FET High Side Switch

[mikep]

Does anyone have experience with FET high side switches such as Fairchild FDC6324, FDC6330 or the Vishay equivalent SI3861?

Is there a better alternative? The idea is to switch on and off a 12V load using a 5V signal. These switches seem better than just a simple FET.

[dkinzer]

[spamiam]

I have been working with high-side switching recently. I have needed to switch into a 50mH, 6 ohm load. It is surprisingly not simple.

The issue with the FET's, in my case I am using a MOSFET is getting it to turn on "hard". It takes 10V above the source voltage. Since this is a high side switch, the source voltage at the switch is going to be well above ground.

I ended up using a LTC1154 high side driver. It has a charge pump that brings to gate drive up to about input + 10V. Unfortunately, these are not cheap. $3-$5. and it is hard (now) to get them in DIP form.

I have been looking for a cheaper high side driver with a charge boost (and no external R's and C's), but nothing jumps out at me.

With this type high side driver you can use about any N-channel MOSFET and have rather low ON resistance. For low current application I find the BS-170 pretty convenient.

-Tony

[EDIT]

Here is a list of some of the high side drivers I have considered, though I eventually settled on the LTC1154

MIC5011
LTC1255/1154/1155/1161/1477
IR331x
IPC6011
MAX1614
MC33198
HIP5080A/4081A

-T

[mikep]

[spamiam]

[spamiam]

I am attaching a SPICE circuit to show the comparatively poor switching speed of a circuit similar to the FDC6330.

The SPICE program is a really nice one from Linear Technology. I use it all the time to play with circuits to get a general idea about how they will work. I still blow out stuff, but it is much less frequent. Actually the models are a very close approximation of what I find in real-life with an oscilloscope. Unfortunately, I do not have the training to predict failure modes. The last time, it was pretty impressive, a small electrolytic cap EXPLODED! Its guts shot all over and I saw small flames!

website: http://www.linear.com/designtools/software/switchercad.jsp

This circuit uses a 5KHz squarewave signal as the input. A relatively large R1 (10K as specified by Fairchild) only slowly pulls up the P-FET due to gate capacitance. This causes relatively slow turn-off. A side effect of this causes the FET to spend a while partially conductive. This increases heat dissipation.

If the switching frequency is quite low, then this will not be as much of an issue.

Mike, what switching frequency were you planning?

-Tony

[mikep]

[mikep]

[spamiam]

[DocJC]

Kind of an old thread, but for what its worth...

I used IR3310 Power Mosfet High Side switches in an automotive application. They worked well. I used 2N2222's to buffer the 5V uC up to the 12 switching signal for the devices. Rds is 7 mOhms, and I fused each of the 8 channels at 20 Amps each, well within spec's.

The Control Console used a ZX-24. This power controller unit did not require a ZX. If I re-did this board I might swap the chip used for a ZX-328n, which I have not yet played with.

JC

[spamiam]

JC, It looks very well done! Compact too. What do you use it for?

-Tony

[DocJC]

Hi Tony,

Not to hijack the thread, but...

I drive an emergency vehicle. The unit switches the light bar, dash light, strobes, siren, radio, etc., power on and off. MUCH easier than routing lots of heavy gauge wires through the firewall to a bank of monster 10-20 A toggle switches.

Lots of lessons learned on this project.

One was that I set the external resistor on the high side switch for the mini-light bar on the roof for 15 A, or whatever it was, (2007 project). It worked fine on bench testing. Didn't work at all when installed, (major hassel, let me repeat, MAJOR hassel). Other channels worked. Pull it out from behind the grill, unseal the case, carry the scope to the garage...,

To cut a long story short I had not allowed for enough surge current on start up. The two halogen light rotators draw lots more as a transient at startup with cold filaments, and the motors also have a start up peak. All of them, of course, draw their maximum at the same time, on power up. The switch would shut itself off. This one does not auto restart, you have to cycle its power. Turn off, turn on, high peak, instant off. Repeat. Took me a while to figure it out...

JC

[spamiam]

[DocJC]

I swore at myself a few times... Then I uninstalled the entire thing, back to the basement workbench, more testing, and changed the Imax shut down limit (resistor).

I had measured the steady state current for the various sub-components, but I had not allowed for enough safety margin for the start up. Especially with four devices on that one circuit, all with high initial transients.

I did not, then, have a storage scope, so measuring the start up transients would have been a separate, (albeit doable), project, which I did not want to undertake. Only built one of them, not a production device.

I blew the fuse in my multimeter measuring the currents, by the way. Very tough to get a replacement fuse. And the fuse did not match the manual, and did not match the writing on the meter's board... I also found a cap with a free floating lead inside the meter when I took it appart to fix it... Funny how simple projects cascade... (Needless to say it wasn't a Fluke!).

Unlike current LED light bars, and strobes, the old fashioned halogen rotator's brightness is a function of the supplied voltage. Brighter is better. It is easy to see a light bar at dawn, dusk, and nighttime; but more difficult in bright sunlight, and it impacts one's intersection warning & safety. Hence a series resistor, (or RC), is a sub-optimal choice for that reason alone, plus the size of a suitable power resistor, (potentially x 8 ), heat dissapation, etc.

The Imax in the high side switch is the primary overcurrent fault detector/corrector. The fuses are a back-up, and incredibly difficult to access one installed.

More than you probably wanted to know, but I enjoy the (rare) opportunity to chat about it.

JC

[spamiam]