Author Topic: Fuse selection (blow times)  (Read 2893 times)

0 Members and 1 Guest are viewing this topic.

Offline SimonTopic starter

  • Global Moderator
  • *****
  • Posts: 18118
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Fuse selection (blow times)
« Reply #25 on: August 23, 2019, 08:16:15 pm »
For a 1206 package, I'm skeptical it's practical as the contact surface area isn't there for steady 20-40A. You need heavy copper PCB but a tiny 1206 pad is just too small for long term high currents and over 1W dissipated.



I'm thinking 210µm, it's only a dollar more than anything else.
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 8175
  • Country: us
Re: Fuse selection (blow times)
« Reply #26 on: August 23, 2019, 08:52:29 pm »
The motor drive looks after itself. If you put your hand on the hub even at low speed (you don't do that at 4'700rpm) it will cut out. The only thing i have had so far which is the reason for rethinking fuse strategy is that i had fans running with one fan unplugged for some testing. that unplugged fan was being driven backwards. When i plugeed it in it blew up but did not take the fuse out. But the last fuse was PCB track. i think i will bow my head to the fuse makers and admit they know more than me.

Would overcurrent protection have prevented the fan from blowing up?  If so, in this event, whether an intermittent connection or an unplugging during service, which would be preferable--a blown up fan or a irreparably damaged power distribution board?

If, as you say elsewhere, the wiring is adequately protected otherwise, and you don't really want to protect anything or do anything except meet your customer's demand, I suppose the answer is the least likely fuse to ever blow, in this case the slo-blo model.
« Last Edit: August 23, 2019, 09:49:15 pm by bdunham7 »
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Offline SimonTopic starter

  • Global Moderator
  • *****
  • Posts: 18118
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Fuse selection (blow times)
« Reply #27 on: August 23, 2019, 09:52:54 pm »
why I did was not representative of real use and could never happen. If the fan was damaged it would have been blown mosfets, not sure any fuse can deal with that. My issue with too slow a blow is that the epoxy starts to melt....
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 8175
  • Country: us
Re: Fuse selection (blow times)
« Reply #28 on: August 24, 2019, 04:26:02 pm »
why I did was not representative of real use and could never happen. If the fan was damaged it would have been blown mosfets, not sure any fuse can deal with that. My issue with too slow a blow is that the epoxy starts to melt....

That's good though--it shows that it didn't give up without trying!  :)
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 7678
  • Country: ca
Re: Fuse selection (blow times)
« Reply #29 on: August 24, 2019, 08:11:06 pm »
As i already said it is to satisfy compliance and the loads are unlikely to blow a fuse but the customer wants one. So a fuse they get that will never be needed but it's there.

No, "the customer is always right" does not 100% apply to engineering. The customer does not know the safety standards, or have experience doing product development. Blindly following what they want can be a noose around the engineer's neck. You have to explain and negotiate the product's requirements with a customer, sales/marketing, upper management etc. Sometimes you have to firmly say "no" to their demands because otherwise you end up making something unsafe or unreliable.

If you use non-replaceable fuses, your product is only as good as the fans.
A lemon fan- mosfets that short, bearings failing, water getting in, plastic blades cracking and flying off- mean overcurrent and your fuse pops. A guy puts a new module in and it also pops a fuse. Hmmm. I have seen this, and by time they figure the (fan) load is the problem, there is a box of dud modules needing costly repairs at no fault of the module.
I would also question say 4 of 20A fans and 80A feed for a board, it's not practical due to the heat generated by connections, pcb traces, and components. I had a hard time designing for high ambient temperatures with high currents and potting. I don't know your details but be careful you don't make a silly product.  Something that fails a lot can sink a company.
Possibly an electronic fuse by monitoring load current, and MCU with trouble codes would be better.
 

Offline SimonTopic starter

  • Global Moderator
  • *****
  • Posts: 18118
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Fuse selection (blow times)
« Reply #30 on: August 25, 2019, 08:26:56 am »
As i already said it is to satisfy compliance and the loads are unlikely to blow a fuse but the customer wants one. So a fuse they get that will never be needed but it's there.

No, "the customer is always right" does not 100% apply to engineering. The customer does not know the safety standards, or have experience doing product development. Blindly following what they want can be a noose around the engineer's neck. You have to explain and negotiate the product's requirements with a customer, sales/marketing, upper management etc. Sometimes you have to firmly say "no" to their demands because otherwise you end up making something unsafe or unreliable.


Lets leave the customer politics out of it. I am replacing an existing item that has already been use with satisfaction although clearly it has not yet had to operate.

Quote
If you use non-replaceable fuses, your product is only as good as the fans.
A lemon fan- mosfets that short, bearings failing, water getting in, plastic blades cracking and flying off- mean overcurrent and your fuse pops. A guy puts a new module in and it also pops a fuse. Hmmm. I have seen this, and by time they figure the (fan) load is the problem, there is a box of dud modules needing costly repairs at no fault of the module.

Correct. So as already stated the fans are very reliable. They will shut down automatically on any overload and do soft starting. I have never seen any fail. The fans are IP68 so water is not an issue in creating any electrical problems. You also cannot simply replace the fuse module, it's too burried. The fans are rated for 40'000 hours at full speed and these are speed controlled down so would potentially last longer.

Quote
I would also question say 4 of 20A fans and 80A feed for a board, it's not practical due to the heat generated by connections, pcb traces, and components. I had a hard time designing for high ambient temperatures with high currents and potting. I don't know your details but be careful you don't make a silly product.  Something that fails a lot can sink a company.
Possibly an electronic fuse by monitoring load current, and MCU with trouble codes would be better.

Yes indeed, currently 2 layers of 70µm are used. I am looking to use 210µm. The problem is that I will also have my MD comparing my new cost to the old cost. I think i will also consider 0.8mm PCB to help move the heat to the second layer and make vias shorter and i am considering having some pins soldered in a pad to take the corrent through.

No electronic fusing is not possible for the same reason the fan mosfets can blow before the fuse does so can electronic protection blow with ease. You have to have mechanical circuit breaking somewhere.
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 7678
  • Country: ca
Re: Fuse selection (blow times)
« Reply #31 on: August 25, 2019, 04:08:50 pm »
I've learned to look at the whole picture, anything off in the original requirements can make things end badly. I trust the fan's design will never change and they can withstand automotive -ve overvoltage transients without failure...

I think heat will be your largest hassle, as everything wants to make a watt or so of heat. If you need to run in elevated ambients say over 70C then it is very difficult to keep it small and cheap. Slow-blow fuses have high resistance and will operate hot.

People I know in the automotive industry are (low volume) using aluminum Mcore pc boards for heat sinking. High production volumes like ECU's using thin FR-4 bonded to the die-cast metal enclosure. There are many high-tech thermal adhesives. It can be cheaper to spend more bom dollars on lower RDS-on mosfets (=cooler) to save more on the enclosure and assembly labour.

Silicone rubbers I find are terrible for cooling, like wrapping your board in a blanket. You lose all convection cooling and the encapsulant acts as an insulator. Hard epoxies are better at conduction heat transfer. You can also use the product's flying leads as heatsinks out the wires.

Most all fuses are characterized and rated in free-air, and when potted their properties change significantly. The potting compound provides conduction-cooling so the fuse's trip current becomes much higher with longer trip times. The curves are no longer valid after potting. Some fuse end caps/housings have air gaps and encapsulant can seep in to the element and make the fuse never really trip. The engineers never notice, but regulatory approval agencies that are on the ball will refuse. Fuses that are already epoxy encapsulated have an internal air space, potting compounds cannot get inside so their characteristics do not change and they are suitable for potting. Littelfuse engineering is very helpful if you have any questions.

Small SMT fuses usually have like a solder mask over top, so potting does not get inside but does hamper heating of the element, and the trip curves are shifted. You can see construction under a microscope, the element is on ceramic substrate.
Your fuses can only protect the wiring and pcb against fire and melting, so seemingly high trip currents I think are fine. The hold current must not be too high so an unhealthy fan drawing 50% more current is not allowed to keep going.

If you have an MCU you can monitor the output voltages and know if a fuse has tripped and blink an LED or something. It's just a few extra resistors. I found cooling fans are almost mission critical and the system (engine, pumps etc.) overheats so a fault or alarm was helpful.
« Last Edit: August 25, 2019, 04:10:28 pm by floobydust »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf