How to wire up PWM control chip to DC DC boost circuit

Status
Not open for further replies.

Thread Starter

lex4794

Joined Jun 3, 2021
12
Hello, I need some help connecting an 8 pin current mode PWM controller to my boost converter circuit. Specifically this one. I have used controller IC's with integrated switches to make boost converters before, however, the amount of power I need to run through the circuit requires an external MOSFET switch. A diagram of my boost converter circuit is below, I have tested the circuit with another IC (with an integrated switch and only 5 pins) and it works. I can figure out where some of the connections go (for the IC in question) but the purpose and use of the other 4 pins, COMP, VFB, RT/CT, and REF are beyond my understanding, I have read the entire datasheet. A diagram of the controller IC is also located below. Where and how do I connect the other 4 pins to the circuit? Thanks in advance!
Controller ICCircuit Diagram
 

Papabravo

Joined Feb 24, 2006
21,159
I think the other four pins are essential to the proper operation of the device. I can pull the datasheet and help you out. Are you expecting to get 56V @ 4 amps or 225 watts out out this supply, with an input power of 279 watts @ 9.3 V to 378 watts @ 12.6 Volts without worrying about those other four pins? If so I don't recommend it. Is the load an LED string of some description? Also tell me about the inductor, the MOSFET and the diode. I'm not sure I know how the 5K pot is wired either. Could you elaborate? Would it be capable of shorting the output?

EDIT: I have the datasheet and using this chip is not a simple matter of wiring it up. It will require some sophisticated design decisions and a printed circuit board. These power levels are not suitable for implementation on a breadboard. Using it without understanding it is a recipe for disaster.
 
Last edited:

Thread Starter

lex4794

Joined Jun 3, 2021
12
I think the other four pins are essential to the proper operation of the device. I can pull the datasheet and help you out. Are you expecting to get 56V @ 4 amps or 225 watts out out this supply, with an input power of 279 watts @ 9.3 V to 378 watts @ 12.6 Volts without worrying about those other four pins? If so I don't recommend it. Is the load an LED string of some description? Also tell me about the inductor, the MOSFET and the diode. I'm not sure I know how the 5K pot is wired either. Could you elaborate? Would it be capable of shorting the output?

EDIT: I have the datasheet and using this chip is not a simple matter of wiring it up. It will require some sophisticated design decisions and a printed circuit board. These power levels are not suitable for implementation on a breadboard. Using it without understanding it is a recipe for disaster.
No the output is more like 54V 3 amps but I errored on the side of caution. I calculated the inductor size for this specific IC to be 2.32 uH. I should also point out that the peak current draw happens over a short time period (don't need to use 26,000 lumens often) , most of the time the LED will operate at no more than 40-60 watts. The LED in question is a Lumix COB LED. I used a 5k pot in place of the shunt resistor detailed in the datasheet. MOSFET. The diode is a Schottky rectifier diode 35A. Would I just be better off building a 555 timer circuit in a PWM configuration? I don't really think I need the feedback functionality of these chips. I guess I wrongly assumed that a dedicated IC designed for the purpose of controlling a DC-DC power system would be less of a hassle. Also yes I know not to use a breadboard to test this at full power, I am using 14 gauge wire just to connect it to the batteries.
 
Last edited:

Papabravo

Joined Feb 24, 2006
21,159
No the output is more like 54V 3 amps but I errored on the side of caution. I calculated the inductor size for this specific IC to be 2.32 uH. I should also point out that the peak current draw happens over a short time period (don't need to use 26,000 lumens often) , most of the time the LED will operate at no more than 40-60 watts. The LED in question is a Lumix COB LED. I used a 5k pot in place of the shunt resistor detailed in the datasheet. MOSFET. The diode is a Schottky rectifier diode 35A. Would I just be better off building a 555 timer circuit in a PWM configuration? I don't really think I need the feedback functionality of these chips. I guess I wrongly assumed that a dedicated IC designed for the purpose of controlling a DC-DC power system would be less of a hassle. Also yes I know not to use a breadboard to test this at full power, I am using 14 gauge wire just to connect it to the batteries.
I'm sorry but this is not a run of the mill 555 type PWM generator. It is a current mode controller. You did not read carefully, and did not understand the datasheet. Let's start with some basic concepts. What frequency do you expect the internal oscillator to run at. the reason I ask is that it needs a specific R and a specific C to select the frequency and you have no such components on your drawing. It also has provisions for slope compensation which is necessary to run at duty cycles over 50% which you intend to do. If you don't connect the feedback and compensation pins the chip is going to behave strangely with those inputs floating.

Yes, for your purposes you'd be better off using a 555
 

Sensacell

Joined Jun 19, 2012
3,432
The way you have this configured, the current through the LED is NOT measured or controlled.

This is a magic smoke generator.
 

Thread Starter

lex4794

Joined Jun 3, 2021
12
I'm sorry but this is not a run of the mill 555 type PWM generator. It is a current mode controller. You did not read carefully, and did not understand the datasheet. Let's start with some basic concepts. What frequency do you expect the internal oscillator to run at. the reason I ask is that it needs a specific R and a specific C to select the frequency and you have no such components on your drawing. It also has provisions for slope compensation which is necessary to run at duty cycles over 50% which you intend to do. If you don't connect the feedback and compensation pins the chip is going to behave strangely with those inputs floating.

Yes, for your purposes you'd be better off using a 555
I understand that this is not a 555 type PWM generator, I was curious to see if I could use it to output a PWM signal to a MOSFET based on the input to the feedback pin much like This IC except with an external switch. Clearly, I read and realized did not understand the datasheet hence why I posted a thread. Thank you for clarifying that this is far more complicated than I expected, I will be using a 555 timer.
 
Last edited:

Papabravo

Joined Feb 24, 2006
21,159
I understand that this is not a 555 type PWM generator, I was curious to see if I could use it to output a PWM signal based on the input to the feedback pin much like This IC. Clearly, I read and realized did not understand the datasheet hence why I posted a thread. Thank you for clarifying that this is far more complicated than I expected, I will be using a 555 timer.
I was actually torn between trying to help you use it and some other alternative. I realized you would have one hell of time troubleshooting this thing if anything went south, and when I determined that the datasheet was actually lacking some important information and even I would have a hard time with this chip without a good deal of time on my hands to experiment with alternative methods. This is probably because TI (whose datasheet I pulled) might actually be a second source rather than the primary developer of the chip. I don't know how you could successfully use this chip with just the TI datasheet unless you'd used a similar one in the past or there is an application note somewhere that fills in the missing details.
 

Thread Starter

lex4794

Joined Jun 3, 2021
12
I was actually torn between trying to help you use it and some other alternative. I realized you would have one hell of time troubleshooting this thing if anything went south, and when I determined that the datasheet was actually lacking some important information and even I would have a hard time with this chip without a good deal of time on my hands to experiment with alternative methods. This is probably because TI (whose datasheet I pulled) might actually be a second source rather than the primary developer of the chip. I don't know how you could successfully use this chip with just the TI datasheet unless you'd used a similar one in the past or there is an application note somewhere that fills in the missing details.
Yea, I think I started to realize that I was biting off more than I could chew after I read your first post. It was worth a try though. Thank you for your help.
 

ronsimpson

Joined Oct 7, 2019
2,988
The UC3843 is not quite the right part for this job but here is my first try.
First do not run the 4ALEDs from a 54V supply. They need current not voltage. In this case I am regulating the current through the LEDs not the voltage.
3.5A and about 55 volts for the LEDs.

1622781806995.png
Red trace is the D voltage on the FET. Blue is the LED current near 4A. Green is the LED voltage. Blue-light is the current in the inductor.
1622782088505.png
 

Thread Starter

lex4794

Joined Jun 3, 2021
12
The UC3843 is not quite the right part for this job but here is my first try.
First do not run the 4ALEDs from a 54V supply. They need current not voltage. In this case I am regulating the current through the LEDs not the voltage.
3.5A and about 55 volts for the LEDs.

View attachment 240420
Red trace is the D voltage on the FET. Blue is the LED current near 4A. Green is the LED voltage. Blue-light is the current in the inductor.
View attachment 240421
Thank you for this really cool circuit schematic, however, it is far beyond my capabilities. Is there anyway for me to close the thread or mark it as solved?
 
Last edited:

Ian0

Joined Aug 7, 2020
9,668
The UC3843 is not quite the right part for this job but here is my first try.
First do not run the 4ALEDs from a 54V supply. They need current not voltage. In this case I am regulating the current through the LEDs not the voltage.
3.5A and about 55 volts for the LEDs.

View attachment 240420
Red trace is the D voltage on the FET. Blue is the LED current near 4A. Green is the LED voltage. Blue-light is the current in the inductor.
View attachment 240421
Just curious - what's the purpose of Q2? It's not normally in a UC3842 circuit.
 

Ian0

Joined Aug 7, 2020
9,668
There are similar parts such as AL8853 which are designed for current-drive to LEDs and have lower reference voltages.
Your R5 will dissipate 9.1W.
 
Status
Not open for further replies.
Top