Author Topic: Mosfet as blocking diode-How? (Read 33558 times)

djsb · « **on:** March 26, 2011, 11:27:01 am »

Someone at work mentioned they where using a MOSFET as a blocking diode. I think the application is solar cells. How is the Mosfet connected and how does it work in this config?
Thanks

David.

Simon · « **Reply #1 on:** March 26, 2011, 11:37:36 am »

maybe a P chanel mosfet ? perhaps with a voltage divider supplying the gate so that if the panels drop below a certainly voltage the mosfet is switched off. Or perhaps there is a full blown control circuit to triget it ?

Simon · « **Reply #2 on:** March 26, 2011, 12:32:42 pm »

You can get mosfets with no internal diode

Zero999 · « **Reply #3 on:** March 26, 2011, 12:47:13 pm »

You can't get MOSFETs without a diode.

When a MOSFET is used as a blocking diode, the internal diode is normally used, it's just bypassed by turning the MOSFET on, so the forward voltage is much lower.

Simon · « **Reply #4 on:** March 26, 2011, 12:48:54 pm »

oh right true I can't think of a part. I thought they were specificly built in.

There must be more than just a mosfet involved then to control it

Zero999 · « **Reply #5 on:** March 26, 2011, 02:03:27 pm »

Quote from: AcHmed99 on March 26, 2011, 01:00:11 pm

Integrated SSR's for several amps start to get pricey for DC/AC. The one I linked to is $6 bucks or so and its only good for a couple of amps at 25C.

It's possible to make an SSR yourself but it's a totally different topic.

oPossum · « **Reply #6 on:** March 26, 2011, 03:01:26 pm »

Low drop rectifier using MOSFET and LTC4412

Note that the MOSFETs are connected "backwards" due to the internal parasitic diode.

Zero999 · « **Reply #7 on:** March 26, 2011, 06:19:47 pm »

You only need one MOSFET, N-channel high side will do.

Drive to the gate can be provided using an oscillator and voltage doubler which is protected against reverse polarity by a normal diode.

I'll post the circuit when I've got more time.

AdShea · « **Reply #8 on:** March 27, 2011, 01:23:02 am »

I've used P-Channel Fets as blocking diodes on MCU boards so they can write to EEPROM on power down. The intrinsic diode on the FET forward biased when the board is initially powered. Initially the gate is pulled high by a high-value resistor. Once the chip boots, it pulls the gate low to reduce the drop when powering the rest of the board. The MCU monitors the supply voltage. If it drops below 4.5V, the MCU releases the gate. If the supply doesn't return after a short time (a few ms), then the chip turns everything off and writes it EEPROM before the 100uF cap on board drops too far.

In this case, the P-FET is used as a diode which can be bypassed by the lower drop of a few mOhms of a fully enhanced FET.

Zero999 · « **Reply #9 on:** March 27, 2011, 08:55:13 pm »

Here's how an N-channel MOSFET can be used to provide high side polarity protection, it's a bit of a fiddle though.

Simon · « **Reply #10 on:** March 27, 2011, 09:13:51 pm »

so how does the mosfet stay on when the 555 switches low to charge the capacitor ? or does the gate capacitance of the mosfet sort that out (hence the 10M resistor) ?

Zero999 · « **Reply #11 on:** March 27, 2011, 09:56:24 pm »

Yes, the voltage doubler charges the gate capacitor. The 10M resistor discharges the gate capacitor when the power is disconnected.

DrGeoff · « **Reply #12 on:** March 27, 2011, 11:08:37 pm »

Mosfets have been used to replace diodes in high current switching applications that require a bridge rectifier without the 2Vf drop (and power/thermal issues that come with that).

http://www.thetaeng.com/FETBridge.htm

scrat · « **Reply #13 on:** March 28, 2011, 12:35:51 pm »

Why shouldn't this work?

oPossum · « **Reply #14 on:** March 28, 2011, 03:58:39 pm »

It won't prevent the load from discharging into the source.

If the source is a PV panel and the load is a battery then it is useless. The MOSFET has to be switched on only when current can flow from source to load. A circuit using LTC4412 or similar chip will do that.

Zero999 · « **Reply #15 on:** March 28, 2011, 05:52:31 pm »

Quote from: scrat on March 28, 2011, 12:35:51 pm

Why shouldn't this work?

It will work if you can find a p-channel MOSFET with the internal diode connected anode to source (you won't).

Leo Bodnar · « **Reply #16 on:** March 28, 2011, 06:04:18 pm »

There are quite a few ready-made ICs that when paired with external MOSFET create almost ideal diode.

Here are few examples:
http://www.farnell.com/datasheets/85443.pdf
http://cds.linear.com/docs/Datasheet/4412fa.pdf
(internal MOSFET) http://www.farnell.com/datasheets/321107.pdf

E.g. google for "ideal diode controller" or suchlike.

Cheers
Leo

RayJones · « **Reply #17 on:** March 28, 2011, 08:02:24 pm »

Search for "Ideal Diode" on Linear Technology's website.

(edit: Whoops, that's what I get for posting from the bottom of page 1 $:-\$ )

scrat · « **Reply #18 on:** March 28, 2011, 09:35:23 pm »

Quote from: scrat on March 28, 2011, 12:35:51 pm

Why shouldn't this work?

Quote from: oPossum on March 28, 2011, 03:58:39 pm

It won't prevent the load from discharging into the source.

If the source is a PV panel and the load is a battery then it is useless. The MOSFET has to be switched on only when current can flow from source to load. A circuit using LTC4412 or similar chip will do that.

I see, I was missing the "blocking" diode purpose. What I posted could be good for reverse polarity protection, not reverse current blocking.

Quote from: Hero999 on March 28, 2011, 05:52:31 pm

It will work if you can find a p-channel MOSFET with the internal diode connected anode to source (you won't).

If I think of an H-bridge high-side p-MOS, it has to have the diode connected that way...
However, I checked this, and it confirms the diode is connected as I have drawn (ex. see http://cache.freescale.com/files/analog/doc/train_ref_material/ANALOGPOWERICTR.pdf?WT_TYPE=Training%20Reference%20Material&WT_VENDOR=FREESCALE&WT_FILE_FORMAT=pdf&WT_ASSET=Documentation).

Jon Chandler · « **Reply #19 on:** March 28, 2011, 10:02:07 pm »

Using a MOSFET for reverse polarity protection actually works pretty well. If the MOSFET is carefully selected the voltage drop is minimal. Either P-channel or N-channel MOSFETs may be used depending on where they are placed in the circuit.

This was discussed on the SparkFun forum sometime back.

I did some tests with a RadioShack MOSFET which wasn't the best choice but it definitely protects again reverse polarity.

See the entire thread here

Zero999 · « **Reply #20 on:** March 28, 2011, 10:04:54 pm »

Quote from: scrat on March 28, 2011, 09:35:23 pm

If I think of an H-bridge high-side p-MOS, it has to have the diode connected that way...
However, I checked this, and it confirms the diode is connected as I have drawn (ex. see http://cache.freescale.com/files/analog/doc/train_ref_material/ANALOGPOWERICTR.pdf?WT_TYPE=Training%20Reference%20Material&WT_VENDOR=FREESCALE&WT_FILE_FORMAT=pdf&WT_ASSET=Documentation).

None of the circuits with p-channel devices, in the paper you posted have the diode connected as per your schematic. All P-channel devices have the cathode connected to source and all N-channel devices have the anode connected to source.

scrat · « **Reply #21 on:** March 28, 2011, 11:20:15 pm »

Quote from: Hero999 on March 28, 2011, 10:04:54 pm

Quote from: scrat on March 28, 2011, 09:35:23 pm
If I think of an H-bridge high-side p-MOS, it has to have the diode connected that way...
However, I checked this, and it confirms the diode is connected as I have drawn (ex. see http://cache.freescale.com/files/analog/doc/train_ref_material/ANALOGPOWERICTR.pdf?WT_TYPE=Training%20Reference%20Material&WT_VENDOR=FREESCALE&WT_FILE_FORMAT=pdf&WT_ASSET=Documentation).
None of the circuits with p-channel devices, in the paper you posted have the diode connected as per your schematic. All P-channel devices have the cathode connected to source and all N-channel devices have the anode connected to source.

Oh... It' just that I've used your schematic (cut and paste on the image), and I didn't flip the MOSFET, since I'm not that familiar with p-channel (it's a little more rare). I just wrote explicitly "D" and "S" not to confuse... Sorry.
My circuit is the same as the second from Jon Chandler.


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: Mosfet as blocking diode-How? (Read 33558 times)

Share me