250mA DC fuse with a wide temperature range

Go To Last Post
28 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi all,

I need to measure DC current in the range of +/- 250mA and I need some kind of short cut protection.

I use an 1 Ohm shunt and measure the voltage drop with the ADC of an ATtiny24. I have 4 IOs left for free useage.

Can anybody suggest a short cut protection which triggers somewhere in the range of 250 to 500mA and operates in the temperature range of -40°C to +125°C.

It seems like standard fuses won't work here.

Thanks for any idea!

Regards
Sebastian

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

One or two diodes connected parallel to shunt resistor.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The current source is a 12V Pb accu. The diodes wouldn't withstand a short cut. In the case of a short cut, the circuit must be cut off.

Regards
Sebastian

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
It seems like standard fuses won't work here.

Combining a fuse with the shunt protection diodes should be a good solution. You should have a fuse in the circuit anyway. Lead acid batteries can supply a lot of "magic smoke" very quickly.

Tom Pappano
Tulsa, Oklahoma

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

This is my best idea so far...

My plan is to monitor the current with the ADC. If the current is too high, I make the FETs high impedant by toggeling the gate voltage.

Usually it should be enough to only use the N-FET but if the user has accidentily exchanged current+ and current- the intrinsic N-FET diode will be forward biased and the smoke will arise (So unfortunately the LT4356H won't work).

Thus I have added the P-FET as reverse battery protection.

Now the current can only flow from current+ to current-. First I wanted to be able to measure positive current as well as negative current. But that is only a nice to have feature.

Hmmm, I could use the analog comparator as well to detect a short cut.

What do you think about this circuit? Maybe you have an even better idea.

Thanks for all your replys!

Regards
Sebastian

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The response of your circuit will be far to slow. I suggest the circuit Tom wrote: a diode (or 2) over the shunt, and a fuse. Diodes can take quite some beating as long as the duration is short.
Another important safety-issue: add resistors in the lines from shunt to diff input of the AVR. 1k will do.

Nard

A GIF is worth a thousend words   They are called Rosa, Sylvia, Tricia, and Ulyana. You can find them https://www.linuxmint.com/

Dragon broken ? http://aplomb.nl/TechStuff/Dragon/Dragon.html for how-to-fix tips

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I've tested the circuit a little bit.
One advantage of this circuit is that the 1 Ohm shunt also works as degeneration of the N-FET. So the short cut current is limited to approximately 2.2A (ZXM64N035L3). When using the analog comparator of the ATtiny24, I'm able to switch off the FETs within a few microseconds. So I think the circuit will be save. Isn't it?

Nard, I was just too lazy to plot some resistors into the schematic. Yes, I was planning to add resistors to the ADC input and I will add some caps to the ADC inputs as well.

Thanks for the feedback!

Regards
Sebastian

EDIT: The problem with the standard fuse is that it is only certified to operate at temperatures up to +85°C. But my application will be placed inside a climate chamber which operates at temperatures up to +125°C. The idea is to place one modul inside the climate chamber which monitors several parameter, to reduce the number of cables which must come out of the climate chamber. The climate chamber is very big, so reducing the number of long cables gives some good advantages. Actually we're not going over +85°C. But all components on the board are certified to work up to +125°C proberly. I didn't want to limit the max temperature of the board by using a fuse. Maybe someday we have to go higher than +85°C...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

IMO it's important that protectioncircuits (especially when a Pb-battery is the pusher ;) ) do their job under any circumstance, because errors never come on their own: usually they are accompanied by several co-errors.
It's a way of thinking .... this kind of safetystuff.
F.i.:
- during power-on reset, the AVR has no control ... yet
- if some dirty EMI occurs (is it a car-circuit?), there is no guarantee that the AVR will be able to act: it could be upset itself.

You can still use your circuit, but then to prevent a fuse-blow .... if it can be taken care of by the AVR, that saves a fuse-change.

Nard

edit: I didn't read your edit when posting ... I will give it some more thoughts ..

A GIF is worth a thousend words   They are called Rosa, Sylvia, Tricia, and Ulyana. You can find them https://www.linuxmint.com/

Dragon broken ? http://aplomb.nl/TechStuff/Dragon/Dragon.html for how-to-fix tips

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think I will add pullup/down resitors to the FET gates. So during reset the current path will be open. We will use the board only for climate tests, not for EMC tests and the car motor will never run during the tests. The danger of a short cut is mainly when connecting my board to the ECU. The worst case is that my board blows up while the car cables are still protected by the car fuses (which are only available from several ampere up).

I'm looking forward your reply!

Regards
Sebastian

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Here my approach: discrete logic handles the overcurrent situation; status can be read back by AVR. AVR can reset the circuit by pulsing the clock-line. If there is still overcurrent-situation, the (high active) reset will take action and overrule the AVR-pulse.

(note: sketch is not complete: f.i. you may need a proper gate-driver for the fet, decoupling, ensuring that 5V is available ALL THE TIME, a proper power-on reset for the FF, etc)

Nard

Attachment(s): 

A GIF is worth a thousend words   They are called Rosa, Sylvia, Tricia, and Ulyana. You can find them https://www.linuxmint.com/

Dragon broken ? http://aplomb.nl/TechStuff/Dragon/Dragon.html for how-to-fix tips

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Wow, thank you!!!

Hmmm, the flip flop will trigger if the shunt voltage is higher than approximately 3V, right? In this case a current of 3A is neccessary to trigger the fuse. BUT as I wrote above (2:12 PM) the circuit itself limits the max. current to approximately 2.2A. So the fuse will never trigger. And a continious current of 2.2A means 5W power dissapation for the 1W shunt resistor and 21W for the FET - I guess I will see smoke...

However I could use a real analog comparator and set the compare voltage to 300mV (= 0.3A). Maybe you can give me an advice.

Once again thank you!

Regards
Sebastian

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

HCT-logic : a "1" @ 2V or more.

The powerdissipation during the short is no problem IMO as it's duration is short .... very short. My suggestion for the fet: FDP6035 or alike. FDS6612A if you want a SMD. 8A continuous, so it's a bit small.

Addition of a discrete comparator is also feasible. But a simple NPN-transistor will be fine too. Of course the reset-input of the FF must be low active.

Rev1_1 attached

Nard

Edit/added:
If you really need that 300mV treshold: add a pull-up resistor of appr. 15k to +5V on the base of the NPN-transistor.

Edit2:
the FDS6612A will work fine as the overcurrent will never reach that value. When there is 5A flowing, the source of the fet will sit at 5V :) and therefor will turn off. Intrinsic safety-feature :)

Edit3: oops, that 100u is a bad choice. Hold on ...
A TL7705 (or a more modern version of it) will be a better choice for the reset

Attachment(s): 

A GIF is worth a thousend words   They are called Rosa, Sylvia, Tricia, and Ulyana. You can find them https://www.linuxmint.com/

Dragon broken ? http://aplomb.nl/TechStuff/Dragon/Dragon.html for how-to-fix tips

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Noticed also a typo in the fet partnumber ...

So rev1_2 it is :)
.. and edited

Attachment(s): 

A GIF is worth a thousend words   They are called Rosa, Sylvia, Tricia, and Ulyana. You can find them https://www.linuxmint.com/

Dragon broken ? http://aplomb.nl/TechStuff/Dragon/Dragon.html for how-to-fix tips

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Looks good, but D1 is actually redundant. The collector of the transistor will provide the blocking in this OR gate arrangement.

Tom Pappano
Tulsa, Oklahoma

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You're right Tom.

Attached another alternative: no integrated circuits, but all discrete. Small, yet effective.
Comments are welcome :)

Nard

Edit: there is no need for a schottky for D2: a minimelf 1N4148 is OK. And use the ADC to read back the pin: it's not logic level.

Attachment(s): 

A GIF is worth a thousend words   They are called Rosa, Sylvia, Tricia, and Ulyana. You can find them https://www.linuxmint.com/

Dragon broken ? http://aplomb.nl/TechStuff/Dragon/Dragon.html for how-to-fix tips

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Nard, your last post is amazing!!! I'll go this way. R1 and R3 will cause an almost constant current offset of 300µA., right? So I can simply sub the offset from my result. Maybe I should decrease R1 to 820 Ohm. If the temperature is high VBE will drop and the circuit might trigger on a current less than 250mA.

You have put a lot off good ideas and effort in your solution. Thank you!!!!!

Regards
Sebastian

EDIT: Why do you think the output at Q3 isn't digital. Doesn't D2 make the output digital?

Last Edited: Sat. May 23, 2009 - 05:33 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The standrad fuses are chertified to 85 C. But the worst thing you can expect from the fuse at higher temperatures is to blow at a lower current. So you may have to use a 1 or 2 A fuse. Diodes parallel to the shunt could give you some extra current the shunt can withstand.

There is also a way to build an all elektronik fuse from a MOSFET with limited voltage at the gate. Reverse voltage protection is probably best done by a simple diode. But I would still like to have the normal fuse as a second safty.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You're welcome :)

Sebastian wrote:

Quote:
Maybe I should decrease R1 to 820 Ohm
Or increase R3 to 18k or more. Note that it will only affect the trip-level.

Quote:
R1 and R3 will cause an almost constant current offset of 300µA., right?
Almost. R3 is doing the actual current-source-job. But you will hardly notice the difference. It's 1 pro mille or so.

As a sidenote: In general I always use a calibration formula on ADC results: value = gain*counts + offset

Quote:
Why do you think the output at Q3 isn't digital. Doesn't D2 make the output digital?
Nope. The highest voltage on the collector of Q3 will be appr. 1.4V: D2 and Q2be. So if you have an ADC-pin spare, use that one. Recommended treshold: 800mV

Kleinstein wrote:

Quote:
But I would still like to have the normal fuse as a second safty.
I second that.

Bard

A GIF is worth a thousend words   They are called Rosa, Sylvia, Tricia, and Ulyana. You can find them https://www.linuxmint.com/

Dragon broken ? http://aplomb.nl/TechStuff/Dragon/Dragon.html for how-to-fix tips

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Could also add a properly sized capacitor with the pullup to Q3, to limit the turn on speed after a fault.

EDIT: typo...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kleinstein wrote:
The standrad fuses are chertified to 85 C. But the worst thing you can expect from the fuse at higher temperatures is to blow at a lower current. So you may have to use a 1 or 2 A fuse.
Diodes parallel to the shunt could give you some extra current the shunt can withstand.
Yes, the fuse will blow earlier, but the datasheet says nothing about how much. However, protecting the shunt via diodes might work well @25°C but at high temperatures I would expect a high error current flowing through the diodes. (Correct me if I'm wrong)

Kleinstein wrote:
There is also a way to build an all elektronik fuse from a MOSFET with limited voltage at the gate. Reverse voltage protection is probably best done by a simple diode. But I would still like to have the normal fuse as a second safty.
As I noted above, the N-FET with the 1 Ohm shunt connected to the source works as current source of approximately 2.2A. So using an 1-2A fuse as you suggested doesn't really give extra protection. The fuse might withstand this current several minutes and in this time the FET is already dead (20W power dissapation). And there are still the car fuses which will trigger at several ampere. So the only dying part would be my board in the worst case.

Plons wrote:
Nope. The highest voltage on the collector of Q3 will be appr. 1.4V: D2 and Q2be. So if you have an ADC-pin spare, use that one. Recommended treshold: 800mV
Ah, now I see. Thank you!

Plons wrote:
As a sidenote: In general I always use a calibration formula on ADC results: value = gain*counts + offset
Yes, and in addition I will oversample the signal to gain even more accuracy.

Regards
Sebastian

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I understand your reluntancy on the recommended fuse. But try to see it as an extra protection in case everything else fails, .... then it makes sense. Car-fuses are there to prevent the wiring from going up in flames, .... not to protect the circuits attached.
If I was in your position Sebastian, I'd put a 1A FastBlow fuse in series, heat it to 125 degr C, and test at what current it blows. It will most likely withstand 500mA easily. And under normal circumstances the electronics will take care of opening up the current loop before the fuse has even *considered* to blow :lol:

Cheers

Nard

A GIF is worth a thousend words   They are called Rosa, Sylvia, Tricia, and Ulyana. You can find them https://www.linuxmint.com/

Dragon broken ? http://aplomb.nl/TechStuff/Dragon/Dragon.html for how-to-fix tips

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Nard, you're right. Too much theorie for a tinker like me. I took an 1A fuse and a pocket lighter, heated the fuse up while pumping 500mA through the fuse. The fuse survived just fine and I guess it was even hotter than 125°C.

On a side affect, I just noted that I can't use VCC as ADC reference when using differential inputs. So I have to switch to the ATtiny261 which provides an internal 2.56V reference. So I can even use an 0.5 Ohm shunt for the current measurement which heats up less durig a short cut.

Regards
Sebastian

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

S-Sohn wrote:
Can anybody suggest a short cut protection which triggers somewhere in the range of 250 to 500mA and operates in the temperature range of -40°C to +125°C.

It seems like standard fuses won't work here.


You can get fuses with a specified operating temperature range of –55ºC to +125ºC.

Some Littelfuse examples of –55ºC to +125ºC fuses:

251 Series - PICO® II Very Fast-Acting Subminiature Axial leaded fuse: http://www.littelfuse.com/series...

262 Series - MICROâ„¢ High-Reliability Very Fast-Acting Subminiature Plug-In Fuse: http://www.littelfuse.com/series...

265 Series - PICO® High-Reliability Very Fast-Acting Subminiature Axial Leaded Fuse: http://www.littelfuse.com/series...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thank you AndersAnd.

Regards
Sebastian

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

S-Sohn wrote:
This is my best idea so far...

My plan is to monitor the current with the ADC. If the current is too high, I make the FETs high impedant by toggeling the gate voltage.

Usually it should be enough to only use the N-FET but if the user has accidentily exchanged current+ and current- the intrinsic N-FET diode will be forward biased and the smoke will arise (So unfortunately the LT4356H won't work).

If you had read (or even glanced :p) at the datasheet you would have seen this:

Quote:
Reverse Input Protection to –60V

And then further down:

Quote:
Two MOSFETs can be connected back to back to replace an inline Schottky diode for reverse input protection. This improves the efficiency and increases the available supply voltage level to the load circuitry during cold crank.

Admittedly, being a linear device it is expensive. And only really advantageous if you need the resettable capability + over-voltage protection.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

thmjpr, to be honest: I really didn't looked at the part very closely. The main reason is that it isn't available at any of the standard distributors of our company. To get parts from our standard distributors is fairly easy but ordering from other distributors is really a pain. I'd had to discuss with a lot of people and collect several signatures until I (maybe) could order that chip. That's mainly because the company I'm working for is going through a hard time (like almost any other automotive company) and we have to save money by hook or by crook.

Maybe I could get some samples right now. But later on, when we need some more devices I might get problems to get the parts.

When I got the link from you, I first checked if the part is available for me. It wasn't. So secondly I took a (really) short glance at the device, saw the N-FET and thought: OK, it wouldn't fit to the requirements anyway.

Sorry, that I didn't check your advice more carefully. I really appreciate your help.

Regards
Sebastian

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

No problem, I completely understand.
May I ask what your main distributor is over in germany?