Announcement

Collapse
No announcement yet.

Harman Kardon AVR 3700 blown standby power supply

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    Harman Kardon AVR 3700 blown standby power supply

    Hey folks, I've come here because it looks like this group has helped a lot of people get Harmon Kardon AVR's back up and running.

    My AVR3700 had it's standby power supply blow out a few weeks ago. I have been trying to get it back up and running, but am pretty much striking out on everything I have tried. My background is just hobby level stuff, mostly dealing with analog electronics. I do a lot of work on car audio amps, but this is my first mains powered receiver that I have attempted to fix. The SMPS's in this receiver are quite a bit more complex compared to what I am used to.

    Below is a list of failed components that have been replaced:

    1) IC91- this is the standby supply SMPS driver, OB2358L, the original had the top of the case blown clean off.
    2) D914- MM1Z18, 18V zener on the supply side of the SMPS driver, looks to be there to clamp input voltage during a fault.
    3) D928- 20V zener, again, appears to be protection during fault condition.
    4) R928,967,972,970,971- These are current sense resistors for the driver. All were toasted, reading 1+ Meg, supposed to be 10 ohms
    5)D901/902- supply primary side of TF91, the standby transformer.

    I have checked every other cap, resistor, diode, transistor, opto-coupler, and the voltage regulators in the standby supply. They all test fine. The new OB2358L tries to start, but it never manages to take off, it resets and tries to start again and again.

    I'm close to my wits end with this thing. I am pretty stubborn when it comes to troubleshooting and repair, but I simply cannot figure out what is keeping this supply from working. Yesterday I decided to lift the cathode on D943 and D921 to see if there was a short somewhere downstream of the xformer. The supply still won't start. I did not really expect it to as D921 breaks the feedback loop, but I started with just D943 lifted and still no pulses. As a last ditch effort I pulled the standby transformer. I couldn't come up with a great way to test it, so I tried to use stepped down mains voltage to drive it. Transformer and leads got very hot and then blew a 2 amp fuse. No AC on the secondary side while I could test. The problem is, this transformer is probably designed for 50+ kHz and 60 Hz may have just saturated the core. By process of elimination I am thinking it is the xformer that is bad, but I don't know if parts are available, and I am also not 100% sure that I did not miss something else while troubleshooting.

    If anyone has any thoughts, please LMK. I'll try to attach the PDF of the SMPS in question, I have the full service manual hosted here, on my website.

    I apologize for the long post, but I wanted to try to be thorough explaining what I have checked so we would not have to reinvent the wheel...

    Thanks,
    Jason
    Attached Files

    #2
    Re: Harman Kardon AVR 3700 blown standby power supply

    The high fault current can damage the transformer's primary winding. I do a ring test.
    What do you notice that indicates the OB2358L is trying to start?
    Breaking the feedback loop puts the OB2358L at risk of failure. The only thing saving it would be the Over Voltage Protection on its VDD pin. Or Open Loop Protection.
    do you detect any fluctuating voltage on the secondary rails?
    Last edited by rievax_60; 07-16-2017, 09:56 AM.

    Comment


      #3
      Re: Harman Kardon AVR 3700 blown standby power supply

      Hi, thanks for your reply! I'm attaching some screen shots of the waveforms on each of the OB2358's pins. The name of each SS is the location they were taken. All were referenced to PG.

      I have a similar thread to this one going on DIYaudio.com, and Eva is trying to help me out, she has been more than excellent, but it is a struggle for me. Every little bit I know is self taught and she quickly gets over my head... I am hoping that I might find a source here that might have some other insight, or just be able to dumb things down a tiny bit for me

      Anyhow, she told me that my waveforms are indicative of this type of switcher trying and failing to start. I am supposed to capture the waveforms w/ a faster timebase on her request, but I have not had time to get them yet.

      So, long story short, I don't know what I don't know. I do a lot of work on car audio amplifiers. I am very familiar with their SMPS as well as the audio side of the amp. I know that this little switcher in the Harman Kardon amp should not be hard to sort out, but I am having real trouble wrapping my head around it.

      To your other questions, there is between .6 and 1.6Vdc on the cathode of D943 and D921 while the switcher is trying to start. When I lifted D921 the switcher appeared to survive. I knew there was risk, but I bought 10 on these and installed a socket just in case I pooch a few.

      Thanks for taking a look at this, I appreciate it... Now I'm off to figure out how to do a ring test.

      Sincerely,
      Jason
      Attached Files

      Comment


        #4
        Re: Harman Kardon AVR 3700 blown standby power supply

        What voltage do you have on D914? it should be at least 12 volts, if its not check R926 & R927 ( you may need to lift one end to check them properly), they could be open, you need about +12 volts to start the ic, once it starts, D911 supplies the voltage to keep it running

        I would also lift R935 to isolate that part of the circuit
        I don't think the transformer is bad.
        Last edited by R_J; 07-16-2017, 11:42 AM.

        Comment


          #5
          Re: Harman Kardon AVR 3700 blown standby power supply

          D914 has about 13Vdc when power is applied and it slowly falls to somewhere around 8Vdc then it goes back up and falls again. The first two waveforms above were taken at VDD and VDD-g on the OB switcher. R926/R927 are both good. I initially misread the schematic and thought they were supposed to be 1K2, I pulled them, saw they were actually 1.2M so I reinstalled them.

          R935 is a good idea, I have tested it, but I had not thought to lift it yet. I will give it a try. It will probably be this coming Friday before I have time off to work on it again. I'm stuck working a bunch of OT at my real job.

          Thank you!
          Jason

          Comment


            #6
            Re: Harman Kardon AVR 3700 blown standby power supply

            R935 supplies voltge to ic92 via q903/d918/d920 which is not need to run standby but could be loading down the voltage at D911, I would change c918 just to be sure its ok.
            If you have another zener for D914 it might be worth changing, if that voltage drops too low the ic just turns off (under voltage lockout)

            Comment


              #7
              Re: Harman Kardon AVR 3700 blown standby power supply

              Thanks RJ, I have replaced D914 already, it failed during the initial fault. Are you recommending I change it again just in case it has been damaged again? It tests fine as a normal diode, but I have not pulled it to reverse bias it and see if it's zener voltage is correct.

              C918 also tested fine out of circuit, it measured IIRC 9.87uF, ESR seemed reasonable as well, but i don't recall what it was and I don't have it noted down. I most likely have a suitable replacement in stock though, so I will swap it out to be safe.

              Thanks,
              Jason

              Comment


                #8
                Re: Harman Kardon AVR 3700 blown standby power supply

                Did you actually check the secondary for shorts? Like RJ already mentioned, IF C918, C921 or R922, R923 are bad, it wont start neither. You need 14.5V on VDD pin for the IC to start.

                Comment


                  #9
                  Re: Harman Kardon AVR 3700 blown standby power supply

                  C918 on the VDD is being charged to a high enough voltage for start up. Voltage rectified from the transformer winding by D911 is supposed to sustain VDD. This is not happening possibly due to an overload somewhere including shorted turns in the transformer.
                  Drive activity is actually stopping before VDD drops to the UVLO threshold. This is due to the 50ms OLP timeout. This point is not important.
                  If not done already, check R932. Remove R935 in case there is a problem in its area.

                  file:///C:/Users/A227/Downloads/OB2358L-On-Bright%20Electronics.pdf

                  Comment


                    #10
                    Re: Harman Kardon AVR 3700 blown standby power supply

                    The first few chopping cycles on the CS pin will show if there is a forward overload on the transformer. The timebase needs to be around 2us/div.

                    Comment


                      #11
                      Re: Harman Kardon AVR 3700 blown standby power supply

                      I'm in a very similar issue, with a 1700 model. We can't make it to have 5V standby voltage on the secondary side of the transformer.
                      I will get back to this, I'm like you. I won't give up. I had/have other stuff to work on right now.
                      The interesting thing is, that in the past, I have repaired a couple of these, only by replacing like two blown IC's a resistor and they were working like new.
                      I always had a (fairly big) resistor blown as well, around the blown ic.

                      Comment


                        #12
                        Re: Harman Kardon AVR 3700 blown standby power supply

                        Originally posted by CapLeaker View Post
                        Did you actually check the secondary for shorts? Like RJ already mentioned, IF C918, C921 or R922, R923 are bad, it wont start neither. You need 14.5V on VDD pin for the IC to start.
                        Checking the secondary- I have not, I do not know how at the moment. I tried the little transformer on 12Vac 60Hz, much to low frequency. I intend to drive it with a function generator when I have my next day off, but other than that I don't know how to do the "ring test" mentioned above, I'm still trying to figure that out. C918 tested good, I do not believe I tested C921. I will. R922 and R923 were tested and are good.

                        Originally posted by rievax_60 View Post
                        If not done already, check R932. Remove R935 in case there is a problem in its area.
                        I have checked both R932 and R935, both measured good in circuit. I will double check when I work on this Friday. As mentioned above, I did not think to lift R935, I will add that to my list.

                        Originally posted by rievax_60 View Post
                        The first few chopping cycles on the CS pin will show if there is a forward overload on the transformer. The timebase needs to be around 2us/div.
                        This may be really dense, but what does "forward overload" mean, does that mean that the primary side is drawing too much current? I will get more waveforms on Friday.

                        Originally posted by tibimakai View Post
                        I'm in a very similar issue, with a 1700 model. We can't make it to have 5V standby voltage on the secondary side of the transformer.
                        If I am not mistaken the AVR 1700, 2700, and 3700 all use the exact same setup for their low voltage supplies. The easy fix would be to find someone selling the PS board, my guess is they are on everyone's favorite auction site... I'm more than a little worried that I may have destroyed the little xformer with my crude attempt to test it. Hopefully it will have survived.

                        Thank you all for your help. I will try to work through the suggestions methodically and post results.

                        Jason

                        Comment


                          #13
                          Re: Harman Kardon AVR 3700 blown standby power supply

                          Originally posted by rievax_60 View Post
                          I do a ring test.
                          I think I understand. Pulse the coil and watch the back EMF oscillations on the scope. I'm still reading about it, apparently a lot of people build instruments to do it. Is it really as simple as momentarily applying DC to the windings and then watching the EMF decay on the scope?

                          Thanks again folks. I am learning a lot!

                          Sincerely,
                          Jason

                          Comment


                            #14
                            Re: Harman Kardon AVR 3700 blown standby power supply

                            Originally posted by jacampb2 View Post
                            I think I understand. Pulse the coil and watch the back EMF oscillations on the scope. I'm still reading about it, apparently a lot of people build instruments to do it. Is it really as simple as momentarily applying DC to the windings and then watching the EMF decay on the scope?

                            Thanks again folks. I am learning a lot!

                            Sincerely,
                            Jason
                            It depends on available equipment. I have an audio signal generator with a low impedance output. If I apply say a 1khz square wave to the transformer primary through a 10nf capacitor, I expect to see clean looking damped oscillations at every transition of the square wave. If I then set the generator's frequency to match the frequency of the damped oscillations, the series resonance will cause a very large signal to develop across the primary.
                            The 1khz test signal from the oscilloscope can be used. Because it will be high impedance, the oscillation will be more damped. Using a lower value capacitor would help make the result more meaningful.
                            http://www.brats-qth.org/training/ad...s/scopeani.gif

                            Comment


                              #15
                              Re: Harman Kardon AVR 3700 blown standby power supply

                              Whenever the MOSFET turns on, the HVDC is applied to the transformer's primary. This is called the On-Time, or Forward Stroke. The voltage that is induced on other windings should not cause any current flow because the polarity should cause diodes to be reversed biased.
                              The only current through the primary and the MOSFET should be an upward ramping of primary inductive current.
                              At least for the very first time the MOSFET turns on at start up, the current should ramp up from zero. This current causes a proportional voltage signal across the CS resistors. The IC senses this voltage and is supposed turn off the MOSFET at 0.8v.
                              If there is a forward overload caused by a shorted diode(snubber diode included) or shorted windings, the current will abruptly shoot up to a higher than usual current whenever the MOSFET turns on.
                              Last edited by rievax_60; 07-17-2017, 11:23 PM.

                              Comment


                                #16
                                Re: Harman Kardon AVR 3700 blown standby power supply

                                That makes sense, thank you for the clear explanation. I read a lot about about these small switchers on breaks yesterday and I think I am starting to get a better handle on the operation.

                                A question that is a bit off the topic of the repair, and more about the design decisions in this PS-- Do you have any idea why the manufacturer would design a more complex switcher when they could have easily gotten away with a small mains transformer, a bridge rectifier, a little bit of filtering and a linear regulator? Is it really that much cheaper to go to a switcher? I realize the transformer is the most expensive part, and that the high frequency switchers can use a much smaller and likely cheaper transformer, but it seems like the road that they took on this one is unnecessarily complex.

                                Thank you again for taking your time to help.

                                Sincerely,
                                Jason

                                Comment


                                  #17
                                  Re: Harman Kardon AVR 3700 blown standby power supply

                                  In the past, a small mains transformer is all that was needed just to keep the control circuitry alive. I would suspect that there is more going on inside of modern receivers while in standby needing more power.
                                  Last edited by rievax_60; 07-18-2017, 04:44 PM.

                                  Comment


                                    #18
                                    Re: Harman Kardon AVR 3700 blown standby power supply

                                    Small, cheap, low power, high efficiency. Pretty much everything has switching power supplies nowadays so parts are always readily available.

                                    They are also flexible (output voltage/current can be adjusted to a degree without changing the transformer windings at all) and are able to do universal 100-240VAC input so there is no need for different parts for different countries.

                                    That makes a lot of reasons, really.
                                    Originally posted by PeteS in CA
                                    Remember that by the time consequences of a short-sighted decision are experienced, the idiot who made the bad decision may have already been promoted or moved on to a better job at another company.
                                    A working TV? How boring!

                                    Comment


                                      #19
                                      Re: Harman Kardon AVR 3700 blown standby power supply

                                      Hi, I did not get as much time as I had hoped today. My wife took my girls school shopping, so my son stayed with me and wanted to build lego robots...

                                      Anyhow, I talked him in to hanging out in the shop for a little while. I tested both R932 and R935, they are ok. I lifted R935, and it is still no go. I reinstalled the transformer and I tried to get more detailed waveforms. I can try again if I still did not get enough detail. They are all still referenced to PG, but I adjusted the zero level to get as much detail as possible in the screen shot. I assume that the waveform detail is more important at this point than knowing the exact voltage of the peaks. The waveforms on Vdd-g and Vdd both had zero set at approximately -9V. The waveforms captured on the drain terminals was with the zero at -330V.

                                      Current Sense


                                      Feedback


                                      Drain


                                      Drain 5ms


                                      Hopefully this will give the information needed. If not I should have the day off sunday and with any luck will be able to get some time alone to do a little better job getting the new traces.

                                      Thanks,
                                      Jason
                                      Attached Files

                                      Comment


                                        #20
                                        Re: Harman Kardon AVR 3700 blown standby power supply

                                        The CS and Drain waveforms show that the primary winding looks like a short circuit. Time to ring test the transformer.

                                        Comment

                                        Working...
                                        X