching it at 8 separate places, but I can try...)
So my question is... just how big is the problem of floating gates?
Since I am not doing "fast switching," just clicking my turn signals on
or off at a typical rate, will a MOSFET gate close well enough if not
connected to ground? Granted, I can do some tests once I get the chance
to put these little FETs on a board, but I thought I'd see if anyone had
an opinion...
--
Mike Librik, LCI #929
Easy Street Recumbents
(512) 453-0438
45th and Red River Streets, thereabouts
Central Austin
in...@easystreetrecumbents.com
www.easystreetrecumbents.com
www.urbancycling.com
Willie Hunt
A floating MOSFET gate will more or less stay at the voltage where it was left. Thus if you turn in on, it will stay on. If you turn it off it will stay off. However, I do not recommend banking on this as small leakages caused by all sorts of issues with your insulators (PCB's, solder flux, moisture, etc) will cause the gate to charge up further or discharge further which could leave the MOSFET operating in a linear mode and possibly over heat it depending on the size of the MOSFET and the load. The SOT23 sized MOSFET's can only handle a watt or so before they fail. So, if you turn on a 30mA LED, no problem, but if you turn on a 12 watt 12 volt tungsten, the will likely die. Also, note that you need several volts on the gate with respect to the source to turn on a MOSFET fully. Connecting the gate to drain will not turn it on fully and it will overheat with sufficient load.
Willie
________________________________
--
To subscribe to this list send a BLANK email to bikecurren...@topica.com
To unsubscribe from this list send a BLANK email to bikecurrent...@topica.com
Steve Kurt
hi Mike,
the quick answer is: yes, a floating gate is a bad thing. It's
possible for noise to pull the gate high briefly, or repeatedly. Not a
good thing. Add a pull-down resistor. You may also consider adding a
small cap from gate to ground.
I'm betting that International Rectifier has some good app notes on this
sort of thing...
Steve Kurt
Willie Hunt
A floating gate in and of itself will not kill a MOSFET. However if noise or ESD discharge cause the gate to source voltage to exceed the gate oxide breakdown then the MOSFET will suffer permanent damage. Some power MOSFET's such as Rohm devices have built in back to back zeners to limit gate voltage and prevent damage. However, most do not have any protection at all. CMOS logic gates normally have power and ground clamp diodes, which provides good protection if the gate is wired to a power supply with bypass capacitors. But a loose chip can be easily damaged because the total chip voltage can exceed the breakdown voltage.
With the small SOT23 MOSFET's the total gate charge that will result in well over maximum Vgs is small. A human body ESD discharge event that is unperceivable to the human can easily destroy the MOSFET. With large FET's such as 30 volt, 100 Amp plus rated (IE Rds on << 0.01 ohms) they can take quite much more abuse, but they too can be damage be ESD and or noise.
Willie
-----Original Message-----
From: Pawel Danielewicz [mailto:dani...@msu.edu]
Sent: Friday, December 08, 2006 5:24 AM
To: Willie Hunt
Cc: in...@easystreetrecumbents.com; bikec...@topica.com
Subject: Re: [BC] MOSFET gate left floating?
Willie Hunt wrote:
> A floating MOSFET gate will more or less stay at the voltage where it was left. Thus if you turn in on, it will stay on. If you turn it off it will stay off.
I suggest experimenting with a single MOSFET outside of the intended
circuitry. You will find that touching the floating gate with your
finger can be used to toggle the MOSFET state. Occasionally, a floating
gate can even kill a MOSFET. On the average, due to higher associated
capacitances, they are better though than CMOS IC inputs.
Pawel Danielewicz
Pawel Danielewicz
Suz...@csiro.au
Mike Librik wrote:
> So my question is... just how big is the problem of
> floating gates?
You absolutely can't leave a MOSFET gate floating. A typical power
MOSFET has about 1nF gate capacitance, and gate leakage measured in
picoamps.
That means that when you turn it on, it'll stay on for a very long time.
It'll also be very susceptible to static discharge.
I've actually tested MOSFET Rds by biasing the gate on with my
multimeter in ohms mode, then measuring across the drain-source. Even
though the gate's not connected when I do the Rds measurement, it's
still got plenty of charge.
So you'll have to either put a resistor grom gate to source, or else
drive them with an active circuit that both sources and sinks. TTL
logic gates are good for low frequency work.
Cheers,
Suzy
jk...@speakeasy.net
Would someone explain what a MOSFET gate is please?
Barry
Pawel Danielewicz
Willie Hunt wrote:
> A floating gate in and of itself will not kill a MOSFET.
Sure, it can provoke the killing by making the MOSFET vulnerable to
stray charges.
> However if noise or ESD discharge cause the gate to source voltage to exceed the gate oxide breakdown then the MOSFET will suffer permanent damage. Some power MOSFET's such as Rohm devices have built in back to back zeners to limit gate voltage and prevent damage. However, most do not have any protection at all. CMOS logic gates normally have power and ground clamp diodes, which provides good protection if the gate is wired to a power supply with bypass capacitors. But a loose chip can be easily damaged because the total chip voltage can exceed the breakdown voltage.
>
Well, I think I never managed to get through with incorporating a CMOS
IC into the circuitry and not blowing the IC on the way. I decided to
give up on the associated power gains and stay on the safe TTL side for
projects carried out from scratch.
> With the small SOT23 MOSFET's the total gate charge that will result in well over maximum Vgs is small. A human body ESD discharge event that is unperceivable to the human can easily destroy the MOSFET. With large FET's such as 30 volt, 100 Amp plus rated (IE Rds on << 0.01 ohms) they can take quite much more abuse, but they too can be damage be ESD and or noise.
>
> Willie
>
> Would someone explain what a MOSFET gate is please?
> Barry
Gate plays a role similar to the base in a bipolar transistor: it
regulates significant currents between the MOSFET source and drain
terminals. In contrast to the base, however, which sinks current, the
gate is isolated, to a very good degree (tens of megaohms), from the
rest of the MOSFET. This is a benefit since there is no power cost but
also means that charge may stay forever in the gate, keeping the MOSFET
in one state of conduction or another. Also, an excessive charge on the
gate can lead to a discharge permanently altering the MOSFET.
Pawel
Steve Kurt
regarding questions about mosfets and how to use them....
I'm fond of International Rectifier's application notes. Of course, I
became familiar with them back in the pre-internet days when databooks
were made of paper and quite valued.
The datasheets and application notes are online nowdays, but still just
as useful. International Recifier's app notes are at:
http://www.irf.com/technical-info/appnotes.htm
Scroll down for some general topics:
AN-936: The Do's and Don'ts of Using MOS-Gated Transistors PDF
AN-937: Gate Drive Characteristics and Requirements for HEXFETŽ power
MOSFETs PDF
AN-955: Protecting IGBTs and MOSFETs from ESD PDF
The other semiconductor manufacturers have similar app notes that can
tell you how to use their parts, and how to avoid the predictable
problems.
happy reading,
Steve Kurt
p.s. some of the best app notes were written by Bob Pease at National
Semiconductor. Very practical, and great info. He even wrote a app
note on how to read and interpret a datasheet! I wonder if that's online?
Willie Hunt
That's being a bit "absolute" ! :)
You certainly can leave the gate floating, but the results may not always be desirable. Back in the early 80's I build some touch circuits where the user's finger completed the circuits to charge or discharge power MOSFET gates. They worked fine for a long time while the humidity was high (this was in Houston TX), but in the winter (lower humidity) the MOSFET gates got damage by ESD. So, I installed 1N4148 clamp diodes to GND and the positive power supply and the circuit was used for years. It would hold the lights state for a day or so, but I typically only left the lights on for a few hours at a time.
Yes, I have routinely used the DVM diode test scale to precharge or reverse charge a MOSFET gate to test it's operation. This also validates that the gate is not too leaky.
Willie
________________________________
Mike Librik wrote:
Cheers,
Suzy
Karen Sutton
From: <jker...@speakeasy.net>
>
> Would someone explain what a MOSFET gate is please?
A mosfet is a metal oxide semiconductor field effect transistor. The gate
is the connection that controls current flow through the transister:
http://www.mtmi.vu.lt/pfk/funkc_dariniai/transistor/mosfet.htm
- Karen