MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Reverse power protection with MOSFETS!
MUFF WIGGLER Forum Index -> Music Tech DIY  
Author Reverse power protection with MOSFETS!
E.A.S.
Here's a circuit I've adapted for some of my designs. Feel free to use it, or part of it for your needs. I first saw an implementation of this in a MAXIM app note.

This is a basic reverse power protection circuit using MOSFETs to provide the protection.
The advantage of using MOSFETs as opposed to a series diode is because you can choose a MOSFET with low Ron so there is VERY little (practically none) voltage drop,
as you would get with a diode, even a schottky.

How it works:
As you see the MOSFETs are connected "Backwards" with the power input on the DRAIN side rather than the SOURCE side. Not shown in the schematic is the "body diode" inside the MOSFET.
The MOSFETs are connected in a way that when correct power is applied, the body diode conducts, the SOURCE is now one diode drop down from the input voltage, and at a level relative to the GATE, to turn on the MOSFET,
this then bypasses the body diode, giving an OUTPUT near equal to the input. By choosing a MOSFET with very low On resistance, this voltage drop can be made to be practically zero.

On my circuit I've added an optional power led / reversed power led, in case you're feeling frisky.

The MOSFETs shown can supply over 1A (1.4A @ 25C) and are little guys. (SOT-23)

Normally for euro, If you are using a 16pin connector and reverse the connector, you will end up shorting the +5V, +12, and internal Gate bus to 0V.
Not good.

Another issue in euro arises when there is a 10pin header on the module and a 16 pin power cable is plugged in, is the possibility of plugging the power in with an offset,
once again shorting one of the power / bus connections to 0V.

My solution would be to use a shrouded 10 pin header. This guaranties, the connector can't be plugged in with an offset. Obviously this isn't perfect because
you only get used of the +- 12V rails. So if you need the CV, GATE or, +5V rail, you're out of luck.

That said, If you really need a +5V rail on your module, you can always locally regulate, down from the +12 anyway.

Basically, you can plug in the power reversed, and wont harm the module, anything else on the power rail, or the power supply itself.
It can also indicate (without smoke signals) when the power is reversed.

-Mike

imcmahon
clever!
kassu
Nice idea. I think I saw something similar on the EEVblog a long time ago, but I have a feeling it used more support circuitry.

E.A.S. wrote:

Normally for euro, If you are using a 16pin connector and reverse the connector, you will end up shorting the +5V, +12, and internal Gate bus to 0V.
Not good.


Hmm, got me thinking, quite a lot of things happen when you reverse a 16-pin connector (let's assume the internal CV and gate are not used)
- The busboards +5V and +12V are shorted together and connected to your modules ground
- Your modudes +5 and +12 are both grounded
- -12V is likely left floating on both sides

So for your module it means +12V and GND are reversed, which would be protected if there are diodes (or mosfets). But for the rest of the system still +5 and +12 are shorted together, and it depends on the power supply what will happen
- If you are lucky the PSU just explodes (or fuses blow)
- Or, if +5 is derived directly from +12, the 5V regulator dies (or is bypassed via a reverse current protection diode), and the 5V rail gets raised to 12V in the entire system. This will certainly kill any digital modules using 5V, potentially quite costly.
Jarno
More info here Geofex
E.A.S.
kassu wrote:
Nice idea. I think I saw something similar on the EEVblog a long time ago, but I have a feeling it used more support circuitry.

E.A.S. wrote:

Normally for euro, If you are using a 16pin connector and reverse the connector, you will end up shorting the +5V, +12, and internal Gate bus to 0V.
Not good.


Hmm, got me thinking, quite a lot of things happen when you reverse a 16-pin connector (let's assume the internal CV and gate are not used)
- The busboards +5V and +12V are shorted together and connected to your modules ground
- Your modudes +5 and +12 are both grounded
- -12V is likely left floating on both sides

So for your module it means +12V and GND are reversed, which would be protected if there are diodes (or mosfets). But for the rest of the system still +5 and +12 are shorted together, and it depends on the power supply what will happen
- If you are lucky the PSU just explodes (or fuses blow)
- Or, if +5 is derived directly from +12, the 5V regulator dies (or is bypassed via a reverse current protection diode), and the 5V rail gets raised to 12V in the entire system. This will certainly kill any digital modules using 5V, potentially quite costly.



Yeah, Ive been trying to think about different ways to fix this shorting together and grounding issue,

I think you could have a MOSFET switch on the two ground lines, that only switch on when there is correct power on the proper rail.

This might add a little bit of resistance the the return path though. depending the the fet used. Havent built it up yet though..

-M
E.A.S.
This adds protection to a 16pin connected module.

Adding to the prior design this adds protection for the power supply itself. Adding protection from shorting the +12 rail to 0V and shorting the +5 volt rail to 0V.

The only thing the circuit does not protect from is shorting the CV Bus output to 0V. Only an issue if:

1) you have modules that bus AND the the CV output circuitry doesn't have protection on its output. (perhaps rare for the circuitry synth modules would use to drive a CV bus)

other additions, would by some ESD protection for the MOSFETs.
a resistor in series with the gate and a small bypass cap from the gate to the drain.

-M

The circuit goes something like this...

E.A.S.
Added a little ESD protection for the MOSFETS..

-M

TheSlowGrowth
The trouble with very low Ron is that noise make its way into and out of a module more easily. The typical Schottky + 10Ohm + electrolytic capacitor arrangement may have significant voltage drop but it also acts as a simple lowpass filter for noise, in either direction.
latigid on
Check out Rossum's power supply solution:

http://www.rossum-electro.com/uncategorized/circuit-protection/

Some will complain that this increases the 0V supply resistance.
E.A.S.
latigid on wrote:
Check out Rossum's power supply solution:

http://www.rossum-electro.com/uncategorized/circuit-protection/

Some will complain that this increases the 0V supply resistance.


Interesting. I actually thought to disconnect pin 9 and 10 as well, but was thinking the 5V would still short to 0V.. but in hindsight the with the (16pin) connector reversed the PSU 0V isnt connected to the 0V of the module so only the shorting is the +12 and +5 together.

I personally wouldn't worry too much about 2 of the 6, 0V connections not being connected.





The optional reverse power LED, only works when using the 10pin header.

E.A.S.
TheSlowGrowth wrote:
The trouble with very low Ron is that noise make its way into and out of a module more easily. The typical Schottky + 10Ohm + electrolytic capacitor arrangement may have significant voltage drop but it also acts as a simple lowpass filter for noise, in either direction.


Which noise are you speaking of?

I guess there's nothing stopping you from choosing a MOSFET with larger Ron, if you really wanted. (or adding a resistor to your design.)

-M
trip
TheSlowGrowth wrote:
The trouble with very low Ron is that noise make its way into and out of a module more easily. The typical Schottky + 10Ohm + electrolytic capacitor arrangement may have significant voltage drop but it also acts as a simple lowpass filter for noise, in either direction.


Using this very setup with; 10ohm resistor, 10u electro cap and S1JL diode I was getting a 600-700mv drop, which is pretty significant. There's no reason you couldn't replace the fuse in this schematic with a resistor (I guess f1/f2 are fuses?), and the 10u caps are already there.
Graham Hinton
You can't polish a turd. Why try?

Any half decent regulator will shut down when shorted to 0V. Common linear regulator ICs do this faster than a fuse can blow. The MOTM 4-way connector may sometimes be mated one pin out and all that happens is one rail is shorted to 0V and the other disconnected.

The problem with mixed analogue and digital systems is the +5V getting either analogue rail shorted to it and taking all the 5V ICs out. This is usually dealt with by having a crowbar circuit on the +5V, which is a power thyristor that triggers on the +5V being exceeded and deliberately shorts the rail to force the regulators to shut down. You have to turn off and fix the fault.

There should be circuitry on a module to protect the rest of the system from a fault on that module, but it shouldn't be trying to compensate for bad system design. Isn't it easier to just put a better connector on your own modules instead of perpetuating a mistake?
trip
Attached eagle library for sot-23 n and p mosfets, as yet untested but should be fine.
trip
The named mosfets in the schematic up top are out of stock everywhere atm with fairly long lead times - these parts seems to be superior as far as I can tell for the N channel:

https://www.mouser.co.uk/ProductDetail/Diodes-Incorporated/DMN3023L-7? qs=sGAEpiMZZMshyDBzk1%2fWiw4EkXjINyPrH7PGy3HZ4FBEKOATXpiIxA%3d%3d

and for the P channel:

https://www.mouser.co.uk/ProductDetail/Diodes-Incorporated/DMP3099L-7? qs=sGAEpiMZZMshyDBzk1%2fWi%252bzdcdBSapG%252bDctwa3X8OfU%3d

They cost a few more pence each than the other named parts, and have Ron of 25 and 65mOhm respectively.

Unless there's something I'm missing on the datasheet that would make them unsuitable?
MUFF WIGGLER Forum Index -> Music Tech DIY  
Page 1 of 1
Powered by phpBB © phpBB Group