Over-voltage Protection

Thread Starter

leonhart88

Joined Feb 23, 2007
118
Hey guys,

So I have a small PCB mount DC-DC converter, which takes any DC input from 18-30V DC and converts it into 5V, with a max output of 300mA. It's a pretty neat device, but it has no over-voltage handling. I'm planning to add a resettable fuse for over-current protection, but I also would like the fuse to blow when there is over-voltage (from environmental factors, or from user errors of inputting >30V DC).

I've read quite a bit online about over-voltage protection and it seems the common thing to do is build a crowbar circuit.

I'm a little confused as to why I can't just put a fuse in series with the source, and then a zener diode to ground? For example, use a 250mA fuse and a 33V zener diode? Under normal voltages, the diode will act as a normal diode and block current flow to ground. When the voltage reaches 33V, the diode will act as a short circuit. There will be a voltage drop of 33V over a tiny resistance, which will probably blow the fuse immediately. I assume the problem would be finding a zener diode which can withstand this high current surge. I know that some zener diodes have very high surge current ratings though. Will it work with just a zener diode and fuse? I may be misunderstanding something and I hope someone can help clear it up.

Thanks.
 

#12

Joined Nov 30, 2010
18,224
A crowbar circuit is just a zener circuit with a helper. The zener is a bit wimpy about decision making. Adding an SCR changes it from things like, "maybe some current flows at 32 volts, and then some more at 32 and a half, and then the Zener has about 5 watts heating it up and the fuse still hasn't noticed anything" to "33 ZAP!"

All it requires is a resistor and an SCR. Pretty simple, really.
 

Thread Starter

leonhart88

Joined Feb 23, 2007
118
Well, if the zener has 5W heating it up, the fuse should have blown because the current that flows through the zener should be flowing through the fuse as well. Maybe i'll read into it more...

I'm also looking at TVS diodes (transient voltage suppression), as they apparently can withstand large currents better than zener diodes. They look very similar though.
 

#12

Joined Nov 30, 2010
18,224
Five watts into a 33 volt zener requires about 152 milliamps. That won't blow a 250 ma fuse.

From the way you're talking, it seems you need Watts Law. Power = current times voltage.
With that and Ohm's Law, you can calculate everything you need for this circuit.
 

Thread Starter

leonhart88

Joined Feb 23, 2007
118
Yeah, that means I would ideally wany a 8W or greater rated zener...which doesn't exist. I thought that the fuses would trip before the zener would be damaged, since they can withstand high currents for short periods of time. However it looks like a 250ma fuse trips in around 1 second.

I'm also not sure about the TVS diodes, as they have a clamping voltage drop that is a little high (if I want a breakdown at 33).

I think I will go with the crowbar route. I was just curious because theoretically a zener sounded like it would work alone.
 

#12

Joined Nov 30, 2010
18,224
TVS's are even wimpier than zeners about making decisions, but that's why this site exists...to clear up the details about parts and circuits.

1n5256B is a 30V 1/2 W zener. Remember that the gate voltage of the scr will be in series with the zener voltage so 30V zener will trip at about 30.5 to 31 volts. The resistor to ground should allow significant current (for the zener) before it gets to .6 volts.

about 8ma. 75 ohms
 

Thread Starter

leonhart88

Joined Feb 23, 2007
118
Thanks #12.

I'm a little confused with the resistor and what exactly it does. What do you mean by it should allow significant current before it gets to 0.6V? How did you calculate an appropriate resistor value?

From my understanding, when the zener voltage is reached, it breaks down and shorts, with a voltage drop equal to the zener voltage (lets say 30V). If the gate voltage for the SCR is 30V, it will conduct and short, causing a short circuit which will blow the fuse. So what exactly does the resistor do other than tie the gate to ground during normal operation?

I have attached the circuit I am looking at. I'm also curious as to why the capacitor is where it is and why it isn't directly connected to the input and ground (like normal filtering for input ripple).

Thanks for your help!
 

Attachments

Thread Starter

leonhart88

Joined Feb 23, 2007
118
Here is my current schematic.

I'm not sure what type of SCR I should use. They have a lot of parameters that are confusing. There are ones with lower/higher trigger voltages. My guess would be that a higher trigger voltage is better because it won't be triggered by noise?

Also, I think I was misunderstanding something in my last post. If the input voltage is 33V and the zener breaks down at 33V, the gate voltage would be 0V because the voltage drop across the zener would be 33V. I am a little stumped on understanding exactly how the gate triggering is supposed to work...
 

Attachments

Adjuster

Joined Dec 26, 2010
2,148
Thanks #12.

I'm a little confused with the resistor and what exactly it does. What do you mean by it should allow significant current before it gets to 0.6V? How did you calculate an appropriate resistor value?

From my understanding, when the zener voltage is reached, it breaks down and shorts, with a voltage drop equal to the zener voltage (lets say 30V). If the gate voltage for the SCR is 30V, it will conduct and short, causing a short circuit which will blow the fuse. So what exactly does the resistor do other than tie the gate to ground during normal operation?

I have attached the circuit I am looking at. I'm also curious as to why the capacitor is where it is and why it isn't directly connected to the input and ground (like normal filtering for input ripple).

Thanks for your help!
The breakdown characteristic of a practical Zener is not perfectly abrupt: there can be significant leakage current flowing at voltages considerably below the expected breakdown voltage. With no resistor added, the SCR could therefore trigger at too low a rail voltage, and the voltage might be very dependent on temperature. The resistor prevents small leakage currents from developing enough voltage to trigger the SCR.

Ideally, the resistor would be chosen to draw the current at which the Zener voltage is characterised, when the gate voltage reaches the right value for triggering. In practice, the resistor value can have some latitude, but should draw at least a few milliamps.
 

Thread Starter

leonhart88

Joined Feb 23, 2007
118
Thanks adjuster. Still a little confused..the zener should act like an open circuit under normal voltages...if there were no resistor wouldn't the reverse current just go to ground? Where would a voltage that can trigger the scr come from?

Also, the scrs I looked at on digikey have a max trigger voltage of 0.8 or 1.25 volts. If the input voltage is much larger than the zener breakdown, wont the gate voltage be a lot higher than this trigger value?

Thanks again.
 

Thread Starter

leonhart88

Joined Feb 23, 2007
118
I think you should place your fuse closer to the power supply, before any components such as your Schottky.
I was thinking of doing this as well. In the end, I don't think it really matters if I pick a schottky diode that can withstand larger voltages/currents than the fuse, since the fuse will just end up blowing.

Thanks.
 

Thread Starter

leonhart88

Joined Feb 23, 2007
118
Regardless of where the fuse is...I'm still stuck on how I'm supposed to calculate the resistor value and still not 100% sure I understand how the voltage at the SCR gate is "made". If the zener diode breaksdown and becomes a short, the gate voltage would just be (supply voltage - zener breakdown). What is confusing is that SCRs on Digikey state a max gate voltage of something like 1.5 volts.

I've also seen a crowbar circuit in a text book that had a resistor between R5 and C1 (in my previous picture).

I've also found an IC that is used for over/under voltage and reverse polarity protection. It looks like it could work. It's the LTC4365 from Linear Technologies. It uses MOSFETs. I'd still like to figure out the crowbar though, but if anyone has experience with an IC like the LTC4365, would be nice to hear.

Thanks.
 

Thread Starter

leonhart88

Joined Feb 23, 2007
118
I'm going to hit the sack. Tomorrow morning I'll Spice it for you so you can see a plot of it working.
Thanks CDRIVE. I found a good document by ON Semiconductor. Apparently the data sheets are misleading...they say max gate voltage/current, but it actually means that you need a minimum of 0.8V at the gate to cause conduction. That makes much more sense to me. Here's the PDF if anyone else is interested (I also attached it):

http://www.onsemi.com/pub_link/Collateral/HBD855-D.PDF

Now I just have to find out what is the maximum voltage that can be applied at the gate. That is, what is the max voltage that this crowbar circuit will work at before the zener or SCR dies. And also if I need to optimize the R and C values somehow, although it seems like the resistor is a pull-down for the SCR gate, and the capacitor is just to filter out noisy signals that may trigger the SCR.

I was also thinking about adding a capacitor before the crowbar to add more filtering...but I don't think I need it.

Thanks.

I'm glad to finally start understanding this...I've spent far too long already confused.
 

Attachments

CDRIVE

Joined Jul 1, 2008
2,219
I used a 30V Zener and an SCR with a gate trigger voltage of 3.55V. As you were told earlier, you can see the Zener begins to conduct below its rated zener voltage. Also note that I didn't use a fuse for this simulation. I found that my spice fuse model blew instantly, which real fuses don't really do. Instead I made a macro (U1) of a solid state breaker designed by our very own Ron H. Since you alluded to wanting a resettable breaker it may interest you.
 

Attachments

Ron H

Joined Apr 14, 2005
7,063
I used a 30V Zener and an SCR with a gate trigger voltage of 3.55V. As you were told earlier, you can see the Zener begins to conduct below its rated zener voltage. Also note that I didn't use a fuse for this simulation. I found that my spice fuse model blew instantly, which real fuses don't really do. Instead I made a macro (U1) of a solid state breaker designed by our very own Ron H. Since you alluded to wanting a resettable breaker it may interest you.
That's cool. You can, of course, reduce the breaker trip current by increasing the value of the sense resistor.
 

edmundopt

Joined May 4, 2011
60
You have a max of 300mA output current ?

if so, these components are way to much heavy dutty for that!

There are SCR's with a Holding Current(IH) of less than 10ma , you will need a resistor that allows that and a Zener with a Izt equal or greater..

example : TIC 106 (5amp SCR), IH of 8mA, 1volt triger
zener BZX85c33, with Izt of 8mA
resistor : R(max) = 30v/0.008 R(max) = 3750, near EIA e24 code = 3300ohm
capacitor : more bigger the more stable versus delay it gets


this is just a cheaper option, do you agree CDRIVE ?
 
Top