Author Topic: Idiot proof Breadboard Over Voltage protection needed.  (Read 2094 times)

0 Members and 1 Guest are viewing this topic.

Offline castingflameTopic starter

  • Regular Contributor
  • *
  • Posts: 100
  • Country: gb
Idiot proof Breadboard Over Voltage protection needed.
« on: February 14, 2019, 12:42:38 pm »
Hey all.

I am in need of a circuit that I can make on vero to use to protect my 3.3v micro-controllers from over-voltage protection on its power lines. 

Scenario: Attaching 12v to the 3v3 supply, breakdown of a 3.3v regulator shorting to 12v, other component failure, human error or plain getting things wrong.

I typically have several voltages on the same breadboard. My power wires are in different colours, I label the breadboard common voltage rails, I label everything. Alas things still go wrong and I am just fed up killing my little BluePills!


I have looked and played with Zeners and tried a few circuits but I am not really happy with any. I need a circuit that goes in between the micro-controller and its power supply so ANYTHING over 3.3v on the 3.3v rail will not get to it.


I know the majority of the time it is my own fault, but it is what it is so I want to try and make a little horizontal circuit on vero that I can just plug into the breadboard between the MCU and its supply.


Thanks for any help.


P.S. Please treat m gently!
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: Idiot proof Breadboard Over Voltage protection needed.
« Reply #1 on: February 14, 2019, 01:47:30 pm »
Fuse + n-MOS + voltage supervisor at around 4v or so, like BD5340. Better pick supervisor with push/pull (CMOS) output like BD5340, not open-drain. Circuit is very simple: N-MOS shorts 3.3V rail to ground blowing fuse when voltage of the supervisor IC is reached. Disconnect is worse than crowbar solution I suggest.
« Last Edit: February 14, 2019, 01:50:11 pm by ogden »
 
The following users thanked this post: castingflame

Offline castingflameTopic starter

  • Regular Contributor
  • *
  • Posts: 100
  • Country: gb
Re: Idiot proof Breadboard Over Voltage protection needed.
« Reply #2 on: February 14, 2019, 03:15:39 pm »
Thanks for your response ogden.

Re: BD5337 - CMOS 3.7v threshold voltage detector

From the data sheet ...'When the voltage applied to the VDD pins reaches the applicable threshold voltage, the VOUT terminal
voltage switches from either “High” to “Low” or from “Low” to “High”. '

I have been trying to wrap my head around what you are saying. I have looked at the BD5340 datasheet and from what I understand it can flip/flip its Vout state when its Vdd pin reaches the 'detection threshold' say 3.7v for the BD5337.

Are you saying that BD53xx Vin =3.3v rail, Vout of the BD53xx is then fed into a MOSFETs Gate. Then the Drain to 3.3v and the Source to the fuse and the other side of the fuse to ground? ... 

This is no doubt wrong and I am unsure where the MCU connects (or is it a global 3.3v rail protection?).

So I can understand that this is a very useful component but my limined knowledge is struggling how to put it all together in a diagram.

Thank you so far.
« Last Edit: February 14, 2019, 03:18:29 pm by castingflame »
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: Idiot proof Breadboard Over Voltage protection needed.
« Reply #3 on: February 14, 2019, 03:46:42 pm »
Thanks for your response ogden.

Re: BD5337 - CMOS 3.7v threshold voltage detector

From the data sheet ...'When the voltage applied to the VDD pins reaches the applicable threshold voltage, the VOUT terminal
voltage switches from either “High” to “Low” or from “Low” to “High”. '

They mean that treshold voltage can be reached from below or from above. You are interested "reached from below" which means that output goes from low to high. which is what you need to drive logic level N-MOS (with VGS treshold <=2V).

Quote
Are you saying that BD53xx Vin =3.3v rail, Vout of the BD53xx is then fed into a MOSFETs Gate.

Yes.

Quote
Then the Drain to 3.3v and the Source to the fuse and the other side of the fuse to ground? ... 

Following circuit is more or less functional equivalent.

https://en.wikipedia.org/wiki/Crowbar_(circuit)


It is easy to "convert" this circuit to one I suggested, just put supervisor in place of LM431 and replace triac with logic level N-MOS (VGS treshold <= 2V) with big enough current capability (10 A or so). Just noticed that supoervisor I suggested, have delay function which is bad for your application. You have to find one w/o delay.
 
The following users thanked this post: castingflame

Offline castingflameTopic starter

  • Regular Contributor
  • *
  • Posts: 100
  • Country: gb
Re: Idiot proof Breadboard Over Voltage protection needed.
« Reply #4 on: February 14, 2019, 04:15:09 pm »
YI had the fuse in completely in the wrong place.

I have some heavy duty logic level MOSFETs in stock.

'Just noticed that supervisor I suggested, have delay function which is bad for your application. You have to find one w/o delay.' Okay, will have a search around.

Thanks again. I'll feed back when I have something.
 

Offline mvs

  • Frequent Contributor
  • **
  • Posts: 370
  • Country: de
Re: Idiot proof Breadboard Over Voltage protection needed.
« Reply #5 on: February 14, 2019, 07:13:08 pm »
Following circuit is more or less functional equivalent.

https://en.wikipedia.org/wiki/Crowbar_(circuit)


This crowbar circuit relies on latching property of TRIAC and you can't simply replace it with MOSFET or BJT.
 
The following users thanked this post: castingflame

Online Doctorandus_P

  • Super Contributor
  • ***
  • Posts: 4004
  • Country: nl
Re: Idiot proof Breadboard Over Voltage protection needed.
« Reply #6 on: February 15, 2019, 12:40:25 am »
Instead of focussing on electronics / crowbar you might want to think about simple hardware.

Glueing your PCB on a piece of wood and adding some decent connectors / cables and a power supply with adjustable voltage / current limit goes a long way.

If you use for example a 10-way connector you can have a few GND leads, dedicated leads for 3V3 and 5V and some free leads for auxilary voltages or signals. If your connecter is keyed so it cannot be plugged backwards you're already pretty safe.

Making a dedicated power supply for the other end of your cable can also be simple. A few of the Chinese SMPS modules with adjustable current limit and some panel meter displays can be enough.

What kind of setup for powersupply and cables do you have now?
Banana plugs and breadboards are .... and unwieldy combination  ;D


 
The following users thanked this post: castingflame

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Idiot proof Breadboard Over Voltage protection needed.
« Reply #7 on: February 15, 2019, 01:18:02 am »
Idiot proof?

You'll just find an ever-more-ingenious way to idiot things up! :-DD

Yeah, I've been there. :P

(Example, crossing the wires after the protection; forgetting to install the protector circuit; etc.)

That said, what kinds of solutions would be most likely to be followed?

The simplest may be using a carefully current-limited supply.  You'll get good regulation, and stable current limiting, with a standard constant current source circuit; putting two, back to back, also gets overvoltage protection (say, cross-wiring the 12V supply to the 3V rail), and doing it on the ground (if it's wired separately from the other supplies) also gets you polarity protection.  If you don't mind buying parts you probably don't have on hand already -- a few depletion mode MOSFETs does this nicely, and are available to over 200V and a few watts dissipation (almost good enough you could cross-wire mains to the poor thing!).

Limiting the current to a few 10s of mA, and not placing a large capacitor on the supply's output, allows the possibility of survival, even if you cross-wire that circuit to VCC or an IO pin.  If more current is needed, you simply don't have any way to assure that, and you'll have to box things up and protect all the wires entering and exiting that box.

Example, using a dev kit with MCU onboard, and protection wired up to all the pins.  Maybe clamp diodes, so excess voltage/current gets dumped into the supply, then the supply itself can be protected with a TVS (if ~5V) or crowbar (if less).

Incidentally, even if lower voltage TVS diodes are available, they really don't perform any better than 5V parts.  To get guaranteed low clamping (like 4.0V for a 3.3V supply -- a lot of 3.3V parts catch fire in the 3.6 to 4.4V range), you need to use an active circuit, potentially with fast response time and very high current capacity (in the event you wire a charged capacitor across it, say -- a 1000uF electrolytic will easily deliver 100A short circuit!).  This is a surprisingly challenging problem.  More often, a latching circuit is used (namely, an SCR -- see above examples), in which case the supply stays shorted until current is removed.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: castingflame

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Idiot proof Breadboard Over Voltage protection needed.
« Reply #8 on: February 15, 2019, 01:30:33 am »
Following circuit is more or less functional equivalent.

https://en.wikipedia.org/wiki/Crowbar_(circuit)


This crowbar circuit relies on latching property of TRIAC and you can't simply replace it with MOSFET or BJT.

Moreover, the TL431 is "N type" (it acts like an ideal NPN transistor), and so is an SCR.  Using a TRIAC (it's both N and P type in parallel) allows, I think in this case, quadrant 3 switching -- that is, negative MT current triggered by negative gate drive.  You don't need another transistor to invert the signal.

You can still add positive feedback to this circuit, incidentally -- this is important to ensure hard switching.  In the zener-triggered version of this circuit, a slowly rising supply may draw noticeable gate current without fully triggering the thyristor (which acts like a BJT at currents below the minimum holding current), therefore dissipating a lot of power in it.  This probably isn't a big deal in practice, in this circuit, because the TL431 has very high voltage gain already, but it can be improved just that little bit more.

To apply positive feedback, put a resistor in series with the TRIAC's MT2 terminal (the one tied with GND).  This generates a voltage proportional to MT2 current.  This also allows the TRIAC to saturate, so also connect a diode in parallel with it, limiting the resistor's voltage drop to ~1V.  Finally, connect the GND side of R2, to this resistor instead.  That way, when the TRIAC just begins to be triggered, the voltage sensed by the TL431 begins to rise, causing it to turn on much faster (<1us?) than it would otherwise.

The value of this resistor should be about, oh, 1/10th the value of R2, and also less than (0.25V) / (TRIAC maximum leakage).  Probably, 100-1k ohms would be a typical value.

A small capacitor can also be connected in parallel with R1, to give some dV/dt sensitivity (compensating for the response time of the TL431 and TRIAC), or in parallel with the TL431 (A-K) to drive the TRIAC directly from dV/dt.  Or the supply can be filtered (RC or LC+damping) to keep high frequencies out of the circuit to begin with; or preferably, a combination of these together. :)

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: castingflame

Online iMo

  • Super Contributor
  • ***
  • Posts: 5570
  • Country: va
Re: Idiot proof Breadboard Over Voltage protection needed.
« Reply #9 on: February 15, 2019, 11:19:58 am »
I've been using zeners in my solderless breadboards. Got a dozen of 3V6, 3V9, 5V6 ones and measured them such the Iz at 3.3V and 5V is a few mA max. It gives you some overvoltage and reverse polarity protection. No problems since then.
Readers discretion is advised..
 
The following users thanked this post: castingflame

Online Doctorandus_P

  • Super Contributor
  • ***
  • Posts: 4004
  • Country: nl
Re: Idiot proof Breadboard Over Voltage protection needed.
« Reply #10 on: February 15, 2019, 12:24:51 pm »
Soldering a Zener directly on the Blue Pill board, combined with a current limited power supply seems the simplest solutions (Apart from a decent cable / connector set).
 
The following users thanked this post: castingflame

Offline castingflameTopic starter

  • Regular Contributor
  • *
  • Posts: 100
  • Country: gb
Re: Idiot proof Breadboard Over Voltage protection needed.
« Reply #11 on: February 16, 2019, 09:55:28 am »
Thank you guys, you have given a great response.  :clap:

As a stop gap, I have just introduced a LD117(3v3) dedicated just for the MCU and another LD117(3v3) for everything else @3v3 and using a 12v supply (my fans are 12v).

I'll just get this little project finished first then look at resolving this, taking your ideas into account, as my next mini project.

Thanks again  :-+
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf