1
\$\begingroup\$

I am making a "lightbar" for my vehicle which consists of 7x 6" diameter spot lights linked together. The specs for each of the individual lights are below:

  • LEDs: 14x 3w OSRAM
  • Max Watts: 67w
  • Input voltage: 9-32vdc
  • Current Draw @ 13.8v: 5.6a

I have just finished wiring the lightbar, which I specced the wire thickness based on some previous bench testing of 2x lights connected to a 40a switching powersupply.

When I did this initial bench testing, 2 lights drew 6.5A @ 12v.

So, when I specced the wiring, I allowed for a total of 28a of peak load (6.5/2)x7 + bit for headroom.

I have just gone to test the whole array and my amp draw has increased considerably and I dont know what to do now, as I have some hard limitations which I need this lightbar to work to.

To bench test the power supply, I plugged in 1 light at a time, to check the amp draw. Instantly, just 1 light was drawing way more amps than my previous bench test!. The next odd thing that happened, was that at light 6, the volts dropped to 7.4 volts, but the amps stayed the same. The lights obviously dimmed as well. 30a was all I could push through the wiring and connectors as I didnt want to start a fire.

  • 1 light: 4.8a @ 12v
  • 2 lights: 8.4a @ 12v
  • 3 lights: 13.8a @ 12v
  • 4 lights: 17.9a @ 12v
  • 5 lights: 17.9a @ 7.4v
  • 6 lights: 22a @ 7.8v
  • 7 lights: 30a @ 8.4v

So, the hard limits I need to work within are:

  1. I use a Switch Pros SP9100 to control lighting circuits, which has a max output of 35a. So ideally, I need the load to be below this. Obviously, I could run a 70a relay to switch more current, but its added complexity.
  2. I dont really want to have to rewire the lightbar, which is currently wired to handle 28a.

So, firstly, I would love help to understand the below questions.

  1. Why did my initial bench test show such a low draw compared to testing today?
  2. Based on the above load @ voltage results, whats the total estimated amp draw the lights would make if they were plugged into my vehicle, which can provide upto 14.6v when running.

for solutions, I have gotten to this point.

  1. Use a 12v-24v step up converter, which will from my understanding, half the amp draw? But, what I am confused about is; A) Does the converter still draw the full watts (7 x 67w = 469w / 12v = 40a)therefore my switch pros 35a output would still need to use a relay circuit?? B) Does the draw of the lights half @ 24v, so I would need a 20a step up like this https://www.ebay.com.au/itm/New-DC-12V-to-24V-20A-480W-Step-up-Boost-Converter-Power-Supply-Module-for-Car-/321846640256 ?
  2. Rip all of the wiring out, beef it up to handle 40a of current and then use a 70a relay, switched from my Switchpros?

Im sure this is all fundamental electrical stuff, but it really threw me how badly I got the draw estimates wrong.

Thanks Ned

\$\endgroup\$
7
  • 1
    \$\begingroup\$ If you use patch cables to bypass all of the lights, you can pass a few amps through the whole wiring that you have done, and then should be able to measure the voltage drop -- start with one amp. My guess is that you have a short circuit somewhere. If the wiring tests out, then remove the patch cable to test one light at a time. Hypothesis is the short is inside or related to light 5. \$\endgroup\$ – MicroservicesOnDDD Jan 12 at 7:29
  • \$\begingroup\$ hmm, thats a good point. I will try this. But surely at 30a, a short circuit would have made it self known in a spectacular fashion. But regardless, an easy test to complete tomorrow! \$\endgroup\$ – Ned Morris Jan 12 at 7:38
  • \$\begingroup\$ One possibility is that the assumption that you only needed 28 amps total was wrong. Perhaps at 13.8 volts their current draw increases non-linearly, or one of the two lamps used in that test is actually not typical in some way. That sudden drop in voltage at light 5 is very interesting. It almost looks like your bench power supply is bottoming out, but then its voltage goes back up. Weird. Try the bulbs in a different order maybe. Maybe one light is not like all the others. I'd worry most about that sudden drop. I'd want an explanation. \$\endgroup\$ – DarylK Jan 12 at 9:18
  • \$\begingroup\$ Just to to clarify, the power supply has two knobs, one for amps, one for voltage. In this scenario, the voltage was set to 12v and amps was what it is (endless knob). Once the 5th lamp is connected, the volts drop. No amount of turning the voltage knob changes the volts. Only way to increase volts was to increase amps. \$\endgroup\$ – Ned Morris Jan 12 at 9:46
  • \$\begingroup\$ Those specs are dodgy as hell. 14 * 3W LEDs would take 42W not 67W, and 5.6A at 13.8V is 77W so there are two porkies right there in the spec. (porkies = pork pies = lies). Given that, you should support up to 5.6A * 7 = 39A, and don't expect too much life from these lamps. Also sounds like your supply is current limiting when you add the 5th bulb. \$\endgroup\$ – user_1818839 Jan 12 at 14:06
1
\$\begingroup\$

Ned, the individual lights are spec'd to accept a wide voltage range (9-32 VDC), so there's obviously a driver circuit in each of them. The driver rapidly turns the light on and off, much faster than you can see, to regulate the average power dissipated by the LEDs regardless of the different input voltages.

This means the actual current draw varies between some peak value, and near zero, hundreds or thousands of times a second. You could see it with an oscilloscope, but it's not detectible by the naked eye looking at the lights.

There is no synchronization between your multiple light units, so when you connect multiple lights in parallel, it's unpredictable how those waveforms will combine. This will create a complex waveform of current draw. You could get huge peak currents far beyond the specified average current. It could also interact with your power supply to trigger current limiting. And finally, your ammeter may not be able to accurately read the average current of a complex waveform.

You may be able to get away with wiring sizes that assume the spec'd average current draw is correct and ignore your meter. But I think that's dangerous. Can you find a friend with an oscilloscope?

\$\endgroup\$
5
  • \$\begingroup\$ @NedMorris -- If Mark is right, a large low-ESR electrolytic capacitor local to each LED light should help. Just please let us know what happened. So many just leave us hanging... \$\endgroup\$ – MicroservicesOnDDD Jan 12 at 19:31
  • \$\begingroup\$ Thanks @Mark Leavitt, a very well explained explanation to the odd situation that I am experiencing. I also think that your correct in your hypothesis that I may see current draw far exceeding the stated max draw x7, as when I was trying to feed more amps, they increased exponentially vs the rate at which the power supply could supply volts. I don't have an ammeter, so draw is just going off what the PSU display was showing as the amp output. I'm not sure if I can access an oscilloscope, but I will reach out. \$\endgroup\$ – Ned Morris Jan 12 at 19:56
  • \$\begingroup\$ The other kicker is I was planning to turn this into a product I sell, but if it's going to draw 40+ amps, it's going to make for a rather bulky installation, if I need to allow for a potential 60a draw, with wiring, relays etc to handle that kind of power. \$\endgroup\$ – Ned Morris Jan 12 at 19:56
  • \$\begingroup\$ @mark could you provide a link to an appropriate option for this capacitor? I would assume that this is just to level out any spikes, but wiring and connections would still need to be rated to 40a +? \$\endgroup\$ – Ned Morris Jan 12 at 19:57
  • \$\begingroup\$ Adding a capacitor just opens another can of worms; it will look like a dead short when you first turn the system on. Your system may still work fine from a car battery, which can handle high peak currents. It just won't work with the bench power supply you have. Before selling it, though, you would have to do thorough testing, because it could interfere with, and even damage, other systems in the vehicle. \$\endgroup\$ – Mark Leavitt Jan 12 at 20:32

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.