Announcement

Collapse
No announcement yet.

Dynasty 200SD low buss voltage

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

  • #16
    Thanks Noel. Yes, this is clearly an out out control hobby... Hopefully I get to a point that is helpful for others.

    Comment


    • #17
      I can see I am at the still eating dirt stage.

      Comment


      • #18
        Yup, I am not an electrical engineer, or an electronics tech, so this is a steep learning curve that's only taken about a couple dozen broken welders and some broken plasma cutters to get to confusing myself and others...

        Comment


        • #19
          Well, I received and replaced all of the parts on the Maxstar today - the two buss capacitors, the inrush circuit bypass relay, and the four 150k resistors. After reconfirming I didn't miss something in the pre-power checks, cinched everything up and turned on the power. There is a faint low pitched buzz on top of the normal high pitch hum from the inverter. All functions work and I ran a half dozen 6010 1/8" rods back to back to see how it went. The only thing concerning me now is the fan never turned on. The fan resistance checks out at 1.5M and the negative lead on RC7 (PC1) stayed open. I know it's fan on demand, but would expect it.to demand the fan at some point. When I pulled the case, nothing was beyond warm.
          Click image for larger version  Name:	IMG_20190928_144625253.jpg Views:	29 Size:	27.4 KB ID:	602550 original capacitors
          Click image for larger version  Name:	IMG_20190928_172324229.jpg Views:	24 Size:	63.4 KB ID:	602551 replacement capacitors and new orange resistors installed with 1/8" standoff from the PCB
          Click image for larger version  Name:	IMG_20190928_192943749.jpg Views:	27 Size:	34.8 KB ID:	602553 case removed after burning some rods, checked the fan voltages
          ​​​​
          Click image for larger version  Name:	IMG_20190928_181946228.jpg Views:	26 Size:	30.8 KB ID:	602552 I didn't fully seat one of the capacitors, but it seems ok.
          Last edited by jjohn76; 10-07-2019, 10:25 AM.

          Comment


          • #20
            For anyone that hears a low pitched buzzing sound, particularly from the lower half (transformer output diode area), it could be that your output diodes are unintentionally hard switching... That appears to be the case on the Maxstar at least. I found a break in connection between the output diode Snubber (resistor capacitor snubber shown in the picture). I checked the output rectifier again, as well as the active clamp (diode check across all four leads, each way). There's not much lead left at all on the capacitor to resolder the connection, so we'll see how to get it going...
            Click image for larger version

Name:	IMG_20190929_082900821.jpg
Views:	79
Size:	61.1 KB
ID:	602566 This is a picture of the underside, with the output rectifier bridge on the right. The gold resistor and blue capacitor soften the hard switches when the output rectifier converts the AC voltage from the transformer to DC).

            Click image for larger version

Name:	IMG_20190929_082936187.jpg
Views:	64
Size:	39.0 KB
ID:	602567

            Comment


            • #21
              Update - I had enough capacitor lead left to solder the Snubber back together. I just ran a half dozen 7018 1/8" rods on a scrap piece (would show the piece, but I welded over test beads from a Haas-Kamp project it looks terrible). The fan kicked on a couple times and no issues.

              My only concern is a buzz/squeal when the machine is on when not loaded. There is a quiet buzz coming from both input inductors (boost inductor and snubber inductor) that turns to a quiet squeal when I close the contactor, and even higher pitched to unnoticeable once I load it, either crowbarring the output (3' steel piece between leads, run amperage from 1 to 200) or running rods. I haven't put any scope traces across the DC bus or inductors (yes, I use differential probes to avoid electrocution or magic smoke from my oscilloscope), not sure if I need to or would know what the switching spikes should look like. The output waveforms match the ones in the technical manual. Any thoughts?

              Here's what it looks like back together. I think I am going to leave it looking like this, except cross out the yellow N and replace it with a K, for Kinda effing good...
              Click image for larger version

Name:	15697858572721495587048.jpg
Views:	72
Size:	73.3 KB
ID:	602576
              Last edited by jjohn76; 09-29-2019, 02:42 PM.

              Comment


              • #22
                So with the Maxstar 200DX good enough for my buddy to abuse/test, I moved on to the Dynasty 200SD. I still don't know how deep the problems run, but just started down the control boards' power supply chain. Before starting that thread, I wanted to compare the boostrap supply resistor arrays between the earlier Maxstar 200 (above) and the later model Dynasty 200. Besides the additional standoff/spacing between the board and resistors I mentioned a couple posts ago, the newer board also has a lot more copper around the resistor arrays pads. Clearly those resistors were putting too much heat into the board, and Miller's design solution was to use more of the copper trace to absorb some of the heat. These 3W resistors are very small compared to other resistors of the same power ratings, I guess it's because they conduct a lot of that heat away from the body into the copper traces. This same issue exists on the clamp board in my Dynasty DX, and I don't remember how much time I spent (wasted) trying to see what was causing the over current through the resistors. Turns out to just be bad thermal management design from my perspective...

                Click image for larger version  Name:	IMG_20191005_065751968.jpg Views:	15 Size:	29.7 KB ID:	602935 Here you can see the shadow of the large copper pads underneath the bootstrap power supply resistor arrays.

                Click image for larger version  Name:	IMG_20191005_065655777.jpg Views:	14 Size:	27.8 KB ID:	602936 Here is the underside of the board showing those same large pads...
                Last edited by jjohn76; 10-07-2019, 10:27 AM.

                Comment


                • #23
                  Updates of the issues I found in the bootstrap power supply and the control boards' power supply. This board is completely removed from the welder and no mains power is used directly because there's no need to risk electrocution.

                  The bootstrap power supply provides 15V to the controller that manages the 30V on the control boards'power supply (this 30V feeds every board, directly or indirectly). It does this by connecting a 150k ohm resistor array (the one in the previous post) in series with a 15V zener diode across the DC buss (there is a little yellow filter cap across the zener diode to smooth out the 15V). So long as the buss voltage is above ~17V DC, this circuit provides 15V. Everytime the power switch goes on, the DC buss charges to somewhere between 162V DC and 325V DC (the peak voltage of the incoming 120VAC RMS and 240 VAC RMS).

                  My zener diode was bad, measured by a multimeter diode tester and also confirmed by putting 30V DC across the DC buss from my power supply. Replacing this zener diode now has 15V going to the control boards' 30V power supply controller on PC10.

                  Click image for larger version  Name:	IMG_20191005_084118610.jpg Views:	0 Size:	56.0 KB ID:	602938 You can see the carnage on this board, with the big hole in IGBT Q11 on the left and R6 in the center. Those are part of the 30V power supply and I'll fix later. Right now, I am working on the 15V bootstrap power supply, which relies on D2 (15V zener diode) and C2 (.1uF filter capacitor) to give a stable 15V to the 30V power supply controller.
                  Click image for larger version  Name:	IMG_20191005_084940093.jpg Views:	0 Size:	42.2 KB ID:	602939 There are a couple ways to check zener D2. Using the diode tester on your multimeter lets you know whether or not it is shorted. I didn't get a picture of that (not enough hands), but here's how I also test it. I put 24V across the DC buss (+ fed into the top portion of the resistor array on yellow clip, - connected to pin RC2-2 on hard to see black clip) on PC2. Anything other than 15V +/- tolerance listed in the datasheet and it's not working right. This one was shorted, so it showed 0V.
                  Click image for larger version  Name:	IMG_20191005_100404593.jpg Views:	0 Size:	57.9 KB ID:	602940 I removed the zener, checked to make sure the board still wasn't shorted across the test points (that would have indicated C2 was likely bad too). Then put the replacement back in. It now shows ~15V across the zener, so we're good to go. I'll check later to make sure the cause of the bad zener wasn't something on board PC10.
                  Last edited by jjohn76; 10-05-2019, 01:22 PM.

                  Comment


                  • #24
                    The 30V regulated power supply is a buck converter directly off the bus voltage. It's common is free floating. The buck circuit consists of IGBT Q11(PC2), inductor L3 (PC2), diode D4 (PC2), and capacitor C7 (PC2). The controller takes up the front quarter of PC10, using a quad nand gate (CD4093BCN) to switch the Q11 gate on and off using current and voltage feedback. Resistor R4 on PC2 is the gate drive resistor, and the gate is biased to the Q11 emitter through R5 (200K ohm) on PC2 to protect Q11 in the case there is no signal from the concontroller. R6 on PC2 is the current feedback resistor. I had trouble figuring out the voltage control until realizing the voltage feedback comes through connector RC1 instead of RC2. It references a 30V zener diode to signal the controller through an optocoupler (hysteresis control). Here's the schematic of the buck converter:
                    Click image for larger version  Name:	IMG_20191005_204734520.jpg Views:	6 Size:	42.1 KB ID:	602978

                    On this Dynasty, gate resistor R4 was open (not visibly damaged though), Q11 was fried, the current sense resistor R6 was toast, the totem pole gate drive transistors Q1 and Q2 were overdrawing current, and the quad nand gate U1 was also overdrawing current. The current feedback circuit is functioning, so I am not sure why Q11 went, but as soon as it did, it probably shorted, which saturated L3 and toasted both R4 and R6 with excessive voltage. No CD4093BCNs on hand, so waiting until next weekend to make progress on PC10.
                    Last edited by jjohn76; 10-06-2019, 11:35 AM.

                    Comment


                    • #25
                      Originally posted by jjohn76 View Post
                      Update -, not sure if I need to or would know what the switching spikes should look like. The output waveforms match the ones in the technical manual. Any thoughts?

                      Here's what it looks like back together. I think I am going to leave it looking like this, except cross out the yellow N and replace it with a K, for Kinda effing good...
                      [ATTACH=JSON]{"data-align":"none","data-size":"small","data-attachmentid":602576**[/ATTACH]
                      Once again a conversation above my pay grid. But the switching spike mention? Are you talking about the reading the heart beating on the screen and trying to identify anomalies? I do have some thoughts on that but my basis for them is diagnosing with automotive scan tool technology. And that's above my grid as well.

                      That said, I could see voltage causing the height of the spike, the collapse being a width, length between a build up, but if it's beating and your getting the wave forms, happy freaking camper I'd say.

                      I'd do a big K. Well done.

                      Comment


                      • #26
                        https://control.com/thread/1026227984#1026227984

                        hysteresis control...doesn't that sound like a serious male medical condition? Seems it's not? Who knew? Not me.

                        Interesting. Very interesting. The more a guy reads the more it starts to make a bit of sense.

                        Comment


                        • #27
                          Noel, yes everytime those transistors turn on and off, it makes a voltage spikes, just like contactors arc. Too high of spikes and the voltage can be higher than the components can handle (transistors and capacitors mostly). The more welding power through it, the bigger the spikes. To lessen the spikes, the designers put snubbers across the switches, which absorb some of the shock from suddenly opening and closing those switches. I just don't know how much this welder should have...

                          Comment


                          • #28
                            I think you’re kickin arse fixing those fancy machines. I also appreciate the work explaining stuff to us dummies. I would’ve almost certainly smashed it with a hammer by now.

                            Comment


                            • #29
                              Thank you Ryan. I am just glad it doesn't have a Miller MVP plug...
                              Last edited by jjohn76; 10-06-2019, 11:36 AM.

                              Comment


                              • #30
                                Waiting for a couple of components to arrive (thanks Franz, had you not mentioned Murphy earlier, I would have been good....) I moved on to one of the two power supplies fed by the buck converter. The first one to tackle is the regulated power supply and boost/buck converter on PC10. The schematic for this beast is one of my first posts when dealing with the last Dynasty. Because I had a current limiter on my power supply, I just hooked up 30V to pins RC1-5 (+30V) and RC1-3 (-Bus). My picture shows it hooked up to the -Bus ground plane and the input lead for voltage regulator VR4. In the picture below, you can see how it isn't reaching 30V and hits the current limit of 200mA, so something isn't good.
                                Click image for larger version  Name:	IMG_20191006_150922175.jpg Views:	1 Size:	76.5 KB ID:	603012

                                I cheated with a Thermal camera, that showed the voltage regulator VR4 hot. So I moved on to the output side of VR4 and powered 15V across it. It hit the current limit too. So I cheated again with the thermal camera and found two diodes on the 15V power supply had shorted. Normally, I check every diode on a board, as these parts tend to tell you what went wrong. These and burnt resistors... So I found a 1N4746 zener diode and a 1N4148 diode. I don't have a 1N4746 to replace it (thanks again Franz), so will see what else I can find that would have taken out the diodes.

                                I found a thermal seek IR camera for my smartphone on eBay used for pretty cheap. It's better than the air spray for finding hot parts. Here's a picture of VR4. After removing zener D19 and diode D20, it's the only thing showing heat on the board when I put power to the board (I tried powering from both the input side and output side of VR4). There's ~1.8ohm across the 30V circuit. I'll have to see what pulling the leads from this component does.
                                Click image for larger version  Name:	img_thermal_1570405296836.jpg Views:	0 Size:	65.4 KB ID:	603013

                                ​​
                                Last edited by jjohn76; 10-07-2019, 10:35 AM.

                                Comment

                                Working...
                                X