Hi, Following the recent IR2110 failure, I wanted to try out with a square wave instead of a sine wave, as you can see in the picture. http://www.flickr.com/photos/79743192@N07/8171444219/in/photostream http://www.flickr.com/photos/79743192@N07/8171472476/in/photostream I am able to view a square wave at the OC1 and OC2 pins of the dsPIC. http://www.flickr.com/photos/79743192@N07/8184610634/in/photostream The only difference in the current setup is that, the IR2110 board has been disconnected and the FET's connected directly to the dsPIC. The dsPIC as you can see is properly decoupled and also has a 1500uF/6.3V electrolytic capacitor on the 5V logic supply to the rail. The dsPIC board is powered from a laptop adapter like SMPS power adapter which provides +5V to the dsPIC . The FET's and the transformer are powered from a SMF batter providing 12V. The power to the FET's and transformer were actually connected, while I connected the FET gates to the dsPIC board. I could hear the faint chatte= r of a square wave from the ferrite core (No Load). Hearing the faint chatter and realizing that probably things were okay and that the power was accidentally ON, I switched OFF the power to the FET's while the power to the dsPIC board still ON. Seeing nothing unusual, I switched the power ON again, but this time there was a loud blast from the 1500uF/6.3V electrolytic capacitor. There was paper and light electrolytic all over the place. After this switched OFF everything, inspected the damage: What remained of the capacitor was only it's rubber base, nothing more. On further inspection, the dsPIC also looked dead. No issues with the FET's or the transformer. At this stage, I wonder what could have caused the 6.3V electrolytic to burst on the 5V logic line. Maybe a better question would be, as to what happened on the 5V logic supply line ? Any ideas, folks ? Regards, Manu On Sat, Nov 10, 2012 at 3:07 PM, Manu Abraham wrot= e: > Hi, > > Yes, I follow what you meant by that the IR2110 is not an analog componen= t > but a device that drives a MOSFET in either Cutoff or Saturation regions, > which is what I am using it for. > > I am feeding a PWM signal from the PIC to the IR2110. The PWM signal > (carrier has a frequency of 102.4kHz) and the modulated signal has a > frequency of 50Hz. Sorry, that I hadn't made this very clear. > > With SPWM, the duty cycle is varied continuously, following a sinusoid > with a small dead time with a modulation index of 80% > > You were right that the Soundcard oscilloscope was doing some kind of > filtering. (looked it up and found it to be Hanning). I turned it off, bu= t the > waveforms looked the same. The issue that the PWM signal can't be > seen is due to the bandwidth of the scope as you exactly described. > > > I wanted to test again, with more pictures and details of the waveforms. > > Weird part, I can't see the same waveform distortion that I had > yesterday night. > > > The only difference from yesterday night's run and today is that: > Somewhere on the internet, found that it would be better, if the ferrite > cores are slightly separated with a small piece of tape to avoid core > saturation and have tried this today, as opposed to yesterday night. > (Thinking that likely the ferrite core is getting saturated ?) > > > Pictures of the setup, also with Power > http://www.flickr.com/photos/79743192@N07/8171444219/in/photostream > http://www.flickr.com/photos/79743192@N07/8171472704/in/photostream > > with Power > http://www.flickr.com/photos/79743192@N07/8171472476/in/photostream > > Interesting thing is that today I have had different results as compared > to yesterday night. I am still clueless as to what might be happening. > > OC1 Signal from the dsPIC: > http://www.flickr.com/photos/79743192@N07/8171455317/in/photostream > Signal zoomed at 0 degrees > http://www.flickr.com/photos/79743192@N07/8171455571/in/photostream > Signal zoomed at 90 degrees > http://www.flickr.com/photos/79743192@N07/8171455471/in/photostream > Signal zoomed at 180 degrees > http://www.flickr.com/photos/79743192@N07/8171483210/in/photostream > > > Ho signal output from the IR2110, without the FETs > http://www.flickr.com/photos/79743192@N07/8171482839/in/photostream > zoomed at 0 degrees > http://www.flickr.com/photos/79743192@N07/8171482767/in/photostream > zoomed at 90 degrees > http://www.flickr.com/photos/79743192@N07/8171510934/in/photostream > zoomed at 180 degrees > http://www.flickr.com/photos/79743192@N07/8171482577/in/photostream > > > With a 60W Resistive Load on the Transformer: > http://www.flickr.com/photos/79743192@N07/8171490216/in/photostream > Signal zoomed at 0 degrees > http://www.flickr.com/photos/79743192@N07/8171462403/in/photostream > Signal zoomed at 90 degrees > http://www.flickr.com/photos/79743192@N07/8171490300/in/photostream > > Note 1: Noticed that the Ho signal is considerably lower in amplitude, > compared to OC1. Looks weird. > > Note 2: Can hear a low chattering noise in the Ferrite Transformer. > > > Ho signal, No Load condition: > http://www.flickr.com/photos/79743192@N07/8171518938/in/photostream > Signal zoomed at 0 degrees > http://www.flickr.com/photos/79743192@N07/8171490381/in/photostream > Signal zoomed at 90 degrees > http://www.flickr.com/photos/79743192@N07/8171518746/in/photostream > Signal zoomed at 180 degrees > http://www.flickr.com/photos/79743192@N07/8171490185/in/photostream > > Note 3: The chattering noise from the ferrite core became slightly louder > > Note 4: The wires connecting the IR2110 and the FET gates are getting > warmer and warmer as time progresses. > > Note 5: At some point, the chattering noise becomes just as faint as with > the load. > > Connected the same 60W lamp Load to see whether any issue at all. > The lamp glowed at half the brightness. > > Note 6: The FET's are also warmer now, as in the previous case, they > weren't warm at all. > > Note 7: Wanted to repeat the 60W loaded condition, after FET's cooled > and thus switched OFF. > > Note 8: Switching ON after 5 minutes, the Ferrite transformer was still > making the lower chattering noise, but no output on the load (load > doesn't look lit up) Wanted to look at the waveforms in this situation. > The Lo signal from the IR2110, which was looking very similar to Ho, > was now looking different. > Lo signal at the FET Gate: > http://www.flickr.com/photos/79743192@N07/8171507765/in/photostream > Lo signal from the IR2110: > http://www.flickr.com/photos/79743192@N07/8171507845/in/photostream > OC2 signal from the dsPIC which is responsible for Lin and thus Lo for > the IR2110 > http://www.flickr.com/photos/79743192@N07/8171507927/in/photostream > > Frustrated by the results, wanted to compare both Ho and Lo wrt > OC1 and OC2, with the FEt's disconnected: > OC1 output from the PIC: > http://www.flickr.com/photos/79743192@N07/8171547854/in/photostream > OC2 output from the PIC: > http://www.flickr.com/photos/79743192@N07/8171547780/in/photostream > Lo output from the IR2110: > http://www.flickr.com/photos/79743192@N07/8171518935/in/photostream > Ho output from the IR2110: > http://www.flickr.com/photos/79743192@N07/8171548068/in/photostream > > > By now, I am really knotted up, looks like the IR2110 is dead/damaged ? > > > Thanks, > > Manu > > > On Sat, Nov 10, 2012 at 11:00 AM, Sean Breheny wrote: >> I must be missing something here. The IR2110 is not an analog >> component - it is intended to turn its outputs fully on or fully off >> at all times. >> >> Are you feeding PWM from your PIC to the IR2110? If so, you should be >> seeing a PWM waveform on the FET gates, not a sinusoid. >> >> By SPWM do you mean PWM whose duty cycle varies sinusoidally? What is >> your PWM frequency? >> >> I think that perhaps your soundcard-based scope is averaging-out a >> huge amount of detail that you really need to see here. I don't think >> the gate waveforms really look sinusoidal at all, if you had enough >> bandwidth to see them in detail. >> >> Sean >> >> >> On Fri, Nov 9, 2012 at 9:36 AM, Manu Abraham wr= ote: >>> Hi, >>> >>> I have been trying to drive an IRFP460 MOSFET in Push Pull mode with >>> SPWM. I am feeding each half (0-180, 180-360) of the sine to each of th= e >>> MOSFETs. I was expecting to get exactly the same signal at each of the >>> Gate inputs of the MOSFET's, but it looks what I do get at the gate is >>> different from the output of the microcontroller (a dsPIC30F2010). >>> >>> The following links to the waveform, using a soundcard on a PC. >>> >>> The dsPIC outputs 0-180 degrees of the sine on Output Compare 1 and >>> 180-360 degrees on Output Compare 2. >>> >>> >>> Schematic: >>> http://www.flickr.com/photos/79743192@N07/8169359354/in/photostream >>> >>> Actual Output Compare Signal from the dspic30F2010: >>> http://www.flickr.com/photos/79743192@N07/8169331015/in/photostream >>> >>> MOSFET gate signal with No Load: >>> http://www.flickr.com/photos/79743192@N07/8169331109/in/photostream >>> >>> MOSFET gate signal with a resistive load (incandescent lamp): >>> http://www.flickr.com/photos/79743192@N07/8169331015/in/photostream >>> >>> >>> Any idea why the signal to the MOSFET gate is distorted at the peak >>> of the sine, in the No Load, or with a resistive Load situation (while >>> the actual >>> signal before the IR2110 is not so) ? If I disconnect the gates of the = MOSFET's >>> from the outputs of the IR2110, then the exact signal can be seen at th= e driver >>> outputs. >>> >>> Thanks in advance for any pointers. >>> >>> Regards, >>> Manu >>> -- >>> http://www.piclist.com PIC/SX FAQ & list archive >>> View/change your membership options at >>> http://mailman.mit.edu/mailman/listinfo/piclist >> -- >> http://www.piclist.com PIC/SX FAQ & list archive >> View/change your membership options at >> http://mailman.mit.edu/mailman/listinfo/piclist -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .