Need help on how to read correct UC2845 PWM signal with oscilloscope !!

[??? ????? ??????]

Hi all , am new to this board .

So what got in hands is a MMA inverter welder 160 A , where the problem is mainly on the chopper area , i have found that the chopper IGBTs are shorted so i have replaced them with new ones and BAM , they shorted again and the unit shut down .

Now i have removed the shorted transistors , and started to analyse deeply where the problem is hanging .

I started this mysterious mission by analyzing the control section that drives the IGBTs , this is done by a UC2845 PWM ic .

So i have checked the VCC pin of it and i see a 15 Volts dc , good input so far , but when it comes to the output , the unit is working with a 60 Khz signal , with a 65% Duty cycle .

So in order to read that signal i used an old HUNG CHANG 5502 20 Mhz oschilloscope using two methodes :

1) connecting CRO GND to IC GND pin and probing the ic output :
result : High signal ( 12 V dc ) with no oscillations

2) without using the GND i only connected the CRO (+) probe to the IC output and here is the signal i get :

So what you think of this results ? i don't see any square wave signal , is this considered bad IC or am not reading the signal the right way ?

 

[chuckey]

In 2 it looks like some mains frequency pick up. You must have an earth or use a dual channel oscilloscope switched to A-B, and have both attenuators set the same.
Have you checked for any safety interlocks that might stop the welder, is there a DC out of some extroadinarily large rectifiers?, or even mode controls (foot pedal to zap!).
Frank

 

[??? ????? ??????]

Hi i have used the GND from pin 5 and checked the output at pin 6 , i see 12 volts DC with 0 Hz.
What should i normally get at the output ? i don't see what maybe causing the chopper to short , the transformer is good and it has thick windings i would not see it as suspect but the IC is what confusing me right now . am just trying to make sure i read the real output signal ( good or bad )

Primary section is good , checked also the board for shorted diodes , but it seems like something is making the transistors work at full duty , and eventually burn .

But that just a guess .

 

[??? ????? ??????]

Any other opinions are welcome

 

[chuckey]

the chip uses external components to time its oscillations, are these OK?. Data sheet :- https://www.ti.com/general/docs/lit/getliterature.tsp?genericPartNumber=uc2845&fileType=pdf There is a "typical" circuit" in this, do the Dc voltages make sense?
Frank

 

[??? ????? ??????]

Yes they all check ok .

 

[chuckey]

From the data shet you should be able to find some pin with oscillations on it. If the components are good and the feedback is OK, no oscillations mean the chip has had it. One other thing, you say the output pin is stuck at 12v. if you can, disconnect it from any other circuits, to see if its feeding 12V out or something else is feeding 12V in.
Frank

 

[??? ????? ??????]

Its hard to disconect the output , this is SMD module unfortunatly .

Any other alternative ?

BTW am doing this just for learning .

 

[chuckey]

Cut through the PCB track. You can bridge it with a piece of wire when finished, just make sure that the track runs from the output pin before branching out.
Frank

 

[??? ????? ??????]

Will do , i hope i get some new results

 

[??? ????? ??????]

I have the square wave now , the IC is good .

As for the settings i have chosen 10 uS /DIV , and i get 3 divs high 3.5 divs low . and 12 V oscillation

Is this the right frequency calculus :

1/(10*10^(-6)s*6.5) = 15.38 Khz ??? technical product specs say that frequency is about 60 Khz