Is this PWM IC output bad ? or its my oscilloscope ?

[BGAmodz]

Hello everyone .

I have analysed the output signal of a UC2845 PWM ic , what do you think about this signal sen at the output , this signal looks like this in every TIME/DIV setting .

Here is a video that describes it :

 

[4pyros]

Looks like background noise. What voltage scale are you using?
Do you have a load on it?

 

[ronsimpson]

It would be nice to know:
volts/div
uS/div
and where are you looking.
Schematic please.

 

[BGAmodz]

Hi again , the inverter am analysing does not have IGBTs on it , and have no load eather , this inverter makes the IGBTs short due to some kind of failure , i think its the IC uc2845 responsible for that as it is the one driving the IGBTs through pin 7 .

The IC is receiving a nice 15 V dc , but not outputting a correct square wave like mentioned on the datasheet .
The signal i read is from pin 7 ( output pin )
Here are the settings i use :
5 ms/div
1v/div
probe X10

NOTE that that flickering problem is not affecting the signal form .
I hope i explained my case very well .

Update :

I obtained the first signal without connecting the probe's GND clip to ground ( thinking its ok ) and also i was putting a light bulb in series with mains , in case some thing bad occurs .
When i remove the light bulb the signal changes a bit :

But what's weird is when i connect the probe's GND clip to ground i get constant high signal without oscillations and nothing like before .

 

[JimB]

Is the frequency compensation set correctly on the X10 probe ?

When I saw that scope trace, my first reaction was - poor high frequency response.

JimB

 

[BGAmodz]

Is the frequency compensation set correctly on the X10 probe ?

When I saw that scope trace, my first reaction was - poor high frequency response.

JimB

Please take a look at my previous post's update .

 

[BGAmodz]

More opinions are very welcomed

 

[rumpfy]

hi BGA,
My feeing is that you have a problem involving the potential of the earth of the CRO compared to the earth of the welder.
Not sure which welder we are talking about, but I mentioned in a post somewhere, that I use a pair of back to back isolating transformers to allow the CRO ground lead to sit on the power supply negative point.
The traffos I use are about 50 watts; you can short out the transformed secondary and the short circuit current is only a couple of hundred milliamp; BUT its enough to do the sort of thing you are trying to do. With the traffos, you wont need the series resistor.

 

[4pyros]

and have no load eather

You may need a load to see the output properly.

 

[BGAmodz]

You may need a load to see the output properly.

I cant put a load , because i don't have any output and also no Chopper transistors , am aiming to see if the IC is outputing the right signal that is driving the gate of the Transistors .

hi BGA,
My feeing is that you have a problem involving the potential of the earth of the CRO compared to the earth of the welder.
Not sure which welder we are talking about, but I mentioned in a post somewhere, that I use a pair of back to back isolating transformers to allow the CRO ground lead to sit on the power supply negative point.
The traffos I use are about 50 watts; you can short out the transformed secondary and the short circuit current is only a couple of hundred milliamp; BUT its enough to do the sort of thing you are trying to do. With the traffos, you wont need the series resistor.

Hi rumpfy , this thread is rerlated to an old thread about chopper shorting due to some misterious fault , i then had the idea to check the PWM IC's output signal to see if its correct or not .

As for the test , i use a 20 MHZ CRO to display the signal , but i get two different results , first one is when putting the probe positive pin only on the output pin without connecting the CRO GND to PCB ground , i get a signal as shown on the youtube video , second one is when i plug the gnd croc clip on the IC's GND pin i see on the CRO a 12 Vdc signal with no oscillation .

Right now am only focusing on how to read the PWM signal correctly , it would be really helpful to get that out of the way .

Thanks in advance .

 

[BGAmodz]

Hi again i have managed to get the right square wave signal display , it is a 30 uS high 35 uS low with a 15.38 Khz frequency according to this calculus : Fout = 1/(6.5Div*10uS/Div)

Is this a good frequency calculation ? because i read on the welder technical data that frequency is 60 Khz with 80% efficiency !!

 

[shortbus=]

You say you used 'pin 7', and called it an output. You might want to go back to the data sheet . On the 8pin version of the chip #7 is the Vcc of the chip. On the 14pin version #7 is the pin to control the freq/duty cycle, Rt/Ct.

 

[BGAmodz]

You say you used 'pin 7', and called it an output. You might want to go back to the data sheet . On the 8pin version of the chip #7 is the Vcc of the chip. On the 14pin version #7 is the pin to control the freq/duty cycle, Rt/Ct.

My bad i was making a mistake by checking the wrong pin , the square wave i got was on pin 6 the output pin . ( SOIC 8 IC) UC2845

So what i did is setting my oscilloscope on 10 uS/Div , 1 Volt/Div , and i got 3 divs high 3.5 divs low on the time axis .

Is the previous frequency calculation correct ? 15 Khz looks low isn't it ?

 

[rumpfy]

Sounds like progress.
Can you describe what you did to get a sensible looking waveform. maybe a picture of the result too?

 

[BGAmodz]

Sounds like progress.
Can you describe what you did to get a sensible looking waveform. maybe a picture of the result too?

Of course i will record a video and share it here .

 

[BGAmodz]

Here is a video using 20 uS/Div , s you can see there is about 3 Divs for the whole period

 

[rumpfy]

That looks OK.
I think you may be located in the US. If so, there are some differences between our power systems.
Outside the US, most power systems use what is called a 'Multiple Earthed Neutral'.
The 3 phase power system has a star connected secondary of 230 Volt per phase, and the star point is connected to ground at the substation. Each subscribers gets one, two or three phases plus a Neutral. At each subscribers switchboard, the neutral is connected to ground. Where you have a SMPS, the full wave rectifier produces a DC output which is NOT tied to ground anywhere; it floats above ground AND the power mains. So to measure waveforms in this type of equipment, you need a CRO which is also floating above ground. This is hard to do, so the best bet is to isolate the SMPS from the mains ground by means of an isolating transformer. Then the earth lead of the CRO test probe CRO can be connected to the DC common.
In the US, the power supply is supplied as a two phase (220 volt centre tapped) arrangement from the distribution supply transformer and my understanding is that the AC supply is not earthed anywhere.
I think your pics in post 4 are due to you have mains borne signals included with the input from the PCB.
You dont say what you did to get a sensible looking CRO display.
The display seems OK. Remember you have the MOSFETS disconnected (I think you said) so what the actual duty cycle should be, is not clear. Some of these switch mode IC's have a minimum (or maximum) duty cycle. If you have the mosfets disconnected, then there is probably no feedback from the voltage/current monitoring part of the power supply, so what you get may not have much relevance; except that it appears to be working. So now go to the next step.

 

[BGAmodz]

Thanks rumpfy

The display is really smooth at high frequencies , but it shows the correct waveform at least in every frequency .
I guess this IC is working as its supposed to be working according to that square wave output .

But do you have an idea on how to protect the new IGBTs from another short , some thing like checking the switching condition with another input voltage lower than the mains , and using some cheap FETs ?

 

[rumpfy]

If you go back to one of my earlier posts to you about using a couple of back to back transformers, you will have a very effective current limiting arrangement which will allow you to test the unit. If you have a short or something wrong, you will not get a high current from the mains which will destroy everything. Even a light globe (60 watt)in series with the mains supply lead will limit the current to a safe value.

 

[BGAmodz]

I guess you mean connecting two identical transformers like for example :
220V>24V then 24V >220 V ??