Author Topic: diy 300W LED supply?  (Read 2482 times)

0 Members and 3 Guests are viewing this topic.

Offline ETITsynthesizerTopic starter

  • Contributor
  • Posts: 48
diy 300W LED supply?
« on: November 18, 2019, 07:07:05 am »


R1 C1 make a step down from 170v to 9v. I figured that a full wave rectifier would give me 120Hz. C1 is also the power supply smoothing cap. then there is a 5v zener for the 5v ref and analog comparator IC supply. the pot is the 0v to 5v reference. the output of the SMPS feeds back to the analog comparator. I have the LM311 collector output switching the 2N6213. I designed it to get 1.5mA into the base of the 2N6213. all the caps and resistors are rated for where they are. I'm not sure if the LM311 is ok because the datasheet has 16mA at 30v. I have 1.5mA at 170v.

my load is 3 100W LED's in series screwed to a common heat sink. still not sure if the LED's have some grounded chassis or what. in the planning stages now. I will test with a dummy load before I connect the LED's.

I think it would be safer to not earth this circuit since it has two mains wires and no transformer isolation. maybe I need a fuse or two?
« Last Edit: November 19, 2019, 04:41:37 am by ETITsynthesizer »
 

Offline Berni

  • Super Contributor
  • ***
  • Posts: 5050
  • Country: si
Re: will this work?
« Reply #1 on: November 18, 2019, 07:26:02 am »
To me it looks like it at least won't explode, but it won't really be doing much in the way of regulation because that transistor will be permanently turned on by a small current into the opamp.
 

Offline ggchab

  • Frequent Contributor
  • **
  • Posts: 283
  • Country: be
Re: will this work?
« Reply #2 on: November 18, 2019, 07:28:00 am »
I am not an expert but isn't this circuit a bit dangerous for you (direct connection to the mains) and for the LED's (no current control) ?
 

Offline ETITsynthesizerTopic starter

  • Contributor
  • Posts: 48
Re: will this work?
« Reply #3 on: November 18, 2019, 03:18:18 pm »
To me it looks like it at least won't explode, but it won't really be doing much in the way of regulation because that transistor will be permanently turned on by a small current into the opamp.

thanks for your help. I think I fixed it now. let me know if there is anything else. obviously this is an unsafe product. I will probably just build it to test and then build the real one with the transformer later.


mouser cart is $16.98 USD for two transistors, a comparator and a zener. most of that is in the transistor for $14.23.
 

Offline Berni

  • Super Contributor
  • ***
  • Posts: 5050
  • Country: si
Re: will this work?
« Reply #4 on: November 18, 2019, 03:34:33 pm »
That is better

You might still want to do some math on your resistor values to make sure you got enough power to turn on the power transistor. Might make sense to add a darlington pair there depending on how things work out. You might also want some direct negative feedback from the opamps output since these sort of inherently unstable pass stages love to oscillate.

Also its a bit confusing that you are trying to drive LEDs with a constant voltage output. Usually it makes more sense to drive LEDs using constant current. You would also want your main storage capacitor to be before the pass transistor for best output stability, unless the reason for putting it there is to get a better power factor.

If you don't know what you are doing its best not to mess with mains since when things go wrong they can go really wrong. The thing would most definitely require a fuse and a grounded case.
 

Offline ETITsynthesizerTopic starter

  • Contributor
  • Posts: 48
Re: will this work?
« Reply #5 on: November 18, 2019, 04:09:00 pm »
not sure exactly what you mean on the darlington but I did another drawing anyway. I checked the math again. min Hfe is 10 so 3A at the collector = 300mA at the base. then I take a SC2383 that has a min gain of 60 and work backwards from 300mA. so I need 5mA to the base of the SC2383. I did something like a darlington there. the collectors are on different supplies. maybe there is a way to do it cheaper with only 2 transistors? will this work like this?

 

Offline Berni

  • Super Contributor
  • ***
  • Posts: 5050
  • Country: si
Re: will this work?
« Reply #6 on: November 18, 2019, 04:17:18 pm »
In this case you might find that the 2SC2383 might get pretty toasty since it will be passing all that base current with a lot of voltage across it.

I meant having a darlington configuration up at the high side where you can sacrifice a bit of extra dropout voltage for having an easy way to push a lot of current into the base of the power transistor.

A simulator like LTSpice is a good tool for designing analog circuits like this since it lets you experiment with it in quick iterations and check how hot something might be getting before you even build it.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: will this work?
« Reply #7 on: November 18, 2019, 06:21:16 pm »
That is better

Add an incessant eye twitch, and that's perfectly said... >:D

ETITsynthesizer: the central problem is, you're trying to make a linear power supply, it seems; so in general you need to drop somewhere on the order of 300W (some fraction thereof, that is), and a single transistor ain't gonna do it.  Moreover, you've drawn the circuit like it's digital -- huge, wide open gains in the transistors, and also seemingly drawn it in current mode -- that is, assuming the transistors are only current amplifiers, Ic = hFE * Ib, but reality is more subtle than this (even if you're just using them for switching, but especially if you're using them for linear operation).

I would strongly recommend buying an LED power supply and studying its design; you can recreate its function at low voltages, where it is safe to touch and explosions are unlikely.  Once you've perfected your knowledge, scale up and build this.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline Gyro

  • Super Contributor
  • ***
  • Posts: 10173
  • Country: gb
Re: will this work?
« Reply #8 on: November 18, 2019, 08:08:03 pm »
Please edit the thread title in your original post to something meaningful. If everybody just put 'will this work?' then the forum would become unusable.
Best Regards, Chris
 
The following users thanked this post: tooki, george.b

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5170
  • Country: ro
  • .
Re: will this work?
« Reply #9 on: November 18, 2019, 09:00:05 pm »
You could use a LR8 regulator to get the low voltage you need, with all the protections and safeties: http://ww1.microchip.com/downloads/en/DeviceDoc/20005399B.pdf

As for everything else, does it really have to be 100v 3A, or could you maybe split it in multiple channels or something like that?
 

Offline ETITsynthesizerTopic starter

  • Contributor
  • Posts: 48
Re: will this work?
« Reply #10 on: November 18, 2019, 11:15:59 pm »
cree xlamp. 37vF. 100W CoB LED. 3 of these in series so that they run current matched. this should simplify the supply. I started working on a SMD current regulated supply but it is not ready to post.
 

Online ejeffrey

  • Super Contributor
  • ***
  • Posts: 4033
  • Country: us
Re: will this work?
« Reply #11 on: November 18, 2019, 11:47:18 pm »
PLEASE EDIT THE THREAD TITLE TO SOMETHING MEANINGFUL
 

Offline rvalente

  • Frequent Contributor
  • **
  • Posts: 736
  • Country: br
Re: will this work?
« Reply #12 on: November 19, 2019, 12:04:08 am »
There are plenty of buck designs from various vendors, why making your life harder??
 
The following users thanked this post: Someone

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5170
  • Country: ro
  • .
Re: will this work?
« Reply #13 on: November 19, 2019, 12:25:12 am »
If you don't necessarily need a linear driver, you could just use a led driver IC for each led.

For example, have a look at MP24833-A : https://www.monolithicpower.com/en/mp24833-a.html
Datasheet: https://www.monolithicpower.com/en/documentview/productdocument/index/version/2/document_type/Datasheet/lang/en/sku/MP24833-A/document_id/2070/

It's 1.12$ each and the rest of the components you'd need probably don't cost more than 2$

You can set the brightness between 0% and 100% by giving 0.67v...1.35v on the enable pin (a plain potentiometer will do) or you can pwm the value (100hz to 2 khz)

The catch is this chip only works with up to 55v, so you'd have to step down 110v AC to lower value... you'd probably be best with a standard 48v power supply, they cost as little as 40$, here's an example: https://www.digikey.com/product-detail/en/mean-well-usa-inc/LRS-350-48/1866-3350-ND/7705033
 

Offline rvalente

  • Frequent Contributor
  • **
  • Posts: 736
  • Country: br
Re: diy 300W LED supply?
« Reply #14 on: November 19, 2019, 11:22:00 am »
There is no sense making a linear driver when you can make an smps.

This IC is very easy to setup and with external mosfets and built in current protection, you can go crazy power.


https://www.infineon.com/dgdl/irs25401pbf.pdf?fileId=5546d462533600a40153567b06662826
 

Offline mikerj

  • Super Contributor
  • ***
  • Posts: 3382
  • Country: gb
Re: diy 300W LED supply?
« Reply #15 on: November 19, 2019, 12:31:07 pm »
Ignoring the problems with isolation, power dissipation and excessive loop gain for the moment, an LED should be driven with a constant current rather than the constant voltage which this circuit provides.
 

Offline rvalente

  • Frequent Contributor
  • **
  • Posts: 736
  • Country: br
Re: diy 300W LED supply?
« Reply #16 on: November 19, 2019, 01:08:36 pm »
Isolation was never a problem to lighting products. HID, fluorescent and even halogens were never isolated (except 12v halogen).

 

Offline rvalente

  • Frequent Contributor
  • **
  • Posts: 736
  • Country: br
Re: diy 300W LED supply?
« Reply #17 on: November 19, 2019, 01:10:31 pm »
 You can replace the m2 mosfet with a fast diode, eliminating the sync rectifier and lowering the cost. I believe any 10A 500v mosfet will work here, very common and cheap
 

Offline mikerj

  • Super Contributor
  • ***
  • Posts: 3382
  • Country: gb
Re: diy 300W LED supply?
« Reply #18 on: November 21, 2019, 12:41:15 pm »
Isolation was never a problem to lighting products. HID, fluorescent and even halogens were never isolated (except 12v halogen).

Isolation can definitely be a problem in lighting products if they are either not grounded or double insulated.  Commerical drivers for high power LEDs are most always isolated.
 

Offline dmendesf

  • Frequent Contributor
  • **
  • Posts: 342
  • Country: br
Re: diy 300W LED supply?
« Reply #19 on: November 21, 2019, 04:57:45 pm »
Even when the datasheet says that the forward voltage is 37V it's not safe to drive a constant voltage over such a led. This is just an indication about the needed voltage range for your constant current supply. If you drive a constant 37V the device will get hot, so for the same voltage it will demand a higher current (as the temperature coefficient is negative) an will probably enter a thermal runaway until it burns.
 

Offline rvalente

  • Frequent Contributor
  • **
  • Posts: 736
  • Country: br
Re: diy 300W LED supply?
« Reply #20 on: November 22, 2019, 01:53:56 pm »
Isolation was never a problem to lighting products. HID, fluorescent and even halogens were never isolated (except 12v halogen).

Isolation can definitely be a problem in lighting products if they are either not grounded or double insulated.  Commerical drivers for high power LEDs are most always isolated.

For a 300W led driver, even if its isolated and in a plastic housing I would still ground it, just like many notebook power supplies do. 
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf