UC3843A: uc3843 skips the working cycles

user5213571

Part Number: UC3843A

Hi. I am making a boost DC/DC converter from 12VDC to 27.8VDC@100mA with adjustable output voltage from 24v@30mA to 30@160mA.

Power source - battery. Depending on the level of charge, it produces a different voltage. When fully charged - 13v, and with a full discharge of 10v. Also, sometimes the device is connected to the power supply unit for 19 volts. When the input voltage changes or when the output voltage is adjusted, the controller skips the operating cycles. It looks like this:

How to fix it? what is it "slope compensation" why is it needed and what does it do and how calculate R7? I have made it but I don't understand. Is feedback compensation necessary if there is an slope compensation? how to calculate the values of feedback compensation elements?

over 7 years ago

0 Benjamin Lough over 7 years ago

TI__Mastermind 26735 points

Hi User,

Thanks for your interest in UC3843A. I would recommend taking a look at section 3 of this application note: www.ti.com/.../slup317.pdf. There is a section explaining slope compensation on page 17. Slope compensation is typically necessary for current mode control devices to avoid sub harmonic oscillation. This can become an issue when you operate above 50% duty cycle.

You will still need feedback compensation even with slope compensation. Page 10 of this document provides some guidance for setting compensation components for current mode control boost: www.ti.com/.../slup340.pdf.

What are the red and teal waveforms in the images you shared? Also could you explain how you are adjusting the input voltage and output voltage in your simulation?

Best Regards,
Ben Lough

0 Benjamin Lough over 7 years ago in reply to Benjamin Lough

TI__Mastermind 26735 points

Hi User,

Were you able to resolve this issue?

Best Regards,
Ben Lough

0 user5213571 over 7 years ago in reply to Benjamin Lough

Prodigy 70 points

red - output voltage on load, teal - output error amp pin 1, lilac - input error amp(pin2), green - isense (pin3), blue - Vg mosfet (pin 6).
The output voltage is controlled via a variable resistor. In a real device, the resistors r3 r4 are one variable resistor at 10kΩ. In the simulator, I just change the ratio of r3 and r4.
The input voltage I do not regulate in any way. The device is powered by a powerful 12 volt battery. And its increased current consumption in this case is not a problem. But I would like to connect it to the 19v source. While I do not do this, because there is a tight current limit. And if the device starts to work in intermittent mode, this source will simply burn.
In the simulator, I just change the source voltage in the settings and simulate with the new parameters. What I've shown is not a transient process, it's a steady mode of operation with constant input and output voltages. I do not change the input voltage during operation.

0 user5213571 over 7 years ago in reply to Benjamin Lough

Prodigy 70 points

no. I read this very useful document www.ti.com/.../slup340.pdf and try to figure it out. I also found one interesting fact. If you change the control circuit of the gate of the FET to this one. That device begins to work stably in those modes in which stability before it could not be achieved. I think that this is all due to current emissions when the transistor is opened. This circuit tightens the opening of the transistor and the current ejection decreases. So maybe it's not about correcting the feedback but in the current needles on the current sensor?

0 Benjamin Lough over 7 years ago in reply to user5213571

TI__Mastermind 26735 points

Hi User,

R2 will slow down the gate drive signal and make the transistor turn on slightly slower. This helps if you have ringing on the drain of the MOSFET due to parasitic inductance and capacitance in the circuit. D2 allows you to still have a fast turn off time.

Also is the controller you are using UC3843A or is it LT1243?

Best Regards,
Ben Lough

0 user5213571 over 7 years ago in reply to Benjamin Lough

Prodigy 70 points

UC3843A. There is no uc3843A in ltspice, i find this lib in internet. Perhaps there is another way to remove the parasitic emission on isens? Will this help? or is it still necessary to make feedback compensation?

0 Benjamin Lough over 7 years ago in reply to user5213571

TI__Mastermind 26735 points

Hi User,

It is still necessary to have feedback. I would suggest increasing C3. As a side note, there are pspice models available for UC3843A available here: www.ti.com/.../toolssoftware

Best Regards,
Ben Lough