Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
Lol good for now until fully tested. I will post results.{insert looks good happy icon here} LOL
It is amassing how much time it takes to be happy with a design.
Being awakened by a dream of a hot transistor chasing after me I had to look at this circuit a bit closer.{insert looks good happy icon here} LOL
It is amassing how much time it takes to be happy with a design.
Here is an SMD version of the pcb with the last version of schematics above. The ground plane (heat sink) can be resized from top to what you desire.{insert looks good happy icon here} SMD.
I updated the smd pcb. see above. Upper tab of top SMD Q is the Collector and will be soldered in to top copper (in this case the ground plane). The tab is there, it is just hidden by other layers. For me to show you different layers I have to create individual layers separately otherwise other layers will mask over them.Note the tab needs to be soldered to the heat sink. So I cleared the solder mask off for the tabs. Maybe should be a larger area.
Look at the FZT792A.
View attachment 98475
I ordered parts to build it. However I have a different take on Ws chip. I believe they used a high speed, low On-Resistance NMOS (0.1 ohms). This internal resistance would dictate the voltage drop. Again like you I am guessing if such high tech, high speed PWM device may not be as inefficient as you assume it to be! If my guess is correct Vdrop=0.0185x0.1=0.2mv which is too insignificant to worry about. We will know once it is tested.There are things I don't know. I believe the bottom of the 5.4 ohm resistor will not be at ground. Because there is a constant current source inside the IC the output can not pull to ground. I don't know but I think it will be 0.7 volts minimum.
Please see my previous notes regarding voltage and current fluctuations and how negative feedback will bring the system back to equilibrium to compensate. Once again it is worth mentioning that although we set initial voltages for operation, the device will keep a constant current thru the LED. See my next post for a new configuration that addresses Vdrop of WS and any fluctuations in Vf and If due to heat and how system will recover and compensate.Today we say a 10% change in LED voltage from 25C to 85C. Is this in your calculations. The voltage was rated at 1.7V but the maximum could be as much as 2.5V and as little as (they did not say). Now I know you will rarely see 2.5V but the LED is still "good" at that voltage.
I have been looking at the original simulation. It is a LED driver but it is not a constant current. The PNP transistors do not allow for the current to be capped and it will ramp continuously even though supply voltage does not change. However the device does not compensate for LED's voltage and current shifts when it gets warmer which was one of the design requirements. To overcome this I went back to NPN transistors and redid the design from ground up. Below please find the new design with simulated "S" curve graph. It also counted for Vdrop of WS2811 (assumed around 0.76V).{insert looks good happy icon here} LOL
It is amassing how much time it takes to be happy with a design.
True. However you are driving a 300ma led in this case. So you need to boost the current and keep it constant at the same time from 18.5ma! The challenge in this case is to have some headroom between Vf vs. supply voltage while considering voltage drop of WS plus two stages of Vbe on top of that. In case of Blue and Green with 9.6v for Vf, that headroom is tight. The fact that I could get a flat current starting close to 11v and going up was the key accomplishment here. For Red with Vf of 6v, you have plenty of headroom.I think the WS2811 has this as the output.
Current is monitored by watching the voltage across R1.
Current is set by Vref and R1.
If what I drew is right then the WS2811 does not care if the LED is 1.5V or 3V or 9V the current will be the same. (as long as there is some headroom)
The output voltage can not be less than Vref (at 18.5mA)
View attachment 98658