EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: MrMetthew on March 16, 2018, 11:09:45 am
-
Hi guys,
PPTC's are awesome parts. When to much current flows through them, they get high-ohmic, when the over current stop they go back to low-ohmic state! Looks to me perfect to protect against powerline overcurrent situations!
What do you guys think as PPTC as a series input power line detection device to stop big peaks on power lines? Any experience? What about the longevity of this solution? I worry because this is a series device so, when/if it after a while does not return to zero Ohm it will make problems. I don't find to many reference designs that use a PPTC, why not?
I specifically like to know about PPTC's, not about other solutions thanks !
-
Polyswitches are used nearly everywhere - pretty much every USB port around has one, for instance.
Do note that they take some time to respond to faults, and have voltage limitations.
-
Hi guys,
when the over current stop they go back to low-ohmic state! Looks to me perfect to protect against powerline overcurrent situations!
Not really. They typically only need about a quarter watt of current flow to maintain that high resistance. Generally power has to be removed completely for them to reset. That is quite a pain in most situations. Their primary use is for someone doing completely stupid and not destroying the electronics.
-
Polyswitches are used nearly everywhere - pretty much every USB port around has one, for instance.
Thanks for the feedback, did not know they were used that often!
-
Not really. They typically only need about a quarter watt of current flow to maintain that high resistance.
Thanks, I'll need to check, simulate and test that! I actually find these PPTC datasheets quite lacking! For example cooldown times to low ohmic states over the different specified operating temperatures are guesswork. I kan only assume going back to low ohmic condition takes longer at 85degC than at -40degC.
-
Yep, and also watch the temperature derating carefully!
They are fine for protecting against things like short circuits on powered IO connectors, but you basically use them where a fuse would otherwise suit, and size them much as you would a fuse @(then derate for temperature).
These are not precision devices, and they can be quite slow, particularly when the power source has limited current capability which caught us once when an external device overcurrent PTC was slow enough that the power supply went into hickup mode first, replaced it with something mosfet based, problem solved.
As to power line shennanigans, what exactly are you trying to protect against? I could see using one (in combination with a butch diode) to catch reverse polarity (maybe), and they are somewhat useful for gross faults as long as you respect the (often painfully low) interrupt ratings, but any time there is big PSC in play, or you are anticipating large voltage excursions they tend to be a poor choice, use as well as rather then instead of a real fuse.
Most of the time (With the possible exception of powered IO ports) I would rather see a real fuse, if sized correctly they basically never fail unless there is a real problem, and unlike the PTC jobbie leave evidence of the problem for you to find. A proper fuse is also very much less temperature sensitive and has a useful voltage and PSC level.
Regards, Dan.
-
Yep, and also watch the temperature derating carefully!
They are fine for protecting against things like short circuits on powered IO connectors, but you basically use them where a fuse would otherwise suit, and size them much as you would a fuse @(then derate for temperature).
These are not precision devices, and they can be quite slow, particularly when the power source has limited current capability which caught us once when an external device overcurrent PTC was slow enough that the power supply went into hickup mode first, replaced it with something mosfet based, problem solved.
As to power line shennanigans, what exactly are you trying to protect against? I could see using one (in combination with a butch diode) to catch reverse polarity (maybe), and they are somewhat useful for gross faults as long as you respect the (often painfully low) interrupt ratings, but any time there is big PSC in play, or you are anticipating large voltage excursions they tend to be a poor choice, use as well as rather then instead of a real fuse.
Most of the time (With the possible exception of powered IO ports) I would rather see a real fuse, if sized correctly they basically never fail unless there is a real problem, and unlike the PTC jobbie leave evidence of the problem for you to find. A proper fuse is also very much less temperature sensitive and has a useful voltage and PSC level.
Regards, Dan.
Well I actually want to protect for relative long high voltage and very energetic peaks generated by big motors and alternators. Right After the PPTC I want to put a good size varistor which will start pulling the energy when it starts clamping in case of overvoltage. The initial energy will go into the varistor giving the PPTC time to come into action. Varistor allone is not a good solution as the peaks remain to long (multiple 100's of ms and +200V) which would force them to dissipate huge amounts of energy. You wont believe the varistors I've seen blown up !
I think I can get a much more efficient protection putting the PPTC in series, allowing me to use way smaller, even smd varistors. But am a bit affraid of the PPTC's since I have no experience with them myself or have seen any field proven designs with this kind of protection. It would be a beautifull solution though in my opinion.
-
Well when I say "huge amounts" of power, I offcourse mean relative to the things I am working on :)
-
Well, if you want a polyswitch rated at 200+ volts, you seem limited to 4A or below trip current. Also, the hold current for these devices tend to be half of the trip current.
Instead of asking "Will X work for my vaguely specified application", try asking "What's the best way to protect my circuit from X and Y conditions with these parameters? I was thinking about PPTCs and varistors".
In this case, let me ask some questions:
* Do you need power to be supplied during the transients - that is, clamped to a safe level - or can it drop out during the transient?
* How much current does the circuitry you're protecting need?
* What is the nominal voltage your circuit sees, without transients?
* Do you have any voltage regulators for the protected load, and if so, what is the minimum voltage margin you have? (For example, if you're powering a 5V load from a 12V lead acid battery, you'll have a 3-4V margin at any time unless you have some monstrously large loads)
-
Well, if you want a polyswitch rated at 200+ volts, you seem limited to 4A or below trip current. Also, the hold current for these devices tend to be half of the trip current.
Instead of asking "Will X work for my vaguely specified application", try asking "What's the best way to protect my circuit from X and Y conditions with these parameters? I was thinking about PPTCs and varistors".
In this case, let me ask some questions:
* Do you need power to be supplied during the transients - that is, clamped to a safe level - or can it drop out during the transient?
* How much current does the circuitry you're protecting need?
* What is the nominal voltage your circuit sees, without transients?
* Do you have any voltage regulators for the protected load, and if so, what is the minimum voltage margin you have? (For example, if you're powering a 5V load from a 12V lead acid battery, you'll have a 3-4V margin at any time unless you have some monstrously large loads)
Thanks for the feedback! Reason I ask about the PPTC is that I want to research this solution specifically! Lets just say a certain solution got shoved my way ;)
*Power can drop during transients, I have power backup-circuitry where necessary, but power should restore by itself after the event
*Power consumed is about 10W-15Wmax, supply voltage is specified from 9V to 36V, so worstcase 1.67A (very very rare but within spec)
*Supply voltage is 9V up to 36V, we only guaranty normal working of the device in this voltage range. I like to start using this as a standard solution so I don't want to check a specific circuit if it can operate under this 9V
The voltage spec was a really good tip. Believe it or not, but I actually noticed that I had missed that the specific PPTC I was eyeing did not comply to this spec! Like they say, hardware is hard :p
-
Two fundamentals first:
1. What about AC?
PPTC's are awesome parts. When to much current flows through them, they get high-ohmic, when the over current stop they go back to low-ohmic state! Looks to me perfect to protect against powerline overcurrent situations!
This is correct at DC: the resistance increases, and there you have it.
That's it, right? Nevermind how long it takes to get there, as long as it happens? Well, no, it has to open before, say, wires melt. Hmm.
Okay, so we must consider the AC characteristics. That is, how fast it operates.
So it is a question, not just of current, but of charge. But not just charge, but dissipation, because it is a resistive element, mostly.
Most fuses carry an I^2*t rating, representing this quantity.
2. Applicability:
Polyswitches are used nearly everywhere - pretty much every USB port around has one, for instance.
Do note that they take some time to respond to faults, and have voltage limitations.
Unfortunately, in most of these applications (say, a powered USB hub), the power supply browns out first.
If the power supply can deliver quite a large current (10A+) or is a linear (non current limited) type, then the fuse will become important. Otherwise, it's unlikely to be of any benefit!
The time scale of relevance here is hundreds of milliseconds. Some seconds for modest overloads, like twice the rated current. Maybe tens of milliseconds under a short-circuit fault.
It only takes tens of milliseconds to discharge an SMPS output, so the fuse may not open!
Keep in mind, also, that transistors melt in 100 microseconds or less!
PTCs have three other ratings: maximum voltage and current, and min/max resistance. You use these by first selecting a part with Vmax > Vsupply in your circuit (this prevents the fuse from bursting into flames when it opens), then from that list, selecting Rmax such that Inom*Rmax does not drop an inconvenient amount of voltage (this resistance adds in series with your supply -- it better not need to be a precision regulated supply). Finally, select Rmin so that Vsupply / Rmin < Imax. This again prevents the fuse from bursting into flames when it opens.
If you can't find one with high enough Rmin to satisfy both of these, consider adding external resistance, or using another method, like a hot-plugging controller or protected switch.
Note that a protected switch only offers, at best, a few milliseconds of current limiting (before either turning off safely, or burning up and failing shorted), because a MOSFET die does not have much heat capacity. If you need more power, consider using a switching controller: a buck converter that can operate at 100% duty can be set to Vout > Vsupply, to obtain overcurrent and overvoltage functions!
Tim
-
You can use a load switch that has a adjustable off time between power interruptions/low voltage, then adjust that off time to match the worst case thermal mass of the PTC, to ensure that it can cool down before power is restarted.
-
Thanks for the feedback! Reason I ask about the PPTC is that I want to research this solution specifically! Lets just say a certain solution got shoved my way ;)
*Power can drop during transients, I have power backup-circuitry where necessary, but power should restore by itself after the event
*Power consumed is about 10W-15Wmax, supply voltage is specified from 9V to 36V, so worstcase 1.67A (very very rare but within spec)
*Supply voltage is 9V up to 36V, we only guaranty normal working of the device in this voltage range. I like to start using this as a standard solution so I don't want to check a specific circuit if it can operate under this 9V
The voltage spec was a really good tip. Believe it or not, but I actually noticed that I had missed that the specific PPTC I was eyeing did not comply to this spec! Like they say, hardware is hard :p
Then no, a PPTC isn't really a good solution to be honest. You'd need a quite low nominal resistance to cope with the 1.6A load, and it'd be way too slow to open during a transient. They'd be little better than just a series resistor.
I suggest looking at various automotive solutions. Something like the MAX6397 (https://datasheets.maximintegrated.com/en/ds/MAX6397-MAX6398.pdf); this one isn't optimal for your application since the max voltage spec is a bit on the low side.
-
PPTCs and PTC thermistors respond too slowly for protection against power line transients. Like fuses, they are used for safety and not circuit protection.
-
PPTCs and PTC thermistors respond too slowly for protection against power line transients. Like fuses, they are used for safety and not circuit protection.
I think they can actually protect some stuff, like transformers getting hot or motors.
-
PPTCs and PTC thermistors respond too slowly for protection against power line transients. Like fuses, they are used for safety and not circuit protection.
And if I put a varistor right after the PPTC which takes the initial energy so the PPTC has time to open? See my earlier explanation. That should work no?
-
And if I put a varistor right after the PPTC which takes the initial energy so the PPTC has time to open? See my earlier explanation. That should work no?
The PPTC protects against sustained overcurrent situations. Your varistor protects against short overvoltage transients.
If your transients are less than a second long, the PPTC will do absolutely nothing. And if they're longer than a second, they're no longer transients, to be honest.
Remember: The transient energy must be high enough to heat up the entire PPTC to 120+ degrees C - and a lot of it will just go into heating the varistor.
-
I actually use 4A versions on circuits that draw more than 10A for a couple of seconds. They are quite slow. Specs are like nailing jelly to a tree. You can get them to reset just by touching one with your finger. Lead length is critical. Make nice oven temp regulator if 100C is needed.
-
Amen about the specs.
The things have their place, but think most carefully when deciding what that place is!
You might find LTs "Surge stopper" parts together with a mosfet to be a better fit, they are designed to survive automotive load dump. Similar things are available from the other usual suspects.
Regards, Dan.
-
Compared to metal filmament fuses, polyswitches have high normal state resistance and a bigger ratio between the "must conduct" and "must open" condition. They also usually have much lower hold-off voltages. In addition, after tripping once, their normal state resistance increases substantially and permanently. Finally they are much more sensitive to ambient temperature. So basically they are completely inferior to metal fuses in every way except that they don't need to be replaced.
Like most fuses, they are primarily to protect against fire, or other damage from a sustained short circuit. With all fuses, you have to make sure that in case of a short circuit, the power supply can produce enough current to guarantee that the fuse blows. This can be hard to achieve with poly switches. Consider a USB charger with a 5V/2 amp rated power supply that you want to protect with a polyswitch. You will need one rated to carry more than 2 A, maybe 2.5 amp. That fuse may not guarantee to switch until you hit 5-6 amp. If the short circuit output of your power supply is only 4 A, the fuse does nothing. If 4A is safe, you don't need it, if 4A is a problem, the fuse doesn't help.
An area where poly switches are useful is if you have a centeral power supply that fans out a lot. So if you have 10 USB ports on your motherboard, each rated for 2 amp, you need a 20 amp power supply. But you can put a 2.5 amp polyswitch on each port and protect against any single port short circuiting.
-
PTC's are nice devices but they have their shortcomings.
Their nominal series (non-tripped state) resistance is usually inversely proportional to their rated current capacity - meaning that for low current PTC's, their nominal series resistance can be in the order of several ohm. That may not be negligible depending on your application and power draw. For instance, a 100 mA nominal current PTC can have a resistance as high as 10 ohm or over. That will translate to a degraded efficiciency.
Also, their nominal series resistance will gradually degrade, ie. increase, every time the PTC has gotten in the tripped state.
So, in some application where power efficiency is more important than low-maintenance and the probability of overcurrent is low to medium, a classic fuse may be a better solution.
Just my 2 cents.
-
As mentioned above, PTC resistors gradually become more sensitive, every time they're tripped. What happens eventually is their resistance becomes so high, they trip under normal load conditions, rendering the device useless, until they're replaced. I've also encountered PTC resistors which have failied in the permanently tripped state too.
-
Yeah, datasheet gives limits on Rmax after number of cycles -- it's not all that many!
A fuse is a fuse: a protective device, a last resort safety feature. Be thankful that you're getting any cycles out of it at all; design accordingly. :)
If you need many resets, use a semiconductor based limiter.
Tim
-
You can use the temperature sensitivity of the polyfuse to your advantage: when you need both thermal and overcurrent protection, thermally couple the PTC to the component you want to "monitor". If it's getting hot, the overcurrent threshold lowers, even down to almost zero.
-
I've used that on a few occasions, put the PTC in a heat-sinking pour around a transistor collector/drain for a thermal foldback current limit.
This combination only works for a modest range of situations -- too much short-circuit dissipation, and the transistor melts before the fuse opens. (The fuse then opens at fault current, but the circuit won't recover on cycling.) Too little, and the transistor never gets hot enough to bother -- drop the fuse, it's wasted cost.
The other quirk of that being, last time I built and tested this circuit, the transistor dissipated the 1W just fine, and the fuse never opened (it got to ~100C, but not the 150C+ needed to activate the fuse). So this was on the low end, it would seem; but it may still open up at elevated ambient, which is nice to have.
I would recommend this approach for SMTs in the SOT89 to DPAK range, at short-circuit power levels of 2-20W or so.
Tim
-
A certain commercial large-pack li-ion distributed balancer system (can't remember the name right now) uses a big power resistor to shunt charge from the cell, and couples the power resistor with a series PTC. Since power resistors are usually OK with 150 degC temperature, and, unlike semiconductors, do not care about reaction time that much, the PTC works quite well for that purpose. Here, the PTC protects from the stuck-on resistor control (duty cycle too long), too high ambient leading to too high resistor temperature, and so on. IIRC, the PTC was in series with the complete module, causing loss-of-signaling when active, causing a fault detection.
I still hope they don't rely on the PTC regulating the temperature all day long, even though it could work like that in theory.
-
If you need many resets, use a semiconductor based limiter.
Tim
With a fuse as backup, if there's a risk of fire.