Reverse polarity input protection

[raptor1956]

I was working in my shop last night and the lights went out but I still had power so I decided to use one of my battery powered LED light panels I use for video as my light source so I still had power to operate my drill press.  Anyway, I brought the light out and plugged it in and -- nothing so I immediately figured I must have plugged it in backwards and went back to the apartment to get my other panel and plugged that in while I still had light and it worked just fine.  So, I completed my work then went back to investigate my non-function panel.

I quickly discovered that the little through hole fuse on the Drok buck converter was blown and when I jumper-ed it the PS came back to life no problem.  I have another fuse on the power input that I checked first and it was OK so obviously a difference in current rating or fast/slow blow type.  So, I got to thinking perhaps I should have a diode in the input circuit downstream from the main input fuse that didn't blow so that if I plug it in backwards again the diode would cause the input fuse to blow.  Is there a particular type of diode best suited for this?

I should mention that the battery I use is a DJI TB-48 drone battery, 6S, that operates between about 20V and 26.3V with a current loading that is a maximum of about 6A from the battery and a bit over 10A output from the buck to drive 2m of LED strips cut 400mm long.


Brian

[SaabFAN]

If the current isn't too high and you are not concerned about efficiency, just use a Diode in series with the Input, or even a bridge-rectifier (then it'll work regardless of the input-polarity).

At high currents, a P-Channel MOSFET combined with a 2 Transistor difference-amplifier to turn it into a so-called "Ideal Diode" drastically reduces your power-dissipation.
Up to 2A you can get away with some beefy Diode. Above that, I would always chose an Ideal Diode.

[raptor1956]

Actually, instead of putting the diode in series which would impact efficiency I thought about putting it across the inputs so that if I reverse the polarity it will pop the main fuse without effecting the PS that's downstream from it.  When operated with the correct polarity there should be almost no impact on efficiency in this configuration.  I would think almost any of the 1N4XXX diodes would do the trick.  So long as the main fuse pop quickly enough there should be very little heating issues with the diode.


Brian

[step_s]

The Diode in reverse bias across the input can work as an extra protection, but wouldn't it be better to just use a decent size mosfet?
http://electronicdesign.com/site-files/electronicdesign.com/files/archive/electronicdesign.com/files/29/9945/figure_01.gif
Also, you could add a red LED with a resistor dropper in reverse bias on the inputs, so that it will glow red when the polarity is wrong, so you don't freak out when it doesn't light up xD

[raptor1956]

Yeah, those would work as well but I don't anticipate frequent issues that might justify the addition of such a circuit.  In addition, the diode "across" the inputs instead of in series with it would have virtually zero voltage drop when connected properly.  A diode across the inputs would cause the fuse to pop but have almost no effect when properly connected.  I just want a protection against damage and am not worried about needing to replace the fuse.  If I were designing this from scratch that would be one thing, but as the PS is an off the shelf and low cost buck converter from Drok at about $13USD I'm not sure adding components with a small additional circuit board or free wired components is justified.  Just throw a diode across the inputs and be done with it.

Of course, even with the diode the PS would still see some negative voltage in the instant before the fuse popped but we're talking 0.7V or less.  A Shottky diode could have as little as 150mV drop and would probably be preferred.


Brian

[step_s]

The Diode in reverse bias across the inputs would have NO effect, if the input is smooth DC, on the functionality. The voltage drop when in forward bias is not significant here, since the amount of current it would draw would be enough to blow the fuse anyway.
How big a fuse are you using? Adjusting the strength of the diode for the fuse, would be preferable. ALso how powerful is the PSU in total?
Most diodes can withstand avery high pulse-current, and if you find the right one, the diode would most likely survive the short-circuit.

As another idea: Add a set of connectors to the output of the PSU and the input of the buck-converter, so that you can't connect it the wrong way around?

[raptor1956]

The maximum draw in use is about 6A at from 20V to 26V depending on charge state of the 6S Lipo battery pack.  I have a 7.5A fuse on input.  I'd anticipate the current through the diode would be perhaps 50A for a mS or so before the fuse would pop.  That amount of current for that short a time period would have a negligible effect on the diode so I would not expect the diode to fail.

Pretty much any other way and you're going to cause a voltage drop ALL the time and you want as little voltage drop with a battery powered device as you can.

The battery pack, a DJI TB-48 with 129WHr capacity is an excellent match for the LED light panel as I expect over an hour at full LED power and over 2 hours at a lesser setting.  The PS, a Drok buck converter is rated at 100W and 8A and 12A max, but I have an enclosure that is cooled by a fan and the exhaust from the plenum where the PS and other controls are located is vented right past the fins of the PS so it gets active cooling.  I have run the PS at about 140W for short periods but don't expect to run it above 100W for any extended periods as that impacts battery life.  In practice the nominal output of the PS is about 10.5V-11V and less than 100W.

The downside to using the DJI battery pack is that the connection is via a pair of spade lugs in line and the connector I made for it can, unfortunately be plugged in backwards as there's no key to prevent that.  I have made labels that should reduce the chance of future f'ups so hopefully this whole diode thing is just insurance.

One further thing about the DJI battery pack...  This is designed for use with a mid-sized consumer drone and produces a flight time of a bit more than 15 minutes though I tend to limit it to about 12 minutes most of the time.  So, it has excellent capacity and good power handling capabilities with the average draw by the drone in the 350W-500W range.  There is smarts built into the pack to prevent over-charging or over discharging and it also controls the charging and does equalization of all the cells.  I have 8 of these $200 items so I might as well make further use of them with these light panels.


Brian

[step_s]

It sounds like you have it all pretty much figured out already
As i said, the trigger time of the diode isn't much of an issue, since it isn't switching or anything, so a regular high power diode will do.
http://www.farnell.com/datasheets/1661896.pdf?_ga=1.83889177.1224770061.1458764725 this costs around 50cent, so it wont be hard for you to find a good and cheap one

Since you have all this expensive gear, and don't seem to be using it with the drones or whatever it was designed for (?), then why don't you create a setup where all the connectors fit each other.
http://www.ebay.com/itm/20Pcs-10Pairs-XT60-Male-Female-Bullet-Connectors-Plugs-for-RC-Lipo-Battery-New-/111834661402?hash=item1a09dd821a:g:RPgAAOSw8-tWWDQ2
XT60 connectors are so dirt cheap, and are standard connectors for LiPO's anyway, so some quick soldering, and voila!

[Zero999]

Actually, instead of putting the diode in series which would impact efficiency I thought about putting it across the inputs so that if I reverse the polarity it will pop the main fuse without effecting the PS that's downstream from it.  When operated with the correct polarity there should be almost no impact on efficiency in this configuration.  I would think almost any of the 1N4XXX diodes would do the trick.  So long as the main fuse pop quickly enough there should be very little heating issues with the diode.


Brian

Yes, that method is frequently used, especially when the circuit is protected by a resettable breaker or PTC resistor.

[raptor1956]

It sounds like you have it all pretty much figured out already
As i said, the trigger time of the diode isn't much of an issue, since it isn't switching or anything, so a regular high power diode will do.
http://www.farnell.com/datasheets/1661896.pdf?_ga=1.83889177.1224770061.1458764725 this costs around 50cent, so it wont be hard for you to find a good and cheap one

Since you have all this expensive gear, and don't seem to be using it with the drones or whatever it was designed for (?), then why don't you create a setup where all the connectors fit each other.
http://www.ebay.com/itm/20Pcs-10Pairs-XT60-Male-Female-Bullet-Connectors-Plugs-for-RC-Lipo-Battery-New-/111834661402?hash=item1a09dd821a:g:RPgAAOSw8-tWWDQ2
XT60 connectors are so dirt cheap, and are standard connectors for LiPO's anyway, so some quick soldering, and voila!

I have sets of XT60 connectors, but I still need to connect to the battery and it is done using in-line spade lugs.  So, I'd still need an in-line spade lug connector to attach to the battery so there's very little point to add an additional layer of connectors (XT60).

I could use another battery of course, but then I'd need to buy them and that seems foolish as I already have perfectly acceptable batteries that are just about the perfect size, voltage and capacity and they have smarts to prevent over-charge, under-charge and to equalize the cells.  I also have a 4-battery charger.  So, using the batteries I have makes sense and therefore I need to deal with the way the battery connects to the other devices.  In the drone the battery slide into a pocket, the drone has the in-line spade lugs for power as well as another connector for communications with the battery so the drone knows what the battery voltage and temp is.  In fact, the drone can know and I as the pilot can know what the individual cell voltages are.

So, no need to add another layer of connectors as that would solve nothing -- I just needed to properly label the connectors I made to reduce confusion and the addition of the diode to pop the fuse is just insurance.


Brian

[lowimpedance]

Be aware that the fast fuse will most likely take longer than "a ms or so" to open and the 1n4xxx rtypes will be more than likely not survive the surge current. (30A for approx 9mS).
So pick carefully the ratings of the diode in relation to the time/current curve of the fuse you are using and the "insurance" cover will be there when needed.