back to back MOSFET's as a switch

[Simon]

OK, so I have it, each mosfet needs it's own gate resistor and it's own drive transistor.

[szan]

so what you need is just a single ideal diode?

Maybe by adding hall effect current sensor to detect unwanted current flow into battery to drive AC mosfet switch to close and open AC mosfet switch when battery is discharged with given outpuit current.
This could also allow to make more triclky things like allow flow to battery very limited small current eg. for balancing purposes, and limit output current when needed because of we have current sensor capable to measure positive and negative current flow ?!

[DavidAlfa]

Try this. It allows bidirectional current flow. So you can either source current (power from the battery) and sink (From external power, to charge the battery)

[Simon]

Try this. It allows bidirectional current flow. So you can either source current (power from the battery) and sink (From external power, to charge the battery)

Yea, I did some paper sketching and came to the conclusion that this is what I need. The problem is that I have a device on the battery side that when unpowered pulls the gate down and allows backflow. I do not want bidirectional flow through this switch but it is the same function. The problem is that the batter could be at 12V with 36V on the outside, this could bias the mosfets which is bad news. The charging is done by an onboard switch-mode converter that will accept up to 40V and charge the batteries.

[DavidAlfa]

Well, everything is possible  . This is for best efficiency, almost zero loses:




If you don't mind losing some power, you can simply use a schotkky diode to prevent current flowing back.
Ex. Schottky diodes drop ~0.25V,  so at 10Amps it would drop only 2.5W.

[langwadt]

Try this. It allows bidirectional current flow. So you can either source current (power from the battery) and sink (From external power, to charge the battery)

Yea, I did some paper sketching and came to the conclusion that this is what I need. The problem is that I have a device on the battery side that when unpowered pulls the gate down and allows backflow. I do not want bidirectional flow through this switch but it is the same function. The problem is that the batter could be at 12V with 36V on the outside, this could bias the mosfets which is bad news. The charging is done by an onboard switch-mode converter that will accept up to 40V and charge the batteries.

if you don't want bidirectional flow through the switch you need the ideal diode circuit, the switch will happily conduct in both direction once it is on

[Simon]

If you don't mind losing some power, you can simply use a schotkky diode to prevent current flowing back.
Ex. Schottky diodes drop ~0.25V,  so at 10Amps it would drop only 2.5W.

If you can recommend a diode that will drop 0.25V at 10A I would be very grateful  and you would be sitting on a gold mine. My experience is that all Schottky diodes drop 0.6-0.7V at any sensible current flow. Just as silicone diodes claimed to be 0.7V actually drop 1-1.1V at normal flow. The headline datasheet figures are just for when you perform a test with a really small current compared to their nominal. When I was a kid I discovered 0V drop 1N4007 diodes. I put several in series with a 9V battery to reduce the voltage to 5V but my multimeter still measured 9V!

[DavidAlfa]

Yep, my bad, definitely not at 10amps.
I've seen them dropping 90-120mv at the mA range.
However there're still pretty decent diodes, like STPS10L60, 0.48V@10A typical.

[Peabody]

I've tried to follow this thread, but have never understood your circuit or what you want to accomplish.  Do you want current to flow in either direction or not?  Could you possibly post a simplified version of the whole circuit showing where the charger is and which part you're concerned about when it's powered down?  If that part pulls down the gate, does it not also pull down the source?  Also, if a gate is controlled through an N- part, it could not be pulled down unless the part controlling it is powered up.

And finally, if current is really only going to flow in one direction, is it not possible to use a single mosfet with its body diode oriented to block the undesired back flow?  That may seem to be a "backwards" orientation, but current will flow in either direction when the mosfet is turned on.

[Simon]

Basically there is a battery with built in charger. The charger has a mosfet that will block power coming in when not enabled. The battery output is controlled by this double mosfet, both output and charger input are on the same pin. It needs to be a double one because there should not be any output if it has not been asked for and when external power is applied to be put through the charger it must not enter directly to the battery so current must be blocked going both ways. The last schematic is as generic as it gets ignoring another mosfet that feeds the charger.

The issue seems to be that there in unpowered circuitry that is pulling one source or drain down and it doing so will activate the other mosfet.

[BrianHG]

Looking back on some of Simon's comments and schematics, I'm confused.  Simon, is all you want is a super low drop diode, IE an ideal diode?

These things exist...

Example 1, a solar-panel diode: SM74611KTTR from texas instruments.
Up to 15 amps, only a 0.026v drop at 8 amps.
This is nothing more than a mosfet with a built in ideal diode gate driver with charge pump for the gate drive.
Note that the charge pump means that at around 32Khz, for a tiny fraction of time, the Mosfet auto-turns off to check for a voltage drop and to build charge to drive the gate.  The 0.026v drop is the total average voltage drop over the 32KHz time.  So this guy's drop is actually less and it makes almost no heat at full load.

Note that ICs exists specially to drive a mosfet to do the same as the TI solar-panel ideal diode.  I don't remember who, but you would need to Google along the lines of 'mosfet ideal diode controller'.

Also, have you searched FERD diodes.  I bet if not 1, 2 to 4 in parallel will get below 0.25v drop at 10 amps.

[BrianHG]

Here is an 'Ideal Diode Controller' from linear technology:
LTC4357

[dietert1]

Often the two back to back mosfets are wired with sources and gates in common. Combine that with a photovoltaic optocoupler like PVI1050 to drive the mosfets. Very simple circuit with isolation between drive and output - a perfect replacement for a relay contact.
About 10 years ago i found that circuit is a nice mod for output relays in audio amplifiers and active speakers, for reliable operation and better sound.

Regards, Dieter

[Simon]

Yes such an ideal diode could work I need to look into this more. As things stand with parts supply I'll be using relay's that look like they are cheaper and less resistance.

[BrianHG]

Yes such an ideal diode could work I need to look into this more. As things stand with parts supply I'll be using relay's that look like they are cheaper and less resistance.

Yes, relay is cheaper.
Look at the last page of the LTC4357 datasheet.  They have others in the same family which may replace microcontroller control, or control 2 mosfets from 2 power sources with built in logic and status flag outputs including 20v versions designed for heavier currents.

[Simon]

I think if I use mosfets I need something like the attached.

[langwadt]

I think if I use mosfets I need something like the attached.

keep in mind that the right fet will not work like a diode once it has been turned on. So if you turn on the left fet and then add a supply to output it'll feed the battery and you have no  way to stop that

[dietert1]

The back-to-back mosfet relay is lower resistance, if you select the right parts. Typical contact resistance of a new relay often is 10 or 20 milliOhm, easy to beat with mosfets if you keep them cool. While a new relay may perform well, after some time contacts will deteriorate. Mosfets always switch the same.

Regards, Dieter

[Simon]

The back-to-back mosfet relay is lower resistance, if you select the right parts. Typical contact resistance of a new relay often is 10 or 20 milliOhm, easy to beat with mosfets if you keep them cool. While a new relay may perform well, after some time contacts will deteriorate. Mosfets always switch the same.

Regards, Dieter

Have you tried to buy mosfet's recently?

[Simon]

I think if I use mosfets I need something like the attached.

keep in mind that the right fet will not work like a diode once it has been turned on. So if you turn on the left fet and then add a supply to output it'll feed the battery and you have no  way to stop that

Yes. What I am doing is measuring current consumption. as soon as it drops below the known loads idle current the output goes off so you cannot backfeed. If there is voltage on the output already the output will not turn on. Yes I have states galore in the code.

[langwadt]

I think if I use mosfets I need something like the attached.

keep in mind that the right fet will not work like a diode once it has been turned on. So if you turn on the left fet and then add a supply to output it'll feed the battery and you have no  way to stop that

Yes. What I am doing is measuring current consumption. as soon as it drops below the known loads idle current the output goes off so you cannot backfeed. If there is voltage on the output already the output will not turn on. Yes I have states galore in the code.

if you add a supply while the left fet is on it is too late, the right fet will not turn off and the left fet has the body diode int he back feed direction

[dietert1]

The back-to-back mosfet relay is lower resistance, if you select the right parts. Typical contact resistance of a new relay often is 10 or 20 milliOhm, easy to beat with mosfets if you keep them cool. While a new relay may perform well, after some time contacts will deteriorate. Mosfets always switch the same.

Regards, Dieter

Have you tried to buy mosfet's recently?

Yes i got mosfets and PVI1050 optocouplers from Digikey,  a similar use for input protection of nanovoltmeter. To get a start, you could try IRFB4110: https://www.ebay.com/itm/264780371590?hash=item3da623d686:g:-oIAAOSwhe1e-rCd

Regards, Dieter

[Simon]

This is only 3A, I was looking at SO8 mosfet's, all gone.

[DavidAlfa]

For only 3 amps, I'd use the schottky diode.
You have to be careful with that voltage divider at the gate.
Under any circunstance the voltage at the gate must be between 12V and 18V.
Less and it might overheat, more and will break down.

This resistor values will work in the range of 30-48V.
I you're absolutely sure the voltage isn't going to exceed 48V, you can remove the zener.
If the voltage can go below 30V, then it'll need further tweaks.
I'd make an additional security check by reading the external voltage and only turning on the fet if it's within limits.
A simple zener and some resistors...

[Simon]

The battery (with built in charger) will only be connected to it's device or a supplied power supply. If the user wants to mishandle it that is beyond the warranty. I know the voltages I plan to use for the charger so no problem there.