Combined Reverse & Overvoltage Protection

I'm looking for help with a combined reverse and overvoltage protection circuit.

The IC I'm protecting has an absolute maximum input range of -0.3 to 26V. Typical Vcc is 12V. I'd like to cap Vcc at 18V and prevent any reverse voltage. In both cases below, circuit activation will blow a fuse/CB in the supply line.

-----

I'm familiar with the standard SCR/zener-based crowbar circuit for overvoltage, but I'm wondering if it's possible to make it bidirectional by substituting a TRIAC for the SCR [see image below: TRIAC.png].

It appears that the TRAIC gate will have to see ~1.3V before turning on, so with reverse voltage applied, my IC's Vr limit will be exceeded by 1V.

-----

Another alternative I'm considering is the standard SCR crowbar with a parallel Schottky diode in reverse bias [see image below: SCR-Schottky.png].

Many 10A-rated Schottkys [like this] seem to have a Vf of ~500mV, which also exceeds my IC's reverse limit, although not as much.

-----

Questions:

1. Will these two circuits work as I've drawn them?
2. Am I correct in my assessment that both solutions exceed my IC's Vr limit?
3. If so, are there other options that will protect better?
4. Is C1 required with the TRIAC?
5. How do I calculate proper values for R1 and C1 in these circuits?
6. Is it correct to say that overvoltage cutoff = Vtrigger + Vzener?
7. Are 8-10A rated SCR/TRIAC sufficiently robust to blow a 3A fuse/CB without failing?

Thanks.

How about a plain old transil? For example http://uk.mouser.com/ProductDetail/STMicroelectronics/SM15T18AY/?qs=sGAEpiMZZMuNo3spt1BaV5%2fT68rC1k52F0HZ1pK4mlA%3d
You may want to use an anti-parallel schottky, but really that depends on the fault current when you reverse the polarity and the rating of the fuse. This one is unidrectional, so in reverse it acts as a normal PN diode.

Another idea for rerse polarity protection is either a mosfet or a diode in series, depending on how much current the device draws.
edit: like this http://hackadaycom.files.wordpress.com/2011/12/p-fet-reverse-voltage-protection1.png?w=470&h=326

You can use the triac for both directions, it will fire when the gate is the same polarity as MT2, so reverse the supply and reverse the zener .

kubeek said:

How about a plain old transil? For example http://uk.mouser.com/ProductDetail/STMicroelectronics/SM15T18AY/?qs=sGAEpiMZZMuNo3spt1BaV5%2fT68rC1k52F0HZ1pK4mlA%3d
You may want to use an anti-parallel schottky, but really that depends on the fault current when you reverse the polarity and the rating of the fuse. This one is unidrectional, so in reverse it acts as a normal PN diode.

Another idea for rerse polarity protection is either a mosfet or a diode in series, depending on how much current the device draws.
edit: like this http://hackadaycom.files.wordpress.com/2011/12/p-fet-reverse-voltage-protection1.png?w=470&h=326

Click to expand...

Thanks kubeek. I've looked at TVS devices, but I'm a bit dubious of their ability to carry a dead short long enough to blow a fuse or open a breaker without being damaged. The data sheets all show impressive numbers, but they're for transients of at most a few milliseconds.

The system will be powered by a 12V SLA or Li battery, so fault currents could be substantial.

Could you explain "anti-parallel"? I'm unfamiliar with that term.

Thanks for the hackaday image. I had seen that blog post and lost track of where it was. That P-FET solution is an elegant one. Perhaps that, combined with the standard SCR crowbar...

Dodgydave said:

You can use the triac for both directions, it will fire when the gate is the same polarity as MT2, so reverse the supply and reverse the zener .

Click to expand...

Do you mean that the Zener has to be reversed to protect against reverse polarity (i.e. it should "point down"), or that I need a second Zener head-to-head or tail-to-tail with the first? Any chance you could post a quick diagram of your suggestion?

Thanks, Dave.

From what I´ve read TVSs fail shorted, so there should be no problem with overcoming their ratings. Anyway it wouldn´t cost much to buy some and test them straigth with two SLAs in series to get 24V fault voltage and see what happens on the supply rail.
Antiparallel means parallel but the devices are pointing in opposite directions.

I have 2 suggestions. First, put a diode in series with the power supply so that if it is connected backwards, no current flows. Second, don't connect the power supply backwards. A polarized connector might solve this.

The problem with making a circuit idiot proof is that you can never figure out how many ways an idiot can do it wrong.

apqo1 said:

Do you mean that the Zener has to be reversed to protect against reverse polarity (i.e. it should "point down"), or that I need a second Zener head-to-head or tail-to-tail with the first? Any chance you could post a quick diagram of your suggestion?

Thanks, Dave.

Eric

Click to expand...

like this, reverse the zener at the gate terminal

I think #12's suggestion to use a diode in series with the power to prevent reverse circuit voltage is good, provided you can tolerate the forward voltage drop of the diode in your supply voltage. A large Schottky diode will have a forward drop of about 0.5V.

You still need an over-voltage circuit with that.

OK, kubeek solved the reverse voltage protection with the hackaday link. Here's the full post with explanatory video.

I did some searching for a solution on overvoltage protection and came up with this.

With these two solutions combined, I have the attached circuit, which protects against reverse voltage and overvoltage using transistor switching, and without blowing a CB or fuse. BOM as follows:
Any comments or suggestions are welcome. Thanks to all who offered assistance!

I'll try to remember to post here again once I have a board built and tested.


EDIT: This circuit may work in low/constant current applications, but the over-voltage protection portion oscillates badly when feeding a switching regulator. See this post for a solution that tolerates varying current demand without oscillation. Q1, D1 and R1 work fine as-is for reverse voltage protection.

I put the circuit in my last post together tonight and it works great:

• Reverse voltage (tested to -24V) is blocked.
• Overvoltage cuts to zero at 18.55V.

The overvoltage cutoff is slightly higher than the specs of Zener diode D2 would suggest (18.36V max), but I don't have top-tier test equipment so it's definitely in the ballpark.

Thanks again to all who offered ideas on this thread. Your help is much appreciated!

apqo1 said:

I put the circuit in my last post together tonight and it works great:

• Reverse voltage (tested to -24V) is blocked.
• Overvoltage cuts to zero at 18.55V.

The overvoltage cutoff is slightly higher than the specs of Zener diode D2 would suggest (18.36V max), but I don't have top-tier test equipment so it's definitely in the ballpark.

Thanks again to all who offered ideas on this thread. Your help is much appreciated!

Eric

Click to expand...

Looks really good and who made the board ???

Also what's the specs for it ???
Like
min voltage
max voltage

Dr.killjoy said:

Also what's the specs for it ???
Like
min voltage
max voltage

Click to expand...

+1
I'd be curious to see a simulation (or real data) of transient responses and such. Looks wicked clever.

Dr.killjoy said:

Looks really good and who made the board ???

Click to expand...

I have prototype boards up to 5 sq in (which is most of the boards I do) made at OSH Park. This board (slightly more than 0.5" square) cost $1.70 for three copies, delivered via First Class Mail. At $5/sq in, and with 12-day turnaround, I often have boards made for prototyping.

Above that size I was using BatchPCB, but just this morning I discovered that they've closed their doors and are sending all customers to OSH Park.

Dr.killjoy said:

Also what's the specs for it ???
Like
min voltage
max voltage

Click to expand...

I tested the reverse voltage function down to the lower limit of my bench supply (millivolts) and detected no pass through. However, my multimeter was wandering a lot and it's hard to tell what's measurement and what's noise; I presume there was nothing to measure since the meter never settled.

I'm not sure of the upper voltage limits, as there is other protection for really gross transients elsewhere in the system this circuit will serve. Take a look at the transistor and Zener data sheets (BOM on pg 1 of this thread). The circuit could easily be scaled up to handle whatever you need by substituting devices that match your specs. According to the Maxim write-up on this circuit, Q3 isn't critical as it only switches Q2.

wayneh said:

+1
I'd be curious to see a simulation (or real data) of transient responses and such. Looks wicked clever.

Click to expand...

Thanks. I wish I could take credit for it, but I just aped circuits I found elsewhere.

I'd love to post this data but I'm very much an amateur hobbyist and I don't own a scope. Sorry!