Do you know a simple, inexpensive, elegant way of boosting 24V up by just 0.8V, to 24.8V (between 24.5V and 25V)?

My current goal is to power LED appliances but my question is not specificaly related to LEDs (see note below). The question is what would you do when you have a regulated 24V power supply and you need between 24.5V and 25V, and use it with a large amount of power, up to 10A for example?

Is there another way of boosting voltage when the voltage increase is small?

I'm looking for a solution with few parts, and if possible without expensive switch controler. The project is for small production volume (a few hundreds pieces), That's why I want it inexpensive and small, holding on a small size PCB.

  • The power supplied is already voltage regulated and stable.
  • The desired voltage increase is just 4%, betwee 0.5 and 1V.
  • It should sustain 8 or 10A continuously, safety margin included.
  • Current may vary between 0 and the above mentioned maximum.
  • The output voltage must remain stable within the 24.5 and 25V range no matter how much current is drawn. (Very short interruptions can be tolerated or solved by adding a capacitor if needed)
  • Different power supplies, albeit all rated 24V, may be used.
  • They may have different maximum ampere ratings (3A, 5A, 8A...) but will always be constant voltage supplies.
  • They may be build by different manufacturers, all using common practices, but possibly with slightly different schematics and qualities.
  • The boosting device could be sometimes placed at some distance from the power supply, sometimes right next to it, but always near the load.

Any idea?

(Note: I understand that some poeple feel unconfortable with the idea of powering LEDs without knowing the specifications in details, and I'm not against talking about it and providing more informations. And I do appreciate their contributions. Yet, I would like to focus on the methods of boosting voltage as described above. We can talk volumes about LEDs and it will quickely stray off topic.)

  • 7
    \$\begingroup\$ This sounds like an XY problem. Where are you getting the 24v and why do you think you need 24.8v? \$\endgroup\$ – Chris Stratton Jan 5 '18 at 0:13
  • 1
    \$\begingroup\$ That +0.8V won't help either. \$\endgroup\$ – Marko Buršič Jan 5 '18 at 0:27
  • 1
    \$\begingroup\$ I get 24V exactly from the power supply. But the LEDs are optimaly lit at 24.8V (+- 0.2V). At 24V they are abit dim, above 25V they heat too much. Specifications from the manufacturer is not clear: Between 23.8V and 28V. Even if the LEDs were optimal at 24V, this "exact 24V" is annoying because of inevitable voltage drops in complex systems of wire, swtiches etc. And, yes, +0.8V makes a big difference. \$\endgroup\$ – Fredled Jan 5 '18 at 0:42
  • 5
    \$\begingroup\$ Most LEDs are best driven by a constant current supply, not by constant voltage. Can you link to a datasheet for your LEDs? \$\endgroup\$ – Peter Bennett Jan 5 '18 at 0:52
  • 2
    \$\begingroup\$ "Specifications from the manufacturer is not clear" - OK, I'll bite, what specifications? Why not include them in your question? Don't you think it would be helpful to those of us trying to understand your problem? Provide a link, for Heaven's sake. \$\endgroup\$ – WhatRoughBeast Jan 5 '18 at 13:57

Ah. So you're driving LEDs directly with voltage, and the units you've got work best at 24.8 volts. Congratulations - you're about to kill your LEDs. The reason that the manufacturer's specs are not clear is because the operating voltages for different LEDs will vary from unit to unit.

Now, about killing your LEDs. Yes, driving LEDs from a voltage source, particularly at high power, has a very high chance of killing them. The reason is that the forward voltage (which you have measured) will decrease with increasing temperature. This will cause a voltage supply to provide more current at the same voltage, which will dissipate more power, which will cause the LEDs to get hotter, which will cause them to draw more power, etc until one of three things happens.

1) The power supply goes into current limit. This will provide a maximum power level, beyond which the supply simply will not go.

2) The LED manages to come into thermal equilibrium. While possible, this is not something you should count on. If nothing else, the equilibrium point is likely to change as the ambient temperature changes.

3) The system enters into what is called "thermal runaway", where the power dissipated increases until the LEDs fail.

Note that option 1 is equivalent to specifying a constant current drive, rather than constant voltage, and Peter Bennett's comment is exactly correct.

So the answer to the question you should have asked, "Is it a good idea to try to drive my LEDs with a voltage supply of 24.8 volts?" is, "No. Don't do it."

And the answer to the question you did ask is, "Nope. You need a DC-DC converter which will have a very small input/output difference. This will require a switching power supply with all of the stuff you don't want to do."

  • 3
    \$\begingroup\$ @Fredled - biting the hand that feeds is seldom productive longer term. IF your "LEDs" have associated CC regulators then your question should include that 'fact', and anything else that is relevant that you know that you have not yet told us. The "actual question by 1000 cuts" method is liable to produce disatisfaction all round. | If you can provide part numbers and/or sources it is liable to be useful. There are experts 'hovering in the wings' who may assist if the question seems liable to be completely put and stable, and not liable to result in a rebuke in exchange for assistance. \$\endgroup\$ – Russell McMahon Jan 5 '18 at 10:48
  • 3
    \$\begingroup\$ @Fredled If the LEDs have a constant current regulator, why is the supply voltage so critical to the brightness? \$\endgroup\$ – Simon B Jan 5 '18 at 11:04
  • 3
    \$\begingroup\$ @Fredled - No. They are not connected to a constant current source. If they were, input voltage would not matter as long as it was above some minimum voltage. Please provide a link to the data sheet which you are misinterpreting. \$\endgroup\$ – WhatRoughBeast Jan 5 '18 at 13:59
  • 3
    \$\begingroup\$ @Fredled - For the second time, please provide a link to the data sheet. Stop hoarding information and then criticising the ignorance which results. For instance, it appears that you have a CCR, which you did not mention earlier. Provide a complete, detailed description (use the schematic entry tool) and stop jerking us around. \$\endgroup\$ – WhatRoughBeast Jan 5 '18 at 16:48
  • 1
    \$\begingroup\$ @Fredled, you say "Current flowing through the LEDs depends on voltage. They should take 350mA each." While current through a LED does indeed depend on applied voltage, you got this backwards. A string of your LEDs should be fed with 350mA, not that "they should take it". The resulting "forward voltage" will vary depending on LEDs temperature. When LED heats up, its FV drops (effective resistance drops). If you continue to feed it with constant voltage, the current will exceed LED's specifications, rapidly, and they will likely burn out. Controlling voltage will not fix this problem. \$\endgroup\$ – Ale..chenski Jan 5 '18 at 18:18

The range given by manufacturer (23.8 - 28 V) is the range of forward voltages that will appear on the string of LEDs if a constant 10 A current is applied. This voltage will depend on temperature and manufacturing differences. Using a power supply with fixed voltage (and unlimited current) is a sure thing to kill your LEDs, because at different temperature the forward voltage might go down, and current will go over the roof, and the string will overheat and burn out.

Actually, all you need is a special class of power supplies called "LED driver". Example: MeanWell lineup, model HLG-320H-30 will fit. This is almost a regular power supply, but with constant current limit. If the load draws the voltage below the supply set point (30 V in the example), the supply will turn into constant-current (10 A mode), and drop the output voltage to whatever it takes for LEDs to draw 10 A. It could be 24 V, or 25 V, or 28 V, and it will vary with LED string temperature (whichever thermal management is designed into it and ambient air conditions). These "LED drivers" have somewhat less quality in terms of line regulation/ripple etc, and therefore usually are 30-50% cheaper than the similar full-blown power supply modules.

  • \$\begingroup\$ The LEDs are not using 10A. 10A is the maximum current for the entire appliance which connects several circuits together. Each led is driven by a DC-DC constant current regulator and should draw 350 mA each. The regulator can only reduce the voltage not increase it. So, if the voltage should be 24.5V to reach 350 mA, it needs more, 25V for example. If I apply 28V it's too much because the current regulator takes time to adjust and could even not be able to adjust properly, and the LEDs heats quickly. According to experiments I ran, 24.8V is the best voltage to reach the expected effect. \$\endgroup\$ – Fredled Jan 5 '18 at 7:23
  • 3
    \$\begingroup\$ So you didn't formulate your question correctly, then. The question is in in fact not how to boost voltage for driving LEDs, it's about driving a number of circuits with a LED in them. Which changes things up. If you post what type of LED and driver you are using, it would be significantly easier for us to give you an answer that you might find On-topic. \$\endgroup\$ – MrGerber Jan 5 '18 at 8:27
  • \$\begingroup\$ MrGerber: I wrote "LED appliances" ;) \$\endgroup\$ – Fredled Jan 5 '18 at 15:47
  • 1
    \$\begingroup\$ @Fredled, if you have a bunch of individual 350-mA LED strings, each with a dedicated and appropriate DC-CC converter, then you should be able to apply ANY voltage in excess of 25-28 V, and everything should be fine. If you apply 28V and something heats up, then your "DC-CC" converters are designed incorrectly, especially if you mention something like "takes time to adjust" and "not be able to adjust properly". What time? how long? Did you design these DC-CC converters? \$\endgroup\$ – Ale..chenski Jan 5 '18 at 16:56
  • 1
    \$\begingroup\$ @Fredled, you didn't get it. There is no "typical voltage", the strip voltage will vary from 23 to 28 V depending on the LED's chip temperature, and the temperature will go up, from ambient cold to 100-120 deg.C after turning on. \$\endgroup\$ – Ale..chenski Jan 5 '18 at 22:36

If you don't WANT 24.0 volts DC, and instead want 24.8 volts, adjust the power supply. The likely ways to do this include changing a resistor connected to a TL431 adjustable zener, or tweaking a trimpot, or offsetting the V+sense and V-sense pins on a regulator that has sense pins.

Be prepared for unintended consequences. Brightness of a regulated LED might depend a lot on temperature and age, but (because of the regulator) is unlikely to have a uniformly favorable response to minor voltage tweaks. The usual tolerance on power supplies is 5% or 10% (2% for a 'laboratory' supply), and this, circa three percent change, is not guaranteed accurate over any extended time.

  • \$\begingroup\$ Yes, I was thinking about reverse engineering power supplies but it's a lot of work (ok for one or two, not for 100) and I don't want to open the case neither. If it's waterproof, forget it. Not even talking about invaliding the guarantee. The best would be telling the manufacturer, but then you should order big quantities. Not my budget. But thanks for your reply. \$\endgroup\$ – Fredled Jan 5 '18 at 15:53
  • \$\begingroup\$ @Fredled: there's no problem 'invalidating' if you get a supply with an adjust knob, like <amazon.com/MEAN-WELL-S-150-24-Supply-Single/dp/B005T6J4BQ/…> \$\endgroup\$ – Whit3rd Jan 6 '18 at 2:42

Many fixed output power supplies have an output adjustment range of +/-10% or so, which would be more than enough to get you an additional 0.8 V. Separate from that, a switching boost regulator can be a relatively simple circuit with just an inductor (as opposed to a transformer). Linear Tech has the LT1070 and 1270 series of parts. These are in a 5-pin TO-220 package and have the power transistor built in. Somewhere in the series (memory fade) is one with a 10 A switching transistor. Run at it's limit, that would get you about 9 A of output current.

  • \$\begingroup\$ An adjustable power supply is the solution. Unfortunately many types of supply are not adjustable. \$\endgroup\$ – Fredled Jan 5 '18 at 7:32
  • 4
    \$\begingroup\$ Fortunately, many are. \$\endgroup\$ – AnalogKid Jan 5 '18 at 12:19
  • \$\begingroup\$ Yes, I'm redesigning some appliances to fit with adj types. But it's not always possible. I had a look on switching boost regulators, with all components, pcb and stuffs it will be too expensive. The cost of two power supplies, practicaly. \$\endgroup\$ – Fredled Jan 5 '18 at 15:57

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