UC 3842/43 boost converter design

[Saitama]

Hello.
I'm making or at least trying to make a boost converter out of UC3842 IC, but no matter what I do, I just can't seem to do it correctly. I went over the datasheet (linked below) many times and tried to copy other's design, but I still get no results.
The main problem is that the gate voltage on the transistor stays the same. The duty cycle is very small and when I try to change the feedback resistors, it doesn't change. Even the feedback voltage is wrong.
From what I could gather, it's supposed to vary the output, so that the voltage on the feedback pin stays ~2.5V.
Did I choose my components wrong? or could it be that I suck at building boards..

To clarify, here are the specs I'm trying to achieve:
frequency ~100khz
output voltage ~200v
max current ~1A
 
datasheet link: https://www.ti.com/lit/ds/symlink/uc3842.pdf?ts=1683607532917&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FUC3842

[TurboTom]

200W of output power out of a single-phase step up converter at a voltage ratio of almost 20:1 is a challenging job to tell the least. You are looking at an average input current of ~20A give or take a little, and, depending on the configuratoin, a peak-to-peak current ripple of 40A. The problem with a simple flyback is that you need a switching transistor that's rated for high voltage. The consequence is usually that it's got a rather high Rds. Hence, the power dissipation is high (conduction losses). The transistor you selected will dissipate in the ballpark of 50W in this application.

Moreover, the setup with a current shunt (to control the current-mode operation) in the ground node of the switch also adds to the dissipation. Also, a direct link from the shunt to the CS pin of the controller is a no-go. You always want a series resistor and (usually) a capacitor to ground there in order to dampen feed-through pulses from the gate drive. It's also recommendable to place a series resistor in the gate drive node, possibly with another one with a diode in parallel so the switch-off time can be kept shorter than the switch-on time.

I made many charging or HV "gadgets" with the UC3842 and its siblings -- usually, they are quite easy-to-use and work really well in the jobs they are designed for, but this doesn't include a high-power single-phase step up converter.

For this job, I'ld go for a push-pull forward or resonant converter, maybe a royer to drive a high frequency step-up transformer, so the switches can be kept operating at low voltage and hence their on-state resistance is just a few milliohms, dissipating much less power. But you've got to do your homework to get a circuit like this going reliably...

[Saitama]

So, what I need is a resistor and a capacitor at the input of current sense pin and a series resistor with a parallel diode at the gate drive.
Realistically speaking, I probably won't need more than 100mA at high voltages out of this supply, 1A was intended to be kind of like peak output at lower voltage, since I want it to be variable in the end (~50-200v), but the problem is that it doesn't even reach the desired voltage when disconnected from any sort of load. It drops to the supply voltage (minus diode drop) immediately after loading it with a 1K resistor.
I will still look into other types of converters, just like you suggested, but I don't want to leave this project without a bit of success.
Is there anything I can do to make it do it's job?

edit: without load, I got it to output 28v from a 12v supply.

[T3sl4co1l]

FYI, you must have '43 on 12V supply.  '42 is only guaranteed to start by 16V, then can run down to 10V.

How are you measuring gate voltage / duty?  What do the other pins show?

Tim

[TurboTom]

Sorry, but it's not as easy to just add this or that component to your schematic. The datasheet denotes the schematic that you apparently copied, for good reason as "simplified".

You've got to understand the working principle of the current mode controller as described in the datasheet very well. Basically, the IC switches on the MOSFET until the current charging the inductor reaches a "trip value" determined by the shunt and the feedback voltage. After that, the energy stored in the inductor is (completely or partially, depending on the choice of components and operating conditions) transfered to the HV output side. The voltage across the shunt, required to trip the end of charge comparator, can be as high as 1V (as specified in the datasheet). This already indicates that that particular SMPS controller may not be the ideal choice for your application since you waste almost 10% of your input voltage across the shunt.

This won't matter that much in case of low power flyback converters or mains-powered off-line (isolated) switchers, which this series of SMPS controllers was originally intended for, but even a 20W output converter would dissipate more than 2W in the shunt which needs to be much lower resistance than your choice of ~1 Ohm. Select the shunt to drop 1V at the peak charge current that your inductor permits. In case of a 20W inverter, I'ld expect this to be somewhere between 4 and 5A, so a shunt of 0.22 Ohms may be adequate. At high, fast changing currents, circuit layout topology becomes very important, especially in case of the ground node, otherwise you will end up with instabilities.

I'ld recommend to design a real low-power converter first (1 or 2W, maybe also at a lower, safer output voltage) to familiarize yourself with the quirks of such circuits, and after that, consider approaching the higher powered varieties. Read and understand the datasheet and the design hints contained there, it provides a lot of valuable information. And follow @T3sl4co1l's hint regarding the power-up threshold voltage -- use the UC3843 

[Benta]

Basic question: which inductor are you using?

[moffy]

With everything perfect and a 90% on cycle, the best you could get out of 20uH@100kHz and 12V supply, is about 30W i.e. (0.5*L*Ipk^2)*f. That is for well above the 12V for the output like your 200V.

[Saitama]

How are you measuring gate voltage / duty?  What do the other pins show?

I'm using an oscilloscope to measure the gate voltage. It's a little usb scope (owon vds 1022) and the other pins seem to be working fine. I'm supplying correct voltage to the input pins, output frequency is about 100khz as expected, but the problem is with the output and feedback pins.

Basic question: which inductor are you using?

I'm using a toroidal inductor I wound myself. I checked it's inductance with one of those cheap LCR meters you find online and it said 0.02mH.

[Benta]

Basic question: which inductor are you using?

I'm using a toroidal inductor I wound myself. I checked it's inductance with one of those cheap LCR meters you find online and it said 0.02mH.

Inductance is one thing, but have you done any current/field calculations? Will it still be an inductor at 40 A?