Announcement

Collapse
No announcement yet.

Warning about amps with switching power supplies

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

  • Warning about amps with switching power supplies

    I just finished repairing a Carvin SX200 that was a very interesting repair.

    It's possible that I would pass on an amp where all the PCB's, including the power supply and output boards are SMT, but this one had symptoms that seemed to indicate nothing worse than intermittent FX loop jacks and probably an intermittent contact in the headphone jack. I asked the guy if he ever used the headphone jack and would he care if I just jumpered across the interruptor switch to prevent future problems. He said go right ahead.

    Fixed up the jack problems, and along the way marveled at the small size of an amp capable of cranking more than 100W before clipping. Son of a gun, it had a switch-mode power supply. Good thing that wasn't buggered, especially with all the surface mount stuff on that main PCB.

    Unfortunately, final testing showed the amp exhibiting an intermittent high frequency oscillation that seemed to come and go when I wiggled an interconnect cable.

    Long story short, it turned out that the VIPER100 switch mode regulator's legs were almost broken off, and in fact did come completely off with very little prodding. I think the howl was due to the switching frequency getting superimposed on the audio. Also replaced the +15V regulator which was hanging by a thread. I even noticed that one of the SMT zeners had come adrift and instead of tacking on a leaded component, I got it back in place with my long nose pliers and the same soldering iron I use on tube sockets.

    Felt like a real swingin' dick I did. Fixed a switch mode power supply and did SMT rework with my regular amp bench tools!

    Until I attempted to fire everything back up in the usual "power supply issue" mode, by bringing it up on the variac, and then on the lightbulb limiter.

    Drawing lots of current, brightly lit bulb every time, but every voltage was what it should have been, given the limited input voltage. No indication of shorted output transistors, drivers, anything like that.

    Bringing the variac up a little more resulted in smoke coming off an unidentified component that I now think is likely a thermal resistor. It's in-line with the main DC coming off the rectifier bridge.

    Finally decided to chance just plugging the thing in and riding the power switch, and of course it worked fine.

    I guess switch-mode power supplies behave that way in brownout conditions. Really never worked with them before, but techie beware: you can't use your usual soft-start techniques with a switcher!

    In 5 years of taking in amps for repair, this is the first such amp I've encountered, and there are few posts on this forum talking about amps with switching power supplies, so it's probably not a huge issue, but man!
    -Erik
    Euthymia Electronics
    Alameda, CA USA
    Sanborn Farallon Amplifier

  • #2
    Off-line switchmode power supplies are a whole other ballgame to the usual troubleshooting jobs. I've had reasonable success fixing broken switchers in the past, but I've also (rarely!) come across ones that have design flaws that make them unreliable.

    This usually happens in smaller volume products, where the PS may have been a custom job contracted out to a little design house, or nightmare scenario, the engineer who did the rest of the circuitry decides to have a crack at designing the switcher too! After all, how hard can it be... bang... sizzle...

    I know a guy who used to design them for a living, and if I get really stuck troubleshooting one I call him. If you mention the VIPER or Topswitch you can hear him groan.
    "Enzo, I see that you replied parasitic oscillations. Is that a hypothesis? Or is that your amazing metal band I should check out?"

    Comment


    • #3
      I've been working on an old ADA Microtube 200 --- the switching power supply is kicking my tech butt (and spurring some old fashioned learning).

      In the last go around, the power MOSFETs seem to have shorted and blows the fuse (it's either that, or the custom transformer has shorted). Before that, the diodes in the compare circuit were bad, and replacements would gradually bad.
      See the birth of a 2-watt tube guitar amp - the "Dyno Tweed"
      http://www.naturdoctor.com/Chapters/Amps/DynoTweed.html

      Comment


      • #4
        Yeah,
        Also switch mode supplies don't usually favor well with the usual 'bringing up on the variac' approach since they need to oscillate to function. They either won't fire until they reach a design minimum AC (which is enough for them to blow up anyway) or just behave with instability which can also make them blow up. They pretty much either work, blow up, or do nothing & then blow up ;-]

        You pretty much have to pray alot & plug-er-on in. I won't touch high power stuff with switchers in them anymore. glen

        Comment


        • #5
          Boy! If that doesn't remind me of the nightmares I had with the switching supply series that Gallien Krueger put out in the 90's (the "micro" series). What REALLY made it bad was that the shop I was working for at the time was doing warranty work on them.They never really perfected those (as far as I know, at least), and with all the units that would keep breaking down, over and over, that's one of the very few lines of amps that I just refuse to take in for repair any longer.
          Mac/Amps
          "preserving the classics"
          Chicago, Il., USA
          (773) 283-1217
          (cell) (847) 772-2979
          Now back on Chicago's NW side in Jefferson Park!
          www.mac4amps.com

          Comment


          • #6
            I worked with switching power supplies for years. They generally regulate by varying the chipper frequency around the Q point of a smallish transformer ( the transformer is small because the frequency is very much higher than line). To do this they require feedback usually through an opto-isolator. In my experience, most smps failures were due to bad caps or a reference diode, most of the time a bipolar electrolytic, but anything wrong with that isolated feedback loop can cause all sorts of grief. They also must have a load. If you know how they work or have a manual they can usually be jumpered into a low current mode to troubleshoot. But yes, they can be a pain if it isn't a "known" problem and you don't understand them.

            Comment


            • #7
              Originally posted by mac1amps View Post
              Boy! If that doesn't remind me of the nightmares I had with the switching supply series that Gallien Krueger put out in the 90's (the "micro" series). What REALLY made it bad was that the shop I was working for at the time was doing warranty work on them.They never really perfected those (as far as I know, at least), and with all the units that would keep breaking down, over and over, that's one of the very few lines of amps that I just refuse to take in for repair any longer.
              just as point of info, not sure if they still do it, but GK used to sell a linear supply "upgrade kit" for their MB series amps. I have replaced about a half dozen of those over the years...the kit includes a new cab with a transformer mounted inside, and a replacement circuit board...it's basically a drop fit. I have a GK MB something or other coming into my shop in a few days and I think, from what the owner described to me over the phone, that it will be needing that power supply upgrade. I guess I'll find out if GK still sells them if I need one because, like you, if it has crapped out, I am not going to try and fix it.

              Comment


              • #8
                But I though Mackie's Hotwire was revolutionary in its use of SMPS?

                Comment


                • #9
                  They don't regulate by varying the frequency, it's the duty cycle.

                  Problems with SMPS repair are generally down to some of the following:

                  The feedback loop, as others have pointed out, the SMPS will do whatever it can to bring its output voltage to the setpoint, even if that means turning on full into an internal short and self-destructing! Better quality units nowadays have a current limiting circuit to protect the chopper transistors, but that circuit can still go wrong.

                  The feedback loop also means that the line current draw goes up as line voltage goes down. Not only does this make them hard to test on a variac/light bulb limiter, it also means that they present a negative impedance to the line. If the line has a lot of sag (as with the light bulb) the negative impedance can cause oscillation.

                  To add insult to injury, the control electronics are often powered off the switcher's own output as an economy measure. But then how does it start in the first place?! This catch-22 is resolved by the "kickstart circuit": it charges an electrolytic cap off the rectified line through a large dropping resistor, then dumps the cap into the controller chip to kick it into life. Again this foils your attempts to bring it up on a variac, because it needs a fair amount of line voltage before it'll "kick" the chip hard enough for a successful start.

                  So usually nothing happens until the variac is up quite high, then suddenly the kickstart fires and all the components you just replaced instantly blow out again. Don't ask me how I know this! While we're on the subject, switcher blowouts can be violent enough that you'll wish you'd worn eye protection. I remember once getting hit in the forehead by a chunk of MOSFET.

                  Some old SMPS, especially the bigger sizes, had a tiny iron-cored transformer to power the control circuitry straight off the line. I always hated the Rube Goldberg/Heath-Robinson kickstart circuits, so I thought this was a nice touch.

                  In desperate cases, you can power the control circuitry up yourself from a bench PS, bypassing the kickstart. If the output is unloaded, you should then be able to bring up the power circuitry on a variac. But beware: the control circuit's ground rail is often hot. If you don't understand this concept, and at least own an isolation transformer and understand what it does, then probably best to leave SMPS alone.

                  Oh, and finally, if you just can't get a SMPS to work reliably, it's not unknown (though very rare) for them to have design flaws. Especially the kind of short-run custom ones you might find in the music electronics market. That Gallien-Krueger upgrade kit springs to mind :-O
                  Last edited by Steve Conner; 05-12-2009, 10:20 AM.
                  "Enzo, I see that you replied parasitic oscillations. Is that a hypothesis? Or is that your amazing metal band I should check out?"

                  Comment


                  • #10
                    I used to work on a ton of switchers in another industry, but of the ones I encounter in pro audio, I find those large value resistorsthat trickle into the kick start sometimes go open. SO it won;t start.

                    Also, they tend to have all manner of self protection, so if a secondary rectifier is shorted, many of them simply shut themselves down instead of blowing up. Same thng they do when the circuits they power are shorted, such as when output transistors short.

                    Amen to being careful when poweruing the control circuits with a bench supply. They may run on 12v, but the "ground" of that 12v supply is the -170VDC rail - rectified right off the mains.
                    Education is what you're left with after you have forgotten what you have learned.

                    Comment


                    • #11
                      It's funny because I work on alot of switching supplies for my real job and I thought I was pretty good at it until some of these audio amps with switchers started kicking my butt!

                      Comment


                      • #12
                        Originally posted by tedmich View Post
                        But I though Mackie's Hotwire was revolutionary in its use of SMPS?
                        From Mackie's website:

                        "Greg's solution was to create the first modern, lightweight, switching power supply with the classic, purposeful 'limitations' of a well-designed tube amp power supply...."

                        Speculation/translation: he powers his MOSFET hybrid with a switcher that has some "sag" components tacked on.

                        Use of a switcher is not new in the guitar amp world. Mackie's implementation seems to have some new twists.
                        -Erik
                        Euthymia Electronics
                        Alameda, CA USA
                        Sanborn Farallon Amplifier

                        Comment


                        • #13
                          Originally posted by Enzo View Post
                          I used to work on a ton of switchers in another industry, but of the ones I encounter in pro audio, I find those large value resistorsthat trickle into the kick start sometimes go open. SO it won;t start.
                          There's nothing like troubleshooting for an open 1 meg resistor.

                          Comment


                          • #14
                            Or a pair of 560k in series. Or even a pair of 750k.
                            Education is what you're left with after you have forgotten what you have learned.

                            Comment

                            Working...
                            X