High curent opamp driver buffer questions

Thread Starter

craigtharpe

Joined Jan 11, 2017
6
I am looking for help with this circuit. I need to build 35 of an LED dimmer circuit. I am trying to utilize the circuit in this link
https://www.allaboutcircuits.com/te...r-an-op-amp-output-for-higher-current-part-1/.
I used a LMC6482, rail to rail op amp and a TIP120 Darlington transistor as substitutes in the linked circuit. The op amp and transistor share the same 12 volt DC, 2.1 amp, Meanwell power supply. The LED array draws 1.61 amps at 9.8 volts when driven directly from a variable bench power supply. The op amp input is being driven from a 0-10 volt signal. The 0-10 volt signal originates from a remote 0-10 volt dimmer driver. The output tracks the input as expected so in general the circuit works great, except the TIP120 gets extremely hot. I have added the largest heat sink that I can fit to the TIP120 and still cannot control the heat. I have checked the current draw for the circuit while operating and it never exceeds 1.75 amps. What am I missing? Shouldn't the TIP120 suffice for this application? I have tried substituting a MJH6284 Darlington rated at 20 amps, but get the same results. Thanks in advance for any guidance!
 

danadak

Joined Mar 10, 2018
4,057
1.61 amps at 9.8
So The TIP120 is dissipating (12-9.8) x 1.61A = 3.5W

The absolute max Pcollector = 2 W from datasheet. Ta = 25C. 65W if Tc = 25C.

The Tj-c = 1.92 C/W. Tj-a = 62.5 C/W

What is Thermal R for the heat sink ? If its just a simple clip on
heat sink probably explains the sink and TIP120 being hot.

3.5W w/no or tiny sink will definitely cook eggs.

http://www.ti.com/lit/an/slva462/slva462.pdf


Regards, Dana.
 

Thread Starter

craigtharpe

Joined Jan 11, 2017
6
I appreciate the calculations and that link Dana. I started with a tiny clip-on and as you said it would easily "cook eggs." I have been able to get it to stabilize at about 120C with a cobbled up heat sink. I have no idea what the thermal resistance is. I will dig into the linked document and hopefully educate myself a little. I would like to get the temperature down a little more for a little headroom in adverse conditions. Looking at the thermal/power derating on the TIP120 datasheet, it looks like she'll stand about 10-12 watts at 120C. I think I'm going to work on the heat sink optimization some more and go with it. It still feels too hot for comfort.........

Is the 2W rating at 150C Tc and 25C Ta?
 

Papabravo

Joined Feb 24, 2006
21,157
There are three ways in which head is removed from a device. Radiation, conduction, and convection. If the heat sink you have is not big enough, then try a fan blowing on the device or the fins of the heat sink. If that doesn't work then black anodize the aluminum for better radiation. How about modifying the heat sink to have cold liquid piped through it. You get the idea.
 

Thread Starter

craigtharpe

Joined Jan 11, 2017
6
There are three ways in which head is removed from a device. Radiation, conduction, and convection. If the heat sink you have is not big enough, then try a fan blowing on the device or the fins of the heat sink. If that doesn't work then black anodize the aluminum for better radiation. How about modifying the heat sink to have cold liquid piped through it. You get the idea.

Yep. I get the idea, but as Murphy would have it, all this has to fit into a ceiling cavity with limited space and limited make-up air.
 

Thread Starter

craigtharpe

Joined Jan 11, 2017
6
That's why linear dimmers are very rare indeed- the heat sink costs more than all the other parts combined.
That's why linear dimmers are very rare indeed- the heat sink costs more than all the other parts combined.
I believe that! The most surprising thing of all this project has been how little the manufacturers and distributors of LED fixtures know about their products. No one's sure if it's dim-able and if they think it is they are not positive whether it's PWM, CC, CV, maybe 0-10 volt who knows?!?! I think we've finally got a solution and I really appreciate all the help from you guys.
 

Sensacell

Joined Jun 19, 2012
3,432
I believe that! The most surprising thing of all this project has been how little the manufacturers and distributors of LED fixtures know about their products. No one's sure if it's dim-able and if they think it is they are not positive whether it's PWM, CC, CV, maybe 0-10 volt who knows?!?! I think we've finally got a solution and I really appreciate all the help from you guys.
Haha- you will find that people that sell things generally know very little about them, it's not necessary 98% of the time.
 

Thread Starter

craigtharpe

Joined Jan 11, 2017
6
Just out of interest, why did you go with linear control and not PWM?
The fixtures go crazy (flickering and strobing) at at least 2 points in the 0-100% scale. The vendor said "you might have to program around those spots"..........the customer said "I don't think so." That's what I was referring to above. These were sold as PWM dimmable fixtures. I would swap everything out and send the fixtures back, but the holes are cut and everything installed before the problem was discovered and the deadline looms near!
 

dendad

Joined Feb 20, 2016
4,451
You could use a switch mode regulator with filtered output in place of your linear regulator so the lamps still run on DC, but you get rid of the heat from your linear reg.
Or a switch mode pre-regulator to your linear one so the voltage drop across the pass transistor is kept to minimum levels.
Just PWMing the lamps directly is probably beating with their internal PWM if they have them.
What are the lamps? Otherwise, the beating may be a fault with the supply you were using.
 

MrAl

Joined Jun 17, 2014
11,389
Hello there,

PWM would be the way to go, but you do have to be a little careful with DIRECT PWM because there are a few pitfalls. The first is that some LED's are not supposed to be PWM'ed. Check data sheet for info on that and then go to a full buck instead rather than raw PWM if needed.
With a buck circuit you might be able to get away with no heatsink if the buck is high efficiency.

If you are stuck with linear, then you are stuck with a heat sink, probably large.
The power it needs to dissipate is the max power of the circuit element being mounted to the heat sink. The max power is found from multiplying current times voltage during operation where the product is maximum.
If you have a power source of 12v and current max of 2 amps, then the max power is 24 watts, but because the current usually comes down as the voltage across the device comes down and when the voltage is 12v the current is zero amps, that means the max power might only be around 6 watts and could actually be less with a non linear device (like an LED). Given the 6 watts you need a heat sink with at least 6 square inches surface area to achieve a temperature rise of 60 degrees C over ambient, but that 6 sq inches is usually best in a well designed heat sink rather than flat plate because of the temperature distribution that comes from heat conduction and heat sink shape. So what you normally do is find a heat sink that has a rating that fits the job. Since you have maybe 6 watts to dissipate and you want a max temperature rise of 60 degrees (not a bad idea for various reasons) then you need a heat sink that has a max rating of 10 degrees C per watt. You can get better if you want to end up with a lower temperature.

You also need to consider the spatial confinement of the device too though. If it is to be mounted inside a box or wall you need to use a better heat sink and possibly drill air holes. If you dont drill air holes the temperature could build up inside the enclosure. This also means careful testing is a must.
Temperature improvement by way of a fan can be deceiving. The fan brings the temperature down a lot provided it has free air flow, but fans are mechanical and so fail and thus no longer bring the temperature down. This means it is wise to add to the design a way to detect that the fan has stopped working and shut down.
 

be80be

Joined Jul 5, 2008
2,072
There is always better way all the led lighting we been putting in use current source and can be lowered for dimming.
There made for hot places where there is no air and there potted with that brown rubber like stuff to keep water out.

I just moved 4 lights the temp in the place the power supply's are at get's hot It hit 140F by noon time to get out of there LOL.
But these things use PWM for dimming and there RGB for the cool light shows while your getting your car washed the leds inside the car wash
the power supply's inside the ceiling above.
4783931.jpg
 

ebp

Joined Feb 8, 2018
2,332
If the fixtures flicker and strobe when some separate PWM circuit is added, it is a very strong indication that the fixtures already contains active electronics that is doing some sort of PWM. No configuration of LEDs, with or without series resistors, would do this. This would also suggest that attempts at dimming by altering the supply voltage are dubious.

Will the fixtures operate on 12 volts AC?

My suspicion is that the LEDs have a high-frequency switch mode driver to provide constant current. These have become extremely common. Suitably designed drivers of this sort will work if the input is low frequency PWM'd - typically what you would get from phase-angle control of AC mains frequency, so the maximum PWM frequency probably should not exceed 150 Hz or so. Some of the newer drivers will do a very good job of making the input current proportional to the instantaneous input voltage (high power factor - at least when there is current flowing).
 

ronsoy2

Joined Sep 25, 2013
71
I believe that! The most surprising thing of all this project has been how little the manufacturers and distributors of LED fixtures know about their products. No one's sure if it's dim-able and if they think it is they are not positive whether it's PWM, CC, CV, maybe 0-10 volt who knows?!?! I think we've finally got a solution and I really appreciate all the help from you guys.
Try one of these. ebay# 142899028282. I suspect you can fool with the pot connections to get a variable voltage to vary the output. You will not be able to build these for less than you can buy them. Try removing the pot and connecting your varying voltage to the center pin through a resistor. I think you will find a resistor value that will allow your control voltage work to vary the output. Be sure to connect common to the correct place, probably the negative lead of the power supply.
 

MisterBill2

Joined Jan 23, 2018
18,167
I just popped in on this thread, and it must be that the fixtures have internal electronics. That should have been clear from the specifications. A direct connection to the LED devices would allow using an external PWM controller easily.
BUT here is another idea, which is examine one of the fixtures and trace out the circuit of the internal electronics. Probably it will be possible to use those for the controlling and the interface will just be some isolation of some kind. Why invent a controller circuit when it is obvious that you already have one on hand? AND it is already paid for, I think.
 
Top