Help please designing LED alternative lighting?

[enricosavazzi]

Like many semiconductor components, LEDs exhibit a non-linear current-to-voltage ratio (i.e. they do not follow Ohm's law). They are also vulnerable to thermal runaway: fed with a constant voltage, the current increases with temperature, and beyond a certain threshold the LED keeps passing a higher and higher current and warming up more and more, until it self-destructs.

This is one of the reasons why it is better to feed a LED from a current-stabilized power supply (which limits the current to a given threshold value by changing the output voltage accordingly), rather than a voltage-stabilized one (which, in an oversimplified case, tries to keep the output voltage constant by changing the current, and does not care how much current it outputs as long as the output voltage is constant). If you don't exceed the current specification of the LED and provide adequate cooling of the LED, thermal runaway is prevented with a current-stabilized supply.

The max current provided by a current-stabilized power supply in general can be set with a trimmer on the power supply, so you do not need a power supply specifically tailored to a given LED, and can choose a beefier power supply than strictly necessary, to avoid the risk of overheating and shutdown. A separate trimmer to set the minimum current value can also be handy.

A current-stabilized power supply that allows you to change the threshold current between the set minimum and a set maximum with e.g. a potentiometer on a control panel also works as a variable dimmer. A variable-voltage power supply is also a poorer choice as a LED dimmer because of the LED's non-linear voltage-to-current ratio.

[dolmadis]

Thank you Enrico for this clear and informative explanation.

BR

John

[dolmadis]

A current-stabilized power supply that allows you to change the threshold current between the set minimum and a set maximum with e.g. a potentiometer on a control panel also works as a variable dimmer. A variable-voltage power supply is also a poorer choice as a LED dimmer because of the LED's non-linear voltage-to-current ratio.

Is this a "A current-stabilized power supply" please?

http://www.ebay.co.uk/itm/122228964289

As always thanks for the help.

BR

John

[enricosavazzi]

A current-stabilized power supply that allows you to change the threshold current between the set minimum and a set maximum with e.g. a potentiometer on a control panel also works as a variable dimmer. A variable-voltage power supply is also a poorer choice as a LED dimmer because of the LED's non-linear voltage-to-current ratio.

Is this a "A current-stabilized power supply" please?

http://www.ebay.co.uk/itm/122228964289

As always thanks for the help.

BR

John

Yes, according to the specifications you can set both the maximum voltage threshold and the maximum current threshold.

The specified ripple of 100 mV is however a bit of a concern, too high for my liking in a bench power supply. This is not necessarily a show stopper because you would be using it as a stabilized current supply. They do not specify the current ripple, however, which is an even bigger concern. A significant current ripple will cause the LED to flicker.

They do not specify the ripple frequency, either. This seems to be a DC to DC converter, so "ripple" in this context does not refer to the mains 50 Hz or 60 Hz, but to something else, for example a closed-loop ringing or the oscillator frequency of the converter.

[dolmadis]

Thank you Enrico. That is very helpful. I think that I will pass on this because I want to avoid flicker.

But I am floundering when I look for a product because I keep finding ripple of at least that and often more.

I wonder if there is a Forum Member in the UK who could recommend a suitably specified but economically priced Switching Bench Power Supply which provides variable output DC voltage from 0-30V at a variable 0-5A?

BR


John

[dolmadis]

Specification of alternative switched bench supply

Input voltage:AC 110V/220V 50/60Hz
Output voltage:0 ~ 30V
Output current:0 ~ 5A Voltage resolution:100mV
Current resolution:10mA
Power effect:CV=1%+10mV
Effect of load:CV=1%+10mV
Ripple and noise:Vp-p=1%
Voltage display precision:±1%+1digits
Current display precision:±1%+2digits
Volume (mm):81W×165H×220L
Operating environment:(-10~45) Rh<90%

Does anyone have any thoughts on this specification please?

Thanks

John

[dolmadis]

The specified ripple of 100 mV is however a bit of a concern, too high for my liking in a bench power supply. This is not necessarily a show stopper because you would be using it as a stabilized current supply. They do not specify the current ripple, however, which is an even bigger concern. A significant current ripple will cause the LED to flicker.

They do not specify the ripple frequency, either. This seems to be a DC to DC converter, so "ripple" in this context does not refer to the mains 50 Hz or 60 Hz, but to something else, for example a closed-loop ringing or the oscillator frequency of the converter.

I have asked the seller of http://www.ebay.co.uk/itm/122228964289 to comment further on the potential for flicker from ripple.

"you can't have a low ripple on the output voltage if you are trying to control the current. It controls the current by adjusting the voltage across the led. As the temperature changes the led voltage will change."

"looks like the ripple is 100mv at 5 amps it will be less at 3 amps, the switching frequency does not matter as the output is continuous not switched, it uses an inductor and diode so that when the regulator turns off the energy in the inductor supplies the power until that drops then the regulator turns on again , you can add a resistor and cap on the output to get the ripple down"

Does anyone have any guidance on the component choice and execution of "you can add a resistor and cap on the output to get the ripple down"?

Many thanks if you can help.

BR


John

[enricosavazzi]

The second sentence in what the seller says is correct: "It controls the current by adjusting the voltage across the led. As the temperature changes the led voltage will change."

The first sentence, however, is not: "you can't have a low ripple on the output voltage if you are trying to control the current."

What the seller is probably trying to say is that the voltage will vary when the current is fixed. This is right, but has nothing to do with ripple. The power supply can be designed (at additional cost of course) to give a much lower voltage ripple, even in constant-current mode. It could also be debated whether the seller's description of how the regulator works is in fact a switched-mode regulator.

Assuming you need a 3A current through the LED, and that the frequency of the voltage ripple is well above mains frequency (which we still don't know), you could try a 0.47 Ohm resistor, which gives you a 1.41 V drop across the resistor. The resistor must be able to dissipate at least 5 W, probably better to use a 10 W resistor to avoid high temperatures, and allow air to flow around the resistor. Any largish electrolytic capacitor should do, something like 250-500 ?F (or more) 10 V (or more).

To have a lesser waste of power you can use a lower resistance (and possibly compensate for this with a larger capacitance). This, however, will cause a larger current spike when you power on the LED.

Some power supplies have a short-circuit detector that trips at a given threshold (high current and low voltage), and switches off the power. In this case, if the resistance is too small and/or the capacitance too large, when first switching on the LED the current spike caused by the capacitor charging may trip the short-circuit protection. If this happens, try using a larger resistance (which means a larger portion of the output power will be wasted in the resistor, but this is not a big problem in mains-powered LED illuminators).

[dolmadis]

Thanks Enrico. This is a very helpful explanation.

Do you think that an alternative switchable power supply with the following spec would be better and OK?

Thanks

John

Input voltage:AC 110V/220V 50/60Hz
Output voltage:0 ~ 30V
Output current:0 ~ 5A Voltage resolution:100mV
Current resolution:10mA
Power effect:CV=1%+10mV
Effect of load:CV=1%+10mV
Ripple and noise:Vp-p=1%
Voltage display precision:±1%+1digits
Current display precision:±1%+2digits
Volume (mm):81W×165H×220L
Operating environment:(-10~45) Rh<90%

[enricosavazzi]

Thanks Enrico. This is a very helpful explanation.

Do you think that an alternative switchable power supply with the following spec would be better and OK?

Thanks

John

Input voltage:AC 110V/220V 50/60Hz
Output voltage:0 ~ 30V
Output current:0 ~ 5A Voltage resolution:100mV
Current resolution:10mA
Power effect:CV=1%+10mV
Effect of load:CV=1%+10mV
Ripple and noise:Vp-p=1%
Voltage display precision:±1%+1digits
Current display precision:±1%+2digits
Volume (mm):81W×165H×220L
Operating environment:(-10~45) Rh<90%

It does look better. Is this a cased, ready-to-use bench power supply? It is often better to just buy a ready-to-use unit than buying a PC board and having to figure out how to encase it etc.

I normally use flash rather than LED for illumination. I have a cheap digitally controlled bench power supply that I have used, among other things, for LEDs, similar to eBay number 191766020330 (I just picked one at random on eBay, the same model is sold under different brands and at different prices, and this item is not necessarily the cheapest). It is not a high-precision device, but good enough for most general uses.

[dolmadis]

Once again my thanks.

I found this specification on an enthusiast electronics forum where they did not really did not want to give away their source of affordable bench power supplies (cased) on open forum. Very unlike PMN. I did not wish to enroll just to get PM status but reading to other posts I think that I have found the generic product that is badged in many names. I will now continue this research.

Your suggestion from your own experience looks similar in spec but linear which I believe is best for LED control based on further reading.

It has been a long journey for me and I must thank you very sincerely for your explanations and support.

BR

John