0
\$\begingroup\$

I have a query.

I understand that the inrush current is the current that appears during the initial turn ON of the Board so as to charge all the Capacitors present in the board.

But I think that, this initial inrush current not only depends on the capacitors that are present on the board, but also the Power Supply rise time.

I found that there is a difference in the inrush current values when my power supply rise time is 300us and when my rise time is at 100ms.

My questions

  1. What should my the actual power supply rise time be so as to calculate the maximum initial inrush peak current of my board? Is there any sort of standard regarding this Power Supply rise based on application?

Based on the power supplies which one is using, power supply rise times tend to differ. I assume it is due to the presence of the capacitors at the outputs of the Power Supply.

In real life, my PCB board runs on a 12V Lead Acid Battery. How to simulate the Power Supply rise time of the Lead Acid battery using a normal 32V/5A power supply.

  1. How does the inrush current vary when one uses a Bench Power Supply and when one uses a 12V Lead Acid Battery.
\$\endgroup\$
6
  • \$\begingroup\$ How will your circuit become connected to the battery? How long are the wires, what impedance are the wires? \$\endgroup\$ – Andy aka Sep 10 '20 at 9:38
  • \$\begingroup\$ The Wires connected to the battery are 25 sq.mm and 4 feet each \$\endgroup\$ – Newbie Sep 10 '20 at 9:40
  • \$\begingroup\$ I never see a different approach: expect external Power supply's to deliver maximum. Do not expect external Power supply's to have a soft start or such. If you want to control inrush current do it on your board. \$\endgroup\$ – schnedan Sep 10 '20 at 14:56
  • \$\begingroup\$ Yes I agree. But in general, I am asking. Say my power supply maximum voltage for the board is 16V. What should be the ideal Rise Time for a power supply from 0V to 16V or in the case of the 12V Lead Acid battery from 0V to 12V? \$\endgroup\$ – Newbie Sep 10 '20 at 15:17
  • \$\begingroup\$ well without further constraints its zero nanoseconds. seriously. if your load needs more current than available only at turn on, ramping up the power supply might do the job. But then it is determined by the current. Also you might not wan't a fuse to trigger. But there is no general rule: a 16V power supply should have a rise time of X. \$\endgroup\$ – schnedan Sep 10 '20 at 16:55
1
\$\begingroup\$

We've scratched the surface of this in another topic of yours, recently... I mean the questions of inrush and PSU turn-on ramps etc. I have some loose comments:

From the perspective of "rail voltage ramp slope" on power-up, this is a question of what the powered device (power consuming circuitry) requires. Your own constructions or various ad hoc circuits may have no clear / documented requirements (e.g. unless/until you diagnose an issue in that vein) and may technically be pretty tolerant. Note that more complex circuits, such as computer motherboards, may have specific requirements, or industrial standards that they should adhere to, for the sake of market-wide multi-vendor compatibility - such as the ATX spec. A PC motherboard has several input power rails with different voltages, and may need a particular pattern of "power sequencing" = the order and timing and slopes (or mutual timing windows) among the power rails and control and status signals, in order for all the mixed-signal circuitry on the motherboard to come up and start to function correctly. Speaking of the ATX standard, the key power rails have a defined slope of the start-up edge - the duration of the slope from 0 to full voltage should be between 1 and 25 ms = notice that there is a minimum value. I can understand that minimal time to assure a few different things: it helps mitigate adverse effects of inrush (see the next section of this text) and, combined with some thresholds and latencies at which individual circuit blocks come alive in the powered device, it provides for a deterministic "startup sequence" within the circuit.

Speaking of inrush, and "how much is too much": that's a tough question. Often you have no good spec for the powered devices. And the adverse and entertaining effects of inrush typically become apparent in a more complex topology, where a strong central PSU feeds low-volt DC to multiple powered devices. People often design such circuits with the idea that a strong central source will be robust and assure reliable operation - and forget about the inrush effects of switching individual powered devices ON, ignoring the capacitive character of those individual loads... Capacitors at SMPS output can have milliOhms of ESR. The same is true for a self-respecting battery. The input capacitor in a small powered device can have a couple dozen milliOhms of ESR. There's some wiring and contacts in the game... Ohms law applies, so for a couple microseconds, you have a short circuit running at e.g. 24V / 0.1 Ohms = a couple hundred Amps. As for the adverse effects: the current pulse in the PSU path has a very high dI/dt and some nice loop area. Meaning: it radiates far and wide. It affects small-signal circuitry. The corresponding short-term brown-out may not be too much of a problem (the power rails inside devices tend to have a couple milliseconds of holdover slack) but the EM pulse itself can freeze or reset devices or circuit blocks around, even in gadgets that are not directly powered by the culprit rail. Immunity to EM pulses produced by inrush events sure has something to do with EMC testing and norms... IMO the worst adverse effect is the "ground wobble" caused by the inrush event. You can imagine that the short circuit event causes a short-lived brown-out on the rail voltage - but the return path = common / ground wires have their resistance and contacts too. The short circuit on flipping a power switch effectively means, that your powered device has the + and - at equal potential, centered between the +rail and power supply GND. Mr. Kirchhoff plays hide and seek in the web of power return paths, reference grounds and even Protective Earth and chassis grounds in your system. The signal reference potentials suddenly dance around, causing intermittent disruptions to communication, making things freeze, and can actually blow signal transceivers on communication ports that are not resilient enough. Imagine modern industrial process control, including PC computers with USB ports, where the PC's and various peripheral gadgets are DC-powered, the input SMPS in every powered device is a plain buck converter lacking earth isolation, and the interconnect is USB which is never isolated (too fast and too cheap). That's exactly where the power ground doubles as a signal reference potential... now snap an inrush spike of a couple hundred amps into that mesh of GND connections :-)

The honest and correct way to solve this is: implement inrush limiting in your powered device, or using a tiny external gadget (which does not exist on the market). What you want is an active inrush limiting device, semiconductor-based, solid state, featuring instant response, with a safety timer to prevent thermal overload.

Another story from this environment, slightly off topic: RS485 serial ports are sometimes isolated, but often are not. What could happen, right? The RS485 is balanced (differential), so it doesn't even need a reference ground conductor in the transmission line (just some closet purists insist on having a reference ground conductor in RS485 wiring). And both the signal wiring and the power wiring uses connectors of various sorts, including DC barrel jacks and detacheable terminal bars. There's no earth isolation anywhere. No contact sequencing in the connectors. And a popular way to power-cycle some device is to pull its DC power cord and plug it back. Piece of cake. Or, you have the powered device powered by an earth-isolated SMPS DC/DC adaptor, safety class II = no PE terminal on the output. Except that your RS485 has no reference ground either, or does have one, but the connector does not prevent the signal pins from engaging first, before the earth connection has been firmly established. And that external SMPS DC/DC adaptor has non-zero Y-cap leakage. Say 1 mA short circuit, 150V AC open ended. I have some USB/485 dongles around, without earth isolation. Every now and then, someone borrows these, and returns them with "this one is broken. This one's good." While on the oscilloscope, I can see that some of the switches in the transceiver are more or less rotten again. So I replace the line driver chip and let the history repeat itself...

Earths are boring right? Pretty much all at the same potential. Just some safety nazis dabble in strict separation of PE from the PSU return paths and from signal reference grounds... until one day, inrush and grounding bite you in a larger DC-powered system, turning you into another grounding zealot ;-) That is, provided that you can even grasp the broader problem, and the individual stumbling stones...

\$\endgroup\$

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.