LED Strip lights (2 Amps) Short Fade on / Fade Off effect

[rjenkinsgb]

18 @.72

OK, thought it was $12 per board!

 

[ThomsCircuit]

OK, thought it was $12 per board!

The extra copper is $17.00
The mosfet i found slightly better price of .44 with $4.50 ship. (.70) each
I do appreciate everything. I do like saving money but I like creating these things even more.
I should have done this before but I'm going to measure the amp draw with a small meter like this. I think some of the led strips I have were not accurately represented.

 

[transconductance]

Arduino is just a PCB with a particular microcontroller chip on it. The chip will work fine without the Arduino's programmer. You just need to give it power and clock. Many microcontrollers have an internal clock. They are less precise, but for your application it doesn't matter. You do need to pay some attention to proper layout and make sure you have low impedance, small loop area, paths between the bypass capacitors and the microcontroller. Make sure voltage drop caused by the LED doesn't raise the ground voltage of the microcontroller.

Lots and lots of things have microcontrollers in them with no Arduinos in sight. Your TV remote control, electric tootbrush, toaster, toys. They are everywhere. It is normal to program the chip before soldering it to the board but, in small lots and prototypes, you usually provide a connection so you can program it on the board.

I built a very similar PCBA for a different application. But it has 2 MOSFETs controlled by the microcontroller which is all you need. There is a potentiometer which you could replace with a switch. I've attached schematic. Can't find a photo but the board is about 1.5 by 1 inches. I can give you Gerber files if you want them.

 

[ThomsCircuit]

The only place where beefy traces are needed is at the collector and emitter of the Darlington transistor, as well as the PWR input.
Emitter is probably connected to the ground plane, so that's good. But the collector track is carrying alot of current so should be wide. The PWR track from the power input (wherever that is) to the light could be a small copper plane on the top side.

Ive since moved components around. I needed the NPN on the side. I added this small pour from the collector to the lights output pin. The PWR track that went to the lights positive pin was not needed as this circuit is a negative (low) switch so it was removed. My question is this. The method i added the pour planes. I first added a full GND plane indicated by the 4 black arrows. then i drew a smaller pour plane in the top left and assigned to the collectors net. JUST TO BE SURE I DID THIS CORRECTLY. I could take it further and reduce the size of that track, add the pour, adjust the spokes (there will be only 1) from 12 to 20.

 

[ThomsCircuit]

re arranged the NPN. Added additional spokes. Somehow i got a second spoke even though it appears it is isolated.

 

[ThomsCircuit]

I built a very similar PCBA for a different application. But it has 2 MOSFETs controlled by the microcontroller which is all you need.

I do appreciate your suggestion. Im just about to submit this for fabrication. So many have suggested using Adrunio its just a matter of time before my curiosity catches up with me. Thank you.

 

[ThomsCircuit]

At 2A it will be dissipating around one and a half watts, so quite hot and may need a small heatsink.

I measured the amp draw of the LED strips using my multi meter. ive got 3 types. lowest was 1.3A. Another style was 1.8A. and the third was 2.1. Far less than the documented 4-5 amps by the seller. And that was the entire 15 foot spool. My use will be limited to 2-4 feet of leds connected to one Fader Circuit with exception of 2 being the entire reel.

Now about the heat coming from the darlington.
Ive got it set up on a bread board. I connected just one led with a 470r resistor to test and it worked. I connect a full reel rated at 1.3 amps and turned on the circuit again. No problems. no heat build up from the NPN either. I did it a few more times turning it on and off in hopes of getting some dissipation from the NPN but it stayed cool to the touch. Now if ive understood correctly the heat dissipation is only when its fading and not while it fully on or fully off. I then tried a reel that measured 1.9 amps and i got the same result. The NPN remained cool to the touch. I dont know if it matters but i upgraded all the resistors to 1/2 watt. Why do you think this is happening? I mean im glad there is no heat but curious as to why?

UPDATE: I put a thermal temp reader on the Darlington and the temp increased from 79 to 80F during the 1.5 second fade

 

[eTech]

Ive since moved components around. I needed the NPN on the side. I added this small pour from the collector to the lights output pin. The PWR track that went to the lights positive pin was not needed as this circuit is a negative (low) switch so it was removed. My question is this. The method i added the pour planes. I first added a full GND plane indicated by the 4 black arrows. then i drew a smaller pour plane in the top left and assigned to the collectors net. JUST TO BE SURE I DID THIS CORRECTLY. I could take it further and reduce the size of that track, add the pour, adjust the spokes (there will be only 1) from 12 to 20.

Hi
I'm not sure I understand your description or the corner image.
If the collector pin pad and the light terminal pin pad are on the same net, and they are thru hole pins, and you used a pour to connect them together on the top side, then there shouldn't be a trace between them on the top. The pour should be solid area with thermals (wide - 4-spoke thermal) located at each pad. There seems to be a red trace making the same connections but splitting the pour...

The grounded side of the power NPN should be connected to the ground pour (top and bottom) with thermals.

 

[ThomsCircuit]

If the collector pin pad and the light terminal pin pad are on the same net, and they are thru hole pins, and you used a pour to connect them together on the top side, then there shouldn't be a trace between them on the top. The pour should be solid area with thermals (wide - 4-spoke thermal) located at each pad. There seems to be a red trace making the same connections but splitting the pour...

I have found that its fine to draw pours over top of one another. Just to be sure they are applied in order.
This is how it looks now. Yes your right about the trace not supposed to be there. It is hidden and only there visually if i click it.
Ive rearranged the component to improve the number of spokes on the collector pin (2) and i have 6 on the output pin.

The grounded side of the power NPN should be connected to the ground pour (top and bottom) with thermals.

It will. I just have not applied it yet.

 

[ThomsCircuit]

The pour should be solid area with thermals (wide - 4-spoke thermal) located at each pad.

Here are both pours. Top only. If i add the bottom GND pour it will be difficult to see. I resized the pad on the base pin from 70 to 65 and it was enough to get 4 spokes on the collector pin. The emitter pin has 3 spokes.

 

[eTech]

Ive rearranged the component to improve the number of spokes on the collector pin (2) and i have 6 on the output pin.

Really, there's no real reason to add more, or less, spokes than 4. Remember what you are trying to accomplish with this small pour.
It's to provide a large amount of copper connectivity to the pad. Basically, the pour becomes the "wire". Adding more spokes removes copper and increases resistance. While the pads do need thermal relief, its better to stick with 4-spokes and make them wider (remember the reason for wide tracks in the first place).

 

[ThomsCircuit]

Really, there's no real reason to add more, or less, spokes than 4.

I did not realize that it worked that way. "Adding more spokes removes copper and increases resistance." I take it the spokes are not copper but another metal. The color being the same as the pour led me to believe they were the same hence the more spokes the better. Im glad i asked. Ill manually remove two spokes from the other pad. Reason i could not get the four on the collector pad without increasing the layout to six.

 

[ThomsCircuit]

Improved. Wider spokes

 

[rjenkinsgb]

Now if ive understood correctly the heat dissipation is only when its fading and not while it fully on or fully off.

That's if you are using a FET; a MOSFET will dissipate most during the PWM fade and very little while on.

The darlington dissipation will be proportional to the LED current, so continuous while the LEDs are on.

 

[eTech]

I did not realize that it worked that way. "Adding more spokes removes copper and increases resistance." I take it the spokes are not copper but another metal. The color being the same as the pour led me to believe they were the same hence the more spokes the better. Im glad i asked. Ill manually remove two spokes from the other pad. Reason i could not get the four on the collector pad without increasing the layout to six.

The spoke IS copper and are part of the pour. The "gaps" between the spokes "thermally" isolate the pad from the pour while the spokes maintain electrical conductivity between the pad and the pour. Each spoke is a "track" and passes current just like any other track. Therefore, with regard to electrical current, it has the same attributes. A thinner spoke has higher resistance and resists current flow, while a wider spoke has less resistance and will allow more current flow. You also want to keep the spokes evenly distributed around the pad so that heat is evenly distributed.

If you have trouble with clearances, I think there is a setting in the pour dialog.

If you have time, create a small project with a PCB, place some parts and play with the pours and clearance settings. You'll see how they affect how well the pour forms around tracks, pads, and thermal reliefs. The clearance settings are very inportant to know how to use.

 

[ThomsCircuit]

If you have time, create a small project with a PCB, place some parts and play with the pours and clearance settings. You'll see how they affect how well the pour forms around tracks, pads, and thermal reliefs. The clearance settings are very inportant to know how to use.

Your correct. I narrowed the isolation gap by just 1 pt and it added all the spokes evenly around the pad.

 

[ThomsCircuit]

The darlington dissipation will be proportional to the LED current, so continuous while the LEDs are on.

You are right. I left it on for a few minutes and the temp climbed 1 degree every 10 seconds.
After 15 minutes with a1.3amp load the temp topped out at 96 degrees F. A 22 degree increase. After that it would rise and fall 2-3 degrees.

 

[ThomsCircuit]

Plus the PCB tracks need to be wide enough to handle the current; I can't remember how they ended up.

Hi. I have crunched more number and performed a few tests. As i mentioned before the listed current draw and wattage was different from what i measured so there may be a chance i do not need the additional copper if i can widen the right tracks. Hopefully then i can avoid the $17.00 up-charge for the 2oz copper.
Ive added a pour from the collector to the output pin. The shape is not a rectangle but its area equals 190 cubic mm.

Every light with exception of 2 of them draw .5amps of current. The temp of the Darlington increases 14-15 degreesF under this load. Would you agree that the darlington is good for loads under .5 amps?

The other two draw 1amp each. The darlingtom reaches 99F under this load. I personally would use a mosfet for these circuits if not to only reduce the heat dissipation but for the knowledge i have gained from your assistance. (I actually thought a FET and a mosFET were the same thing) My question for these 1amp loads is what tracks would i need to increase to meet the goal necessary?

I mean i looked up what the safe operating temp was for the darlington and 150c is not something im going to subject this circuit to.

 

[eTech]

Hi. I have crunched more number and performed a few tests. As i mentioned before the listed current draw and wattage was different from what i measured so there may be a chance i do not need the additional copper if i can widen the right tracks. Hopefully then i can avoid the $17.00 up-charge for the 2oz copper.
Ive added a pour from the collector to the output pin. The shape is not a rectangle but its area equals 190 cubic mm.

Every light with exception of 2 of them draw .5amps of current. The temp of the Darlington increases 14-15 degreesF under this load. Would you agree that the darlington is good for loads under .5 amps?

The other two draw 1amp each. The darlingtom reaches 99F under this load. I personally would use a mosfet for these circuits if not to only reduce the heat dissipation but for the knowledge i have gained from your assistance. (I actually thought a FET and a mosFET were the same thing) My question for these 1amp loads is what tracks would i need to increase to meet the goal necessary?

I mean i looked up what the safe operating temp was for the darlington and 150c is not something im going to subject this circuit to.
View attachment 135585

Hi

I couldn't help but notice...
Could you identify the pins (C B E) for the darlington NPN?

Looks like the pins might be swapped in error.

 

[rjenkinsgb]

I actually thought a FET and a mosFET were the same thing

Generically, they can be - "FET" covers all variations.
As well as being P or N, can be a JFET (junction FET) or MOSFET - Metal oxide semiconductor; the junction is insulated from the source-drain channel, rather than being like a diode junction.

Plus other specialised variations of those types - Dual gate FETs (a bit like tetrode valves/tubes), GAASFETs, FREDFETs etc.

The 40 thou tracks were OK for around 2.5A if I remember right? No problem for 1A.
If there is a pour, that part will stand rather higher current.

ps. I believe you mean square millimetres, not cubic!