0
\$\begingroup\$

I want to be able to automatically switch on an AC-DC power supply edit for clarity: it's a constant current LED Driver [edited again for further clarity- the driver won’t actually be driving LEDs- this is just a cost effective off-the-shelf option for my specific task] (or potentially multiple supplies and a peripheral). While my steady state current draw will be quite low the supplies have a very high maximum AC inrush current of 50 A. There are a couple potential switching mechanism I might use, but they're all essentially simple mechanical relays, and of course are NOT rated for 50A AC current, and certainly not rated for 130A or so of the maximum inrush I might throw at the system if I switch everything all at once with a single relay. So I will need to mitigate this inrush or I'll most likely fry some components.

I've ruled out using simple resistance to limit inrush. Mostly because I think I need a relatively high power resistor, and to do that cheaply means wirewound, which has some inductance. I think that would be fine for limiting the in inrush but could pose some issues to then bypass the resistor after inrush with another relay since on an AC line I can't just use a simple freewheel diode. A simple solution that immediately gets complicated and requires several components. Correct me if my thinking is wrong here and that a wirewound resistor in series with a LED Driver should not be considered an inductive load and is thus nota problem to switch. Late Edit: I suppose I could use several 10 watt ceramic resistors in series so that there isn't any inductance making relay bypass after inrush unproblematic. This might be viable, and perhaps preferrable to using a SSR, but is definitely clunkier than an NTC Thermistor.

Question 1: I've read a lot of conflicting information about Zero Crossing Solid State Relays. If I can figure out what is and isn't true I feel confident I could make a good selection for my application. So, which of the following statements are true for why these might be a good choice for high AC inrush currents?

a) They dampen inrush by switching near the zero point.

b) They can handle short overcurrent events because there's no contacts to weld.

c) They can be found rated for 100 amps+ so most inrush currents won't exceed their rating.

d) Whether it is AC control or DC control the inrush current handling of the AC output side will be the same.

Question 2: Say a you have a 100 watt power supply that's about 90% efficient. Does this mean that the total power consumption during inrush is about 110 watts? The reason I ask is that I noticed that something like a 5V 1A wall wart supply can have an inrush over 20 A as well, but surely there must be much less power being drawn, right?

Now let's look for an NTC Thermistor! I think these would be my preference as I won't be doing any rapid switching on and off. I'm just a bit lost on sizing. My concern is that I get one that can't handle the load and I blow it up.

enter image description here

Direct data sheet links for the power supply and thermistor for those who want them.

On the left is the power supply I'd like to use. On the right are specs for a thermistor I think might be suited for my task. I chose it based on it's max current and steady state conditions, obviously, but it also seems to be able to handle a lot of energy, which seems like a good thing! I've just been ignoring the capacitance specs. There's a note elsewhere in the data sheet that these figures aren't absolute. I also have no idea what the bulk capacitance of these power supplies is. So I'm trying to tackle this through power/energy rather than capacitance, hence question 2.

For this scenario, let's say we are trying to switch on two of the pictured power supplies plus one unknown 5v 1a wall wart and we want to limit the AC inrush to 10A or less. We need to allow a steady state current of around 3A.

Question 3 Does it seem to be adequately sized? And if a smaller currents are applied, am I correct that the main difference would be that the thermistor won't heat up as much, leading to a bit higher resistance, causing a small voltage drop between AC mains and power supply input, but otherwise doing no harm?

Alternately, I could use separate NTCs for the supplies and use separate relays to switch them on if necessary. I'd still like to cap inrush current to 10A or less per relay though.

I think that's all of my questions for now, trying to keep this as brief as possible and not meander too much. Thanks.

Late edit post note: I'm on 120v mains power. I'm not terribly concerned with tripping breakers but that's another perceived benefit of limiting inrush rather than simply withstanding it. I'm pretty unlikely to ever see the maximum inrush currents given the usage pattern I'll be following, but I'd still like to size everything for the extremes.

\$\endgroup\$
7
  • \$\begingroup\$ I'm not sure but is all this about using off the shelf power supplies where you want to limit their inrush current using external parts that are added whose design is completely unknown and unaccounted for by the original power supply designers? \$\endgroup\$
    – jonk
    Jul 16 at 22:44
  • \$\begingroup\$ I mean, no not entirely. Depending on what is and isn't true about solid state relays, some of my potential solutions may not limit inrush at all, and will instead provide for a switch that can withstand it. At any rate, are you implying that it is inherently unwise to reduce inrush current potential to anything you did not design yourself? Because I might need some convincing of that being true. \$\endgroup\$ Jul 16 at 22:51
  • \$\begingroup\$ I guess I'm just suggesting that being blinded to what's inside a commercial power supply can make the creation of an appropriate external current limiting circuit a more difficult proposal. You may need to collect some detailed behavioral information, first. It's not generally the case that you can just design in a vacuum, so to speak, is all. \$\endgroup\$
    – jonk
    Jul 16 at 23:46
  • \$\begingroup\$ Fair enough. Essentially though, if I’m given a maximum figure from the data sheet as to what I can expect, I should be able to design to accommodate the maximum. So if I have a maximum inrush listed as 50a, and a maximum current rating of 10a on my switching device, I can design with those max figures in mind and know that I’m covered in any extreme event. If my actual use case is less demanding, great! Right? At any rate what I really need is some basic question answered so that I can make better informed decisions about how best to proceed with my design. \$\endgroup\$ Jul 16 at 23:58
  • \$\begingroup\$ The CL-70 would limit inrush to <10 A at 120 VAC, and have less voltage drop across it at operating temperature. However, either should work for your intended purpose. BTW, series-string valve (vacuum tube) devices often used NTC's to slow damage to filaments from inrush current. \$\endgroup\$ Jul 17 at 3:23

2 Answers 2

1
\$\begingroup\$

The peak inrush is when random phase voltage is at the peak or 141% of 120V. So obviously ZCS Triacs will prevent that peak and only experience decayed pulses of 60 cycles (x2) in the 1 second startup time with soft start by design.

which of the following statements are true for why these might be a good choice for high AC inrush currents? (FOR SSR's)

a,b

These designs are made to run 24/7 with a dimmer, so if you have that option, use it.

\$\endgroup\$
1
\$\begingroup\$

FYI, inductance is helpful if anything, but also mostly irrelevant at mains frequencies. It takes 100s µH to noticeably reduce surge/inrush/fault currents. (Which also means line filters are quite safe to place wherever.)

Some recent conversation, if you're thinking about mechanical contacts: Limiting cold-start current of an SMPS by an NTC
Other good threads in the sidebar.

Note that SMPS should already have inrush limiting, which you could reduce a bit further with additional limiting, but it's kind of diminishing returns, wastes more power (may want a bypass relay to short out the limiter after startup), etc.

Yeah, AC type SSRs, with zero crossing, are fine, as long as the surge and fusing ratings are respected. They use thyristors, which can handle quite high surge currents; along with diodes, these types hold the distinction of being the only semiconductors robust enough to carry a fusing rating at all! Check the datasheet.

Zero crossing is better, as accidentally switching on a line peak would draw a larger surge as the voltage rises suddenly. Not much larger, as long as the load does have inrush limiting as is -- the point is to charge over several mains cycles, which means the main capacitor isn't charging very far during the first say 1/4 cycle. But still, it's something. ("Strictly less" doesn't necessarily mean a lot less.)

Nice part about SSRs is they're easier to sequence, say delay by a couple cycles each. If you don't mind putting in some logic or an MCU (or already have one in the project) that is.

There's really no need to worry about inrush in general; nuisance breaker trips take quite a lot of charge, they're sized to handle motor starting for example. Typically curve B, check out MCB trip curves to see what that means.
https://www.electricaltechnology.org/2021/07/tripping-curves-circuit-breaker.html

\$\endgroup\$
4
  • \$\begingroup\$ So to clarify- I need to worry about inrush because my trigger signal to start the power supply will come from some sort of mechanical relay that can handle AC current but can't handle inrush. When you say that you don't need to worry about inrush with an SSR, does this mean that you rate the SSR for your steady state current and you don't worry much about the short overcurrent scenario associated with inrush? 50a max inrush is stated on their datasheet, so it definitely is not limited by the supply. Or do you mean that if I have an inrush current potential of x I need an SSR rated for x Amps? \$\endgroup\$ Jul 16 at 22:58
  • \$\begingroup\$ I clicked the link. I may be wrong about the mechanical relay's ability to handle inrush. I need to do some digging though because I definitely didn't buy Omrons. \$\endgroup\$ Jul 16 at 23:06
  • \$\begingroup\$ @DerekHershman Oh, it definitely is limited by the supply -- without, inrush can be hundreds of amperes; 120V circuit fault current is typically in the low kA. When I say "don't need to worry" I mean in terms of upstream, like the mains and breaker don't care as much as your SSR/relay will. SSR continuous rating may end up higher than apparently needed, dunno, just shop around by surge rating, see what's out there. \$\endgroup\$ Jul 17 at 0:53
  • \$\begingroup\$ Right, limited to no more that 50A, which is the problem of course :). At any rate, yes my concern has very little to do with mains breakers and much more to do with the integrity of my switching mechanisms. \$\endgroup\$ Jul 17 at 0:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.