Current mode regulating PWM

Hi all,

I'm busy trying to repair a monitor (you might have seen my earlier posting on the GND point where I almost blew myself up and have a few questions I hope someone can answer. Its a long one so please bear with me.

When powering on the monitor the voltages on the PSU board keep rising and falling and the relay for the degaussing keeps clicking - presumably from the rising and falling voltages. I've ripped the PSU board out and am working on it separately.

I've attached the circuit diagram I found:
**broken link removed**

I thought I found the problem which was with IC601, a UC3842, a current mode regulating PWM. (The red circle on the diagram) If i disconnect the Vcc from the chip the voltage is stable at 30V, reconnecting the pin causes the voltage to climb to 15V and then suddenly drops to 10V. It rises again to 15V in a seconds and drops again etc etc

I've also tried disconnecting all the pins except GND and Vcc and the problem still occurs - hence why I thought it was faulty. I've now replaced the chip and also the 15V zener (D604) which I also thought might be faulty.

Since the new chip does the same thing I presume there's nothing wrong with it and rather something else is causing it. Unless my logic is completely of course here.

Here's a link to the PWM:
https://www.electro-tech-online.com/custompdfs/2007/07/IP1842_3.pdf

Now onto my question (finally) - has anyone used the above chip before, is it supposed to exhibit the above symptoms I described and am I look completely down the wrong path for the problem?

I know I could take the monitor in for repair, but I'd like to repair it myself and learn as much as possible in the process - and now its become an obsession to fix it as well!

gregmcc said:

Hi all,

I'm busy trying to repair a monitor (you might have seen my earlier posting on the GND point where I almost blew myself up and have a few questions I hope someone can answer. Its a long one so please bear with me.

When powering on the monitor the voltages on the PSU board keep rising and falling and the relay for the degaussing keeps clicking - presumably from the rising and falling voltages. I've ripped the PSU board out and am working on it separately.

I've attached the circuit diagram I found:
**broken link removed**

I thought I found the problem which was with IC601, a UC3842, a current mode regulating PWM. (The red circle on the diagram) If i disconnect the Vcc from the chip the voltage is stable at 30V, reconnecting the pin causes the voltage to climb to 15V and then suddenly drops to 10V. It rises again to 15V in a seconds and drops again etc etc

I've also tried disconnecting all the pins except GND and Vcc and the problem still occurs - hence why I thought it was faulty. I've now replaced the chip and also the 15V zener (D604) which I also thought might be faulty.

Since the new chip does the same thing I presume there's nothing wrong with it and rather something else is causing it. Unless my logic is completely of course here.

Click to expand...

Sorry, but it was never going to be the chip!.

It's no good replacing parts at random, you need to understand how the circuit works - for a start that it's a closed system, and faults anywhere can cause this type of 'tripping' problem. Your first mistake is removing the PSU, it's VERY likely that it won't run without a suitable load, and may well have control signals from the rest of the set?. So for a start, apply a suitable load on the output - 60W 240V incandescent bulbs usually work quite well.

You should download the datasheet for the chip, which will tell you how it needs to be connected, and what each pin does.

just a thought

gregmcc said:

Hi all,

I'm busy trying to repair a monitor (you might have seen my earlier posting on the GND point where I almost blew myself up and have a few questions I hope someone can answer. Its a long one so please bear with me.

When powering on the monitor the voltages on the PSU board keep rising and falling and the relay for the degaussing keeps clicking - presumably from the rising and falling voltages. I've ripped the PSU board out and am working on it separately.

I've attached the circuit diagram I found:
**broken link removed**

I thought I found the problem which was with IC601, a UC3842, a current mode regulating PWM. (The red circle on the diagram) If i disconnect the Vcc from the chip the voltage is stable at 30V, reconnecting the pin causes the voltage to climb to 15V and then suddenly drops to 10V. It rises again to 15V in a seconds and drops again etc etc

I've also tried disconnecting all the pins except GND and Vcc and the problem still occurs - hence why I thought it was faulty. I've now replaced the chip and also the 15V zener (D604) which I also thought might be faulty.

Since the new chip does the same thing I presume there's nothing wrong with it and rather something else is causing it. Unless my logic is completely of course here.

Here's a link to the PWM:
https://www.electro-tech-online.com/custompdfs/2007/07/IP1842_3-1.pdf

Now onto my question (finally) - has anyone used the above chip before, is it supposed to exhibit the above symptoms I described and am I look completely down the wrong path for the problem?

I know I could take the monitor in for repair, but I'd like to repair it myself and learn as much as possible in the process - and now its become an obsession to fix it as well!

Click to expand...

I glanced at the schematic, but it is hard to read. Anyway, I don't see any slope compensation. Also, the RtCt pin, no. 4, should consist of the timing resistor and capacitor connection. I don't know why the other parts are connected, unless it is for the purpose of synchronizing the two 3842 parts. The ISNS pin, no. 3, for current sensing, does not appear to have any slope compensating elements.

Peak current mode converters, especially as the duty factor equals or exceeds 50%, can exhibit subharmonic oscillation if no slope comp is used. Even below 50% duty cycle, an underdamped condition exists in the inner loop (current) where the PWM hunts for the right current value. The output will exhibit large ripple because the servo loop controls the inductor current which in turn controls duty factor. If the current loop is underdamped, the current value will overshoot the target value, forcing an overvoltage at the output. The outer loop (voltage) senses the error and reduces the target current value accordingly which reduces the PWM duty factor. Due to the underdamped current loop, the current then undershoots the target, resulting in an undervoltage output condition. The outer loop corrects by increasing, then overshoots, decreases, then undershoots, etc. etc. This seems to agree with the symptoms you've just described.

Adding slope compensation quashes this tendency. But, the OEM didn't do so and the device worked well until recently, I presume. If you have a better quality schematic, I'll look again tonight. Maybe it's a lot simpler than what I just described. I doubt that I've helped much. Best regards.

Thanks for the tips - thought it wasn't going to be that simple. I've connected the power supply to a load still the same problem. I've got the datasheet but will take a closer look at the chip and see exactly what its supposed to do and why the Vcc drops all the time.

gregmcc said:

Thanks for the tips - thought it wasn't going to be that simple. I've connected the power supply to a load still the same problem. I've got the datasheet but will take a closer look at the chip and see exactly what its supposed to do and why the Vcc drops all the time.

Click to expand...

The chip is powered from a winding on the transformer, so the PSU needs to be running for the chip to work - which in turn requires the chip to be working first. To overcome this, the chip is initially fed by a couple of high value resistors in series, these charge a capacitor up, which then suppies the energy to start the chip, the PSU starts running, and it then feeds itself.

Any short on the secondary will make it repeatedly shutdown and restart, as will anything preventing the PSU supplying the chip.

A common fault is duff electrolytics, you should check all of them with an ESR meter!.

Thanks Nigel. At least I can head in the right direction now. I didn't know that the windings were powering the chip - I wondered why the 2 diodes were connecting back to the transformer though.

I'll dig out my esr meter and keep you posted.

Clude Abraham:
The oscillators in the two control I.C.s are connected together so they will run at the same frequency. A signal from the horizontal section of the monitor enters the power supply through a 2 pin connector and forces the oscillators to run at the horizontal frequency. This does not help fix the supply but may answer a question.

Slope compensation is done at several places. A little is done at the 3842 error amplifier but that amplifier is really not in use. The real error amplifier is a TL431 on the secondary side.

Gregmcc:
I design there things, that does not mean I can fix them. Because the 3842s get power from their own power supply it is hard to see what is wrong. When I first test a design I start out very carefully. I connect a 0-30 volt bench supply to the power supply pin on the 3842. There is a point at the top of the page called F or E, I cannot see clearly. At about 15 volts they should start to work. There should be a square wave on pin 6 (the gate drive pin).

Next I connect a different supply across the main input filter capacitors. (330uf 250 volt X 2) Start out at 5 to 10 volts (current limited). The power FETS should switch on/off. A MOSFETS may be bad. OR You likely have a dead output diode. You will now get a very small voltage at each output. If you have a scope look at the AC side of each power diode. There should be a signal. Increase the supply voltage. I have 300 volt supplies for this type of testing. If you can only get to 30 volts then the outputs should be running at about 1/10 voltage. All the power parts will be working and you may find the bad one.

If all you have is a meter; The power diodes are hard to measure in circuit. You may need to pull the diodes or pull the transformer to measure the diodes with a meter.

Good luck