ebay buck boost controllers. CC CV fail safe??

[user1235423643]

hello, just a quick question about ebay buck boost controllers.

i have had my fair share of blowing them up one way or another. a short on the output will do it..

PROBLEM is, is you set it to 5v and have a 12v input when it blows it will feed 12v to the output..

any ideas how to protect the load on the output in a basic way

thanks

[jeroen79]

A crowbar (thyristor and fuse) on the input side triggered by an overvoltage comparator on the output side.

[Jeroen3]

Most basic way is a big 1500 W TVS diode (SMCJ5.0A) and a fuse (or PTC).

[uncle_bob]

Hi

The old school approach is to put a monster SCR on the beast and fire that when the output goes above 5.5V. Hopefully it is rated for enough of a surge to take out the up stream fuse. Circuits range from a dead simple two resistors on the SCR gate to far more elaborate approaches. They still create a mess. Since it's a race between the fuse and the surge current rating on the SCR, you more or less are finding out which supplier was lying to you about their part.

Bob

[user1235423643]

the output (UBS )5v needs to start around 5v while the current will be anywhere from 100ma to 2.5a

thanks

[uncle_bob]

the output (UBS )5v needs to start around 5v while the current will be anywhere from 100ma to 2.5a

thanks

Hi

It's the "interesting" current at startup on some of these low cost switchers (not much of a soft start) that drives you to a pretty big fuse on the input. That fuse might be 4X larger than you would first guess. Your OVP scheme needs to pop that fuse ... thus it can get pretty big. Yes, there are a lot of variables and they all get into this. Since you don't have much design control over an eBay regulator board, you live with what you get.

Bob

[user1235423643]

this is what we are talking about
i dont think fuses will work to protect voltage spike in a blow condition.

http://www.ebay.com.au/itm/CC-CV-Voltage-Regulator-Automatic-Boost-Buck-Converter-4-35V-to-1-25-25V-LM2577-/141658672421?hash=item20fb83cd25:g:LxcAAOSwv0tU8XPp

[uncle_bob]

this is what we are talking about
i dont think fuses will work to protect voltage spike in a blow condition.

http://www.ebay.com.au/itm/CC-CV-Voltage-Regulator-Automatic-Boost-Buck-Converter-4-35V-to-1-25-25V-LM2577-/141658672421?hash=item20fb83cd25:g:LxcAAOSwv0tU8XPp

Hi

A brute force approach will indeed protect the load with that sort of board. Consider a 20A fuse on the input and a > 20A SCR on the output (parallel with the load). The SCR fires when the output hits (say) 5.5V. While it's on, the output is pulled to near zero volts. The board may catch fire. The fuse may or may not blow. The output will go to zero ....

Bob

[user1235423643]

not sure i understand? 20a fuses?

[macboy]

this is what we are talking about
i dont think fuses will work to protect voltage spike in a blow condition.

http://www.ebay.com.au/itm/CC-CV-Voltage-Regulator-Automatic-Boost-Buck-Converter-4-35V-to-1-25-25V-LM2577-/141658672421?hash=item20fb83cd25:g:LxcAAOSwv0tU8XPp

Hi

A brute force approach will indeed protect the load with that sort of board. Consider a 20A fuse on the input and a > 20A SCR on the output (parallel with the load). The SCR fires when the output hits (say) 5.5V. While it's on, the output is pulled to near zero volts. The board may catch fire. The fuse may or may not blow. The output will go to zero ....

Bob

20A ?!?! With a 12 V input, that's 240 W input capability. With 5 V on the output, and a maximum current of 2.0 A, that's 10 W of output power. Re-examine your fuse choice.

[uncle_bob]

not sure i understand? 20a fuses?

Hi

Take a look at how these beasts pull current at startup. It's not at all pretty.

Bob

[Dave]

No. No no no.

You don't pick the ampacity of a fuse for the turn-on surges. You pick a fuse that is slow enough not to burst immediately on the inrush surge (dive into the datasheet and look at the I²t ratings). 20A fuses would be absolutely insane for a circuit that draws 1-2A when operational.

[Zero999]

No. No no no.

You don't pick the ampacity of a fuse for the turn-on surges. You pick a fuse that is slow enough not to burst immediately on the inrush surge (dive into the datasheet and look at the I²t ratings). 20A fuses would be absolutely insane for a circuit that draws 1-2A when operational.

Indeed. I'd recommend a PTC resistor because it's resetable and takes a long time to blow.

[uncle_bob]

No. No no no.

You don't pick the ampacity of a fuse for the turn-on surges. You pick a fuse that is slow enough not to burst immediately on the inrush surge (dive into the datasheet and look at the I²t ratings). 20A fuses would be absolutely insane for a circuit that draws 1-2A when operational.

Hi

Put a slow blow fuse in series with a crowbar limiter and see what happens sometime ... it's not pretty.

Bob

[user1235423643]

ok, so im still unsure why we are talking about fuses.

i want to protect the output from going to input voltage in the chance of failure

the output is USB 5v  anywhere between 100mah to 3a amp

what i want to protect is VOLTAGE spike over 5v if the reg fails

[uncle_bob]

ok, so im still unsure why we are talking about fuses.

i want to protect the output from going to input voltage in the chance of failure

the output is USB 5v  anywhere between 100mah to 3a amp

what i want to protect is VOLTAGE spike over 5v if the reg fails

Hi

The way you get a spike is when the regulator fails.

When the regulator fails, the only solution is to clamp the output to ground.

When you clamp the output to ground. You need a fuse to separate the input from the power source.

Bob

[FlyingHacker]

ok, so im still unsure why we are talking about fuses.

i want to protect the output from going to input voltage in the chance of failure

the output is USB 5v  anywhere between 100mah to 3a amp

what i want to protect is VOLTAGE spike over 5v if the reg fails

The way you get a spike is when the regulator fails.

When the regulator fails, the only solution is to clamp the output to ground.

When you clamp the output to ground. You need a fuse to separate the input from the power source.

Without the fuse whatever you are using to clamp the output to ground (e.g. a Zener) will burn out rapidly, and become open, thus no longer clamping anything to ground.

[ez24]

Most basic way is a big 1500 W TVS diode (SMCJ5.0A) and a fuse (or PTC).

I do not see one at the OP desired output voltage  ??  but I could be not seeing right

(hope the 1500 w filter gets through this link)

http://www.mouser.com/Circuit-Protection/TVS-Diodes/_/N-axfz1?P=1yw7wqm&gclid=CjwKEAjwpLa5BRCTwcXS6_rpvC4SJACTDQMMFKek5vE1DEn883Roz71pYX0ulss3uIlWX-JEFYs-MBoCft3w_wcB

[uncle_bob]

ok, so im still unsure why we are talking about fuses.

i want to protect the output from going to input voltage in the chance of failure

the output is USB 5v  anywhere between 100mah to 3a amp

what i want to protect is VOLTAGE spike over 5v if the reg fails

The way you get a spike is when the regulator fails.

When the regulator fails, the only solution is to clamp the output to ground.

When you clamp the output to ground. You need a fuse to separate the input from the power source.

Without the fuse whatever you are using to clamp the output to ground (e.g. a Zener) will burn out rapidly, and become open, thus no longer clamping anything to ground.

Hi

.... and ...

If you want the regulator to run fine at 5.25V, but kill the spike completely at 5.5V, there is no zener that will do that. Either you use one that is to low in voltage (and pulls current all the time) or one that is to high voltage (and starts at 5.6V, but does not fully clamp 6.0V).

Bob

[user1235423643]

so can someone confirm the best way and simple to keep the voltage at 5v in the case the reg blows or fails

http://www.ebay.com.au/itm/CC-CV-Voltage-Regulator-Automatic-Boost-Buck-Converter-4-35V-to-1-25-25V-LM2577-/141658672421?hash=item20fb83cd25:g:LxcAAOSwv0tU8XPp

[FlyingHacker]

This PDF is the manual, with schematic, for the old Lambda Overvoltage Protectors from Lambda power supplies. Lambda is one of the best names in power supplies, producing supplies for the likes of Agilent/Keysight/HP and others.

This schematic uses some older transistors, but the idea is the same. Of course it still needs a fuse upstream...

https://www.surplussales.com/Semiconductors/pdf/ps-lmov-2.pdf

Edit: file was too big, link included.

[ez24]

Since the OP keeps saying "simple"  how about a linear regulator at the end of the buck?

Seems that the buck can be set to give 5v at the output of the LDO and then it would stay there if the buck goes to 12v  ??

I think there are LDOs designed to reduce ripple from bucks so it would be a win win (I think)

[uncle_bob]

so can someone confirm the best way and simple to keep the voltage at 5v in the case the reg blows or fails

http://www.ebay.com.au/itm/CC-CV-Voltage-Regulator-Automatic-Boost-Buck-Converter-4-35V-to-1-25-25V-LM2577-/141658672421?hash=item20fb83cd25:g:LxcAAOSwv0tU8XPp

Hi

SCR + 2 resistors + fuse. The fuse is not optional

Bob

[user1235423643]

so far the most simple idea is to use a 5v reg. after the buck. remember its a $3 unit

if the buck goes high and feeds 12v in to the 5v reg it will get hot and cook but it should then break the circuit before letting more then 5v in.

??

[ez24]

All a solid state fuse is, is a high-side transistor (fet or bjt) with a few control components nothing major. I think I've seen intelligent load switches that have everything pretty well integrated.

I searched Aliexpress to see how much a solid state fuse costs and could not find any.  Then I did a Google search and came up with this name "Resettable fuse".  Aliexpress sells 10,000s of these for about 25 cents apiece.  Cheap and simple.

But they are current based, right ?  The OP does not want 12v, he says nothing about current.  I think we are back to a LDO but that is not 100% guaranteed, because if it is "simply" hooked up, there is a chance it could fail and pass through the 12v (per a Google search).

You can increase the reliability with lots of external components as seen in the LDO datasheets.
But then it is not "simple" anymore.

My guess there is NO "simple answer" to the OP's question with great certainty.

If I were to do this, I think I would determine the current under normal conditions and insert a resettable fuse.   I believe this is "simpler" than a LDO.    You just need a set of fuses to use with different circuits.   

How about use a buck -> LDO -> resettable fuse   (seems pretty safe ?)
and the LDO should reduce some ripple.   Does this meets the OP's requirement?