NTC Heat dissipation

[DBoulanger]

It might sound as a silly question but is using a heatsink to cool off a NTC for inrush current limiter a viable approach ?

When looking at the datasheet of some NTC, the manufacturer indicate the temperature the NTC can reach some 200 degrees at its maximum current specification.

Quite a few people are stating that it takes roughly between 30secs and 3 minutes or so for the NTC cool off and get back to its highest resistance at the ambient temperature.

So would a heatsink be something logical to use with a NTC, or does it defeat the overall purpose ?

[DBoulanger]

Hi,

Perhaps explaining a bit what is my goal might help understand my question a bit more clearly.

I have a power tool (grinder) which generates a silly 58A inrush current @ 120VAC.

Quite a few circuit breakers don't like that inrush current amplitude and simply trip.

Once the tool works as planned, it draws about 6-8 Amps, despite the manufacturer is rating the motor at 12A.

To take care of the inrush current, I've been suggested to use a NTC.

What bugs me is the fact that the NTC seems to dissipate a fair amount of heat, so I'm rather concerned at attempting to put it inside the tool itself.  Also, to ensure proper operation, the NTC must cool off enough in order to be somewhat efficient at the next tool startup.

If the startups are too close to each other, the NTC may not have the time to cool off.  Also, with a work tool like that, having to wait 2 minutes between each use is kinda spoiling the fun.

Using some sort of heatsink help, would allow to minimize the waiting time, at least this is what my understanding of the things lead me to believe.

Overall I'm looking for an Inrush Current Limiter approach that is efficient enough and could be re-triggered within a more convenient delay than 2-3 minutes.  Ideally speaking, not having to modify the tool itself would be ideal.  Any idea / circuit to suggest ?

[Cliff Matthews]

I was thinking along the lines of a choke or a run cap (a 25uf can from an AC unit).. But I'll make a guess there's good reading here:
http://sound.westhost.com/articles/inrush.htm

[David_AVD]

How about using a (timed) relay to bypass the NTC after the surge is over?  That way the NTC can start cooling as soon as the relay energises.

[T3sl4co1l]

The NTC is supposed to get hot.  Very hot.  In fact, the hotter it gets, the less power it dissipates!

You can't go too hot, because its materials will break down.  But you want to go up there as well as possible, to keep losses down.  So, put up some heat shields, or wrap it in heat shrink (a common sight), or etc.

Tim

[DBoulanger]

As a matter in fact I was thinking going that route "using a (timed) relay to bypass the NTC after the surge is over".  I would put the circuit in an electrical box that can be fixed on an electrical extension cord.  Ideally, I'd like the circuit to be transformerless.  I searched in the past for something similar but didn't get much success.  Would you happen to have a circuit to suggest ?

[David_AVD]

Whenever I've used a relay to bypass an NTC, it's been controlled via a small micro.

You can get time delay relay modules that are powered from the mains.

[oldway]

Whenever I've used a relay to bypass an NTC, it's been controlled via a small micro.

You can get time delay relay modules that are powered from the mains.

In this case, there is no advantage to using an NTC.
On the contrary, there is a major drawback:
in cases of frequent starts, the NTC does not have time enough to cool down and may no longer make his function of start current limiter.
You must use a power resistor instead of the NTC.

[David_AVD]

Well, it would depend on the duty cycle of the tool use of course.

I think Silicon Chip magazine did a soft starter that went inline with a 240V tool.  Maybe check the archive list on their web site.

[DBoulanger]

I think Silicon Chip magazine did a soft starter that went inline with a 240V tool.  Maybe check the archive list on their web site.

You're right, Silicon Chip magazine did a soft starter.  Unfortunately, the documentation isn't accessible online unless I register to the magazine, which I'm not ready to do only for one article.  As you stated, the kit is designed for 240V.  The actual specs are as follow :

)) Inrush current limiting: <20A  --  I need something that could handle close to 60A inrush current, so this is too low obviously
)) Minimum load power: ~100W  --  The power seems to be a bit on the low side considering the maximum current load specified below. How can they come up with 100W using potentially 10A @ 240VAC ?  Not too sure about this spec anymore.
)) Maximum load current: 10A  --  Close to what I need.  As per the manufacturer, the tool I have is rated at 12A, however I measured 6A while the tool was idling, so perhaps 10A is close to the real thing when actually working.
)) Minimum tool restart interval: 60s recommended  -- It strangely looks like a simple NTC scenario.

Finally, Jaycar and another electronic kits distributor, are selling this circuit as a kit.  They're asking around $50AUD which I find fairly expensive considering that Jaycar is estimating the assembly time at roughly 3 minutes, including the soldering, so I would believe that we can easily count the parts on one hand.  I don't want to negatively comment the kit as such, but based on what I actually have access to, information wise, I'm not really convinced.

[DBoulanger]

Overall, one thing is clear, if a NTC is to be used in this project, either it will need to be bypassed in a timely manner so that it can get out of the picture ASAP and get ready for the next use, or the user will potentially need to wait few minutes before restarting the tool and still benefit from the NTC specs.

The timed relay approach or possibly a SSR, seems to be the logical way to go, minimizing the NTC cooling cycle.

The challenge here might not be immense for some of the forum users, but for some of us who have only basic understanding, things might be different.  The question is how would you proceed to put in place such timed relay and NTC, if you're not allowed to open the power tool and mod it ?

Globally, we have a power cord, a box in which the necessary electronics will be put in, along with the electrical outlet on which the tool will be plugged.  The only switch allowed to start and stop the tool is the one on the tool itself.

At the expense of adding a bit more components, I was thinking of using some sort of current sensor which would be used as a trigger to enable/disable the timed relay block/module.

Also, is it too risky to use a TRIAC based SSR when dealing with fairly inductive loads such as power tools ?  Would I need a snubber and if so, how can I determine the needed snubber specs ?  (I have a few [.5uF-600V / 150 ohms .5W] snubbers collecting the dust, so if I can put them to use that would be good.)

Or is it preferable to use a mechanical relay instead of SSR ?

Some schematics, when proposing approaches, would be greatly appreciated.

Just as a reminder :

- The tool operates at 120VAC, 60Hz
- It has about 58A inrush current
- It draws about 6A when idle and I would expect 10A full load
- The manufacturer rated the motor at 12A, but I doubt it will go that high, considering the type of work to be done.

[David_AVD]

The timed bypass is certainly going to be tricky if the switch has to be on the load side.

[Cliff Matthews]

Just curious... 58a sounds like a lot for locked rotor amps on a small motor. Are brushes and commutator in good condition? I've seen LRA numbers like this on 10,000 btu ac units. What passive solutions have you tried so far? (caps, chokes, etc..) What did you measure?

[DBoulanger]

Just curious... 58a sounds like a lot for locked rotor amps on a small motor. Are brushes and commutator in good condition? I've seen LRA numbers like this on 10,000 btu ac units. What passive solutions have you tried so far? (caps, chokes, etc..) What did you measure?

You're absolutely right, this is a fairly substantial "locked rotor" amps.  I was surprised to measure such a high value for such a small motor after all.  The previous owner sent the tool back to the manufacturer, Dewalt, indicating that he was simply fed up with the tripping breakers situation.  Almost each time he was starting the tool, the circuit breaker tripped.  Dewalt changed the brushes and the power cord.  Once the owner got it back, the same situation was still reoccurring so the problem seems to be at the design level, hypothesis to which the manufacturer won't agree, obviously.

Just by the way the owner was describing the situation, I had a strong feeling that we were dealing with an inrush current situation.  In fact, I experienced a similar situation with my central vacuum cleaner unit few months ago.  Every so often, as soon as the switch was toggled on the handle, the circuit breaker was tripping.  To solve the problem, I ran the appropriate wiring from the unit to the electrical panel, then added a 20A breaker so that the unit would be alone on its circuit, as suggested by the manufacturer.  The manufacturer manual was indicating that a 20A breaker would be more suitable.  Anyhow, since that fix/mod, the breaker never tripped again, so having my wife happy makes me happy too ! 

I tried the tool on the same circuit as my central vacuum cleaner and the breaker doesn't trip.  Still measuring a 58A inrush current with my clamp.  What is spoiling the fun here is the fact that the new owner need that tool at different sites, mostly residential.  Unfortunately, 20A circuits with appropriate breakers doesn't seem to be a common practice.  Most circuits are equipped with 15A and are tripping with that "stupid" tool.  For what it may worth, information wise, the tool is a Dewalt grinder, DW831, type 3.

By the way, I had a similar issue with my 2KVA variac.  Even if the dial is at its lowest, no load connected, the breaker is tripping too frequently.  Using a time delay relay with my variac, wasn't too complicated.  I simply relocated the ICL power switch at a more convenient location on my workbench, so the switch on the variac is always on, I just turn on the Inrush Current Limiter whenever needed.

Being able to use the same ICL device for different tools is a must.  Also, modifying the tools isn't something I want to do as this action will void the warranty instantly.  So yes, having the switch on the load side makes the overall thing a bit more tricky but I'm sure there is a way.

Beside measuring the inrush current with my clamp, I didn't try anything else.  I wouldn't mind attempting some "passive solutions", but I don't know how to determine what would be the appropriate cap or choke to start with.  It might not be of any use, but just in case, I measured the tool's DC resistance at the plug and I have a reading averaging around 1.5 Ohm, which seems rather small.  Just for the sake of it, I also measured the inductance which gives me a reading in the 8mH range.  With an inductance of that value, we come up with an inductive reactance of approximately 3 Ohms @ 60Hz.

Beside these measurements, I don't quite see what else could be measured and useful.

Hoping this is making the things a bit more clear.

[tautech]

My normal remedy and prevention of this sort of problem when running mains circuits close to max or when anticipating large inrush currents is to install C or D "motor rated" breakers.
Their short term over-current ratings invaribly let the start-up surge through and then stay un-tripped quite happily.
Be aware breakers that have experienced on-going tripping likely will not meet spec, so even replacement with the same might even solve your problem. 

A "D" class breaker fixes all of these problems IME.

[DBoulanger]

My normal remedy and prevention of this sort of problem when running mains circuits close to max or when anticipating large inrush currents is to install C or D "motor rated" breakers.
Their short term over-current ratings invaribly let the start-up surge through and then stay un-tripped quite happily.
Be aware breakers that have experienced on-going tripping likely will not meet spec, so even replacement with the same might even solve your problem. 

A "D" class breaker fixes all of these problems IME.

The proposition you're suggesting is making a lot of sense and I do appreciate your input in that regard.

The scenario here is a bit different here.  The owner has a tool and he needs it to do his work.  The work is to be accomplish at different locations, therefore there is no possibility to alter the electrical configuration at any location.  The owner of the tool must come up with some "magic box" that will prevent the breaker tripping at any location he'll be working at.

As far as I can see, most suggestions provided so far can't work right out of the box, simply because it is taken for granted that the ON/OFF switch is at the line side instead of the load side.  The global idea here is to create an ICL "module" that will somehow be "self contained".  Potentially a switch to allow a NTC/PTC/Resitors selection, depending on the magnitude of the Inrush Current to be limited, but this is pretty much it.  Beside that, the device/tool simply needs to be plugged to an electrical outlet fixed on the ICL box.  The ICL then does its little magic when the switch on the device is activated.  The trickiest part here is to avoid any device alteration, because the same ICL "module" can be used for different devices/tools.

As mentioned earlier, device alteration is pretty much out of question, due to the fact that any mod would void the warranty and also, if the tool is not needed frequently, the odds are that the tool will simply be rented and under these circumstances, tool modifications is not an option.

So the project I want to put in place is an ICL that can accommodate different tools/devices without modifying the tools/devices themselves.

In that line of thought, I believe that a circuit using a similar concept to the one I'm showing here, could potentially be considered to trigger the ICL as such.  On the site where I saw this circuit, they're saying that some current would flow through the diodes, thus creating a sufficient voltage drop which would in turn trigger a triac thus powering the trailing sockets when current flow via the control socket.

Keep in mind that what I'm looking here is simply a way to engage the ICL while the switch is on the load side.  Obviously, I won't be dealing with the diodes shown on the diagram here, their specs aren't what I'm looking for anyway.  Just imagine some resistors or anything else that would accomplish a similar job.

Just hoping my explanation is easier to understand by now.

[johansen]

2 ohms resistance (whether resistance or resistance plus inductance i don't care)
sounds like that's a really effiicent motor for 12 amps rated load.
144 times 2 ohms is 288 watts.. man that's actually low, compared to the harbor freight grinder i burned out a couple years back, it probably had twice the waste heat generation potential.

anyhow, i wouldn't complain, use a lighter duty power tool.
58 amps inrush is nothing.
i hope you're not one of those people who put 7 inch diameter grinding wheels on 4.5 inch grinders..

so given ya'll are a bunch of nerds.. you need to find a non linear thermal material to wrap the NTC in.

something that conducts heat better at low temperatures than high temperatures.

oh and a ratio of 2:1 for full load to no load amps.. that's pretty good.
by I^2 R losses that’s a 4:1 ratio.
for a small motor that's not bad at all.

[DBoulanger]

)) anyhow, i wouldn't complain, use a lighter duty power tool.

)) i hope you're not one of those people who put 7 inch diameter grinding wheels on 4.5 inch grinders..

Using a lighter duty power tool wouldn't be as efficient for the type of work the owner is using it for.  With the proper grinder wheel, it's powerful enough to take care of hard pointing.  Years ago, masons used a mixture of Portland and glue to make mortar.  At that time, everything was fine and dandy but when the pointing need to be redone years later, the job is simply a nightmare.  The old mortar is stuck on the bricks like you wouldn't believe.  If you don't pay attention to what you're doing, the brick can break in your hand in pieces and become unrecoverable.  Finding the same type of old brick, same tint, same dimensions, might become a journey on its own.  Using the proper tool for a given job is pure common sense.

The owner has been using grinders for years and no, we're not part of those people who put 7" grinding wheels on a 4 1/2" grinder.  We might be silly at times, but not that stupid.

Overall the timed relay approach seems to be the way to go.  Unfortunately, the fact that the tool can't be modified complicates the setup.