introduction -- current sources -- led lighting

[superfrog]

Hello,

I just joined the forum after quite a bit of looking around for a place to look for design ideas and advice.
I am a design engineer in UK, working mainly on high level simulation.

I have covered during my studies and work most of the parts of digital circuit design, from high level tool design to transistor level design. I even designed a silicon op amp, with quite a bit of guidance.

Sadly, I skipped all the most of the discrete analog bits, and am close to useless in anything that does not relate to RC networks and CMOS logic.

In the next post, I ll briefly explain my goals and ask for comment on a single transistor current regulator for the purpose of early testing of led modules I plan on using.

 

[superfrog]

My goal is to design a led lighting system for plants

It would be a mainly dichromic system in 660nm/455nm.

blue can be done with high power leds, luxeon K2 as these ones for example
Lumileds Lighting | Optoelectronics and Displays | Optoelectronics | LED | LUMILED LED Components |LXK2-PR14-Q00
I ll worry about these ones later on

660nm (far red) has to be generated from 20mA leds, as far as I can research.

My ultimate goal in this project is to get some SMPS design, to handle the load (probably chains of 200 of those leds, at 450V, 20mA), but before that I need to go trough a bunch of steps.

I got to the point where I can generate board designs good enough for non critical designs (like a buch of leds in series), and decent power circuits. I am getting the pcb etching going slowly, even though HCl is proving to be quite hard to get.

I wish to design a wide voltage range linear current regulator for testing modules.

I am thinking about recycling some of the power BJT or mosfets I salvaged from PC power supplies ( I got a nearly unlimited suppy of those from work).

I ll find out how to attach images and I ll attach the design for discussion

 

[Sceadwian]

High level simulation of what? What systems/dynamics, or just general? We'll await your goals and circuit for comment for the current regulator.
Just on a personal note it's very nice to see such a well structured post =)

 

[superfrog]

I do high level simulations of video decoding cores.

I want to design a dichromic led lighting system for plants.

blue can be done with high power leds, I ll worry about those later.
660nm (far red) has to be generated from 20mA leds.

My ultimate goal in this project is to get some SMPS design, to handle the load (probably chains of 200 of those leds, at 450V, 20mA), but before that I need to go trough a bunch of steps.

I got to the point where I can generate board designs good enough for non critical designs (like a buch of leds in series), and half decent power circuits. I am getting stuff for etching slowly.

I wish to design a wide voltage range linear current regulator to test individual led modules to start with. I'd like to use some of the power bjt or mosfets I salvaged from PC power supplies (which I can get for free for personal use from work)

**broken link removed**

Are there any caveats with these types of currrent regulators ?
Current spikes on start, poor regulation?

I was thinking about using voltage regulators with a resistance in series and the adjust pin tied after the resistor, but they do not seem to come with breakdown voltages higher than 125V.

Considering that the transistors come for free, I'd be quite interested to get some use out of them.

As I understand, the gain for the BJT I got is about 10, Vbe=7V

As such, Ibe should boe 2mA for 20mA load current intensity,

With a supply voltage of 200-300V, and a 20V zener, R1 should let 4mA go trough, so with a voltage over the resistor of 180-280, we are talking 45k resistor max. This would need to dissipate 2.37W at 280V. Not impossible, but annoying.

The transistor would have to dissipate 6W which is perfectly in spec, and I have large aluminium heatsinks that should take the abuse without problem.

If I go with the mosfet, I suspect I will be able to greatly increase R1 as gate current should be mostly for charging the gate and compensating for gate leakage, which hopefully is not much on these big ones. I need to research this, but I suspect I should be able to get away with 500K, which would get power under the cheap hobbyist .25W limit.

Any comments?
Am I talking rubbish?
All inputs appreciated

Thanks.

Ps: How do you guys attach pictures to your posts, free web image hosting is somehow suboptimal.

 

[audioguru]

For your transistor to dissipate 6W when its current is only 20mA then you are wasting 300V across it. The 300V should be across the load of LEDs, not across the transistor. If the transistor is conducting then maybe it has only 10V across it and dissipates only 0.2W.

 

[superfrog]

Well, the worse is worth considering

I'd be quite happy to maintain regulation on the whole 0-280 V (that is considering Ron losses and resistor forward voltage and a 300 V input) range forward voltage for the load.
My led strings will be segmented so I might be testing 80-100 V modules at a time. I'd like a single circuit to deal with all possible cases.

Thanks for answering nonetheless.
I'll have a look at the topic tomorrow, now is way past bed time.
Cheers.

 

[superfrog]

I am considering the MOSFET option to be able to avoid the relatively high loss trough R1 in the BJT based limiter, but while my source Wikipedia) seem to imply that all the given circuits can be made with mosfets, I fail to understand how this can work.

The BJT cicuit is based on the fact that Vbe is constant (or at least relatively constant with a temperature related drift), surely in a n-mosfet Vgs is not (or I would not be writing on a computer right now, as CMOS would not work, I have the feeling - that said I might be wrong, and I'd love to be proven wrong).

I was having a look at a design with a opamp comparator to drive the mosfet gate, which seems to make some kind of sense. Mosfet can be used as voltage controlled resistors. Sadly, opamp needs power supply, I am not sure it is all worth the bother.

After going over pretty much every post in the forum relating to constant current sources, I found https://www.electro-tech-online.com/threads/high-power-led-controller.95832/ which has some interesting design ideas. It is pretty much the same as the op amp idea, but more integrated.

I slightly extended the idea to work over the range I am interested in, ie rectified AC, in the following schematic. Is the design correct? The zener voltage could be reduced to accommodate maximum gate voltage, R8 would be used as sense resistor to create drop Vref at 20mA in my case but I guess R = Vref / I in the general case. At 20mA with Vref being 2.5V, Heat losses in the sense resistor are .05W, which is acceptable.

I guess R6 and R7 need dimensioned, and I probably need advice on that, as I believe they are actually the critical bit of the design, in term of responsiveness and stability.

**broken link removed**

Would that work?

Thank you for your answers.