Announcement

Collapse
No announcement yet.

Help needed with fried SWMPS - RING RSC612/RESC612

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

    #21
    Re: Help needed with fried SWMPS - RING RSC612/RESC612

    Hi,
    Many thanks for the replies, I will check both suggestions / lines of thought and post what I find. I may be some time however... LOL.

    Comment


      #22
      Re: Help needed with fried SWMPS - RING RSC612/RESC612

      I have two of the same units, one from 2015 and one from 2020, which was sent by Ring to me when the earlier unit, although out of warranty started playing up - thanks Ring !.

      The fault started as intermittent F03 errors, which according to the manual is is unable to control charge current. I also noted that it would stop charging, showing a full battery, when the battery was not in fact fully charged. I had this on my bench for a bit handling a battery used for an off-grid purpose, the charger then completely failed in the same manner as this thread shows - no mains operation, but would run tests when the battery is connected. You could sometimes see a bit of pulsing of the display for a couple of seconds.

      I have spent some time diagnosing this including reverse engineering a couple of sections of the board to understand its operation, however I had to resorted to comparing between the old and new units to narrow down the fault and eventually fix it as I'd never attempted to fix an SMPS before.

      I have also noted some differences between the units, so I'll list those to help others too

      The next couple of posts will contain what I found out about this. All the notes relate to the 2015 model unless I state otherwise.

      Comment


        #23
        Re: Help needed with fried SWMPS - RING RSC612/RESC612

        Looking at the pictures earlier in this forum, it appears that mandan1 has the later version of the board.

        I can confirm that N5 is an NPN BJT with an HFE of 200, the pinout (using the pin numbers in mandan1's post) is C, B, E. Similarly, Q9, is a PNP BJT with an HFE of 210, the pinout is C, B, E

        In my case, I could see the input filtering and initial AC to DC conversion taking place, resulting in 338V DC but nothing else was operating, the +12 and +5V rails on the low voltage side were at zero.

        I started where mandan1 was, but concluded that the N5/Q9 area provides remote powerup from the small riser PCB on the low voltage side and its purpose is to turn on the 8 pin IC (ST3842) near R21, however as that board had no power it wasn't operating.

        As the board is covered in conformal coating, I tried to scrape it off some of the IC's and semiconductors with a screwdriver and although successful on some devices, I was unsuccessful on U5 where it damaged the top surface making it unreadable. I found that acetone is really good at getting off the conformal coating so that you can see the ID's on the components. It takes a bit of scrubbing with a cotton bud, but it comes off after 30 seconds to a minute. U5 was unreadable, which was a problem.

        To help others with any fixes in the future, here are the components that I managed to identify
        • Devices prefixed N seems to indicate an NPN transistor and Q seems to indicate a PNP transistor, or other semiconductor
        • 8 pin IC near R21 is an ST 3842B https://www.st.com/resource/en/datasheet/uc2842b.pdf
        • Q1, STPS20S100CT, dual diode (on the low voltage side)
        • Q2, ST P9NK70ZFP, N channel SuperMESH MOSFET, ID=7.5A
        • Q3, CS2N60, N channel MOSFET, ID=2A
        • The optos are F817C. All opto are controlled from the low voltage side
        • N1 – BV51, fan control transistor NPN ?
        • Riser board is ST GZQC, also says 324 and 432 on the line below it. The PCB its on says WYH6859CON-A1.PCB
        • U1 AS78L05Z-E1, 5V regulator (Diodes Inc) Input, GND, Output. Feeds the riser board.
        • U4 7550-1, 5V LDO – not sure what this feeds
        • U5 is marked (2020 model) as TDP, then in smaller letters, 36R and J1 on the line below. Believe to be a Diodes Inc, AP3105NA, which is obsolete and has been replaced with AP3125. There are two variants, one has a CTRL pin, this is the one that matches the PCB. https://www.diodes.com/assets/Datasheets/AP3125A.pdf
          variant. The circuit diagram broadly matches the one in the datasheet.
        • N5 is a NPN BJT, HFE=200, Pinout is C, B, E
        • Q6, S8550 PNP, IC=500mA, pinout, C, E, B
        • Q9 is a PNP BJT, HFE=200, pinout is C, B, E
        • Z3 is a 1N47746A, 18V Zener
        • Z5 is 1N4747A, (51 on the second line), 20V Zener
        • M7 diodes are 1N4007's, 1000V, 1A rated
        • C4 is 11nF, doesn't measure correctly in circuit

        Comment


          #24
          Re: Help needed with fried SWMPS - RING RSC612/RESC612

          Now, onto the failure mode.

          Once I'd located the two Zener's and tested them on my Zener tester as OK, I checked them in circuit when powered on. There is what I now know to be VCC, pin 5 on U5, this is the trace below D8 when the silk screen text is the right way up. This was measuring around 17V and I initially thought it was OK as it was close to the Zener voltage of 18V for Z3 (1N4746A). This supply also goes via JP24 to the 8 pin SO8 ST 3842B.

          This pin has two power feeds into it, the initial bootstrap power of 338VDC via R43 (200K) and into Z3 (18V) E6 (why not C6??), which is a 47uF 25V cap and R48 which is a 15K bleed resistor. Once there is sufficient power on U5, it starts firing Q3 which energises the transformer and generates a secondary power feed via the second winding on that transformer. The circuit diagram is very close to that defined in the Diodes Inc AP3125 typical application circuit on page 2 of the datasheet

          When observed on the scope, you see the voltage level climb to 17V and then after a couple of seconds, the supply die to about half that voltage. Pin6 is the output of U5 that feeds via R59 (100R) into the gate of Q3, this shows no signal. I then noted that the traces on the board around this area were discoloured a little, so I removed R43 and capacitor E6, which is a 47uF 25V electrolytic. This tested as 22uF with a massive ESR of 20 ohms, so definitely faulty. R43 was also odd, it was a small 1/8 watt resistor and definitely marked as 200K, however it measured exactly 300K on the meter. There is no indication of discolouration on it and unfortunately no other similar components on the board to compare against. I replaced the resistor with a 1/4 watt 220K resistor and E6 with a 100uF 35V component (see next couple of posts for the reason for that)

          Having replaced both components and re-testing, the VCC rail is now a solid 18V and the gate on Q3 is firing nicely and there is power coming back from the feedback coil on the transformer, so its fixed.

          I guess that as the power rail would have been soft before it went under the threshold where it failed completely, there is a good chance that this is related to the original faults I was seeing with the unit. I've not tested enough yet to see if that problem is also fixed.

          Comment


            #25
            Re: Help needed with fried SWMPS - RING RSC612/RESC612

            Whilst poking around in these two units, I found a number of differences between them, which I'll record here. The assembly datecode is on the back of the unit at the bottom, under the plastic bump case. I have a late 2015 model and an early 2020 model.
            • The 2015 model has a power cable quick disconnect based on a 4 way SAE battery connector, the later model doesn't
            • R17 is 10R 1/8W (2015) and 20R 1/8W (2020). This is adjacent to the transformer
            • R43 is no longer fitted in the 2020 model, this is the bleed resistor across Z3
            • There is visible thermal paste on the power devices on the heat sinks (Q1 and Q2) on the 2020 model, but nothing visible on the 2015 model
            • The layout for R5 and D5 has changed
            • The power into the PWM chips U5 and the SO8 ST 3842B near R21 is via R43 (200K) on the 2015 model and via two 1W 200K resistors R49 and R59 which are before D3, the main bridge rectifier. In both cases, JP24 carries the 18V power between the two PWM chips
            • Capacitor E6 has changed from 47uF 25V to 100uF 35V
            • R110, 3K3, near U4 is present on the 2020 model


            Obviously things only change when there are problems, so its telling that both R43 and E6 are either absent or up-rated in the later board revisions. I wonder if the 25V rating was too close to the 18V Zener voltage, given all the massive spikes from switching ?
            I fitted the larger 100uF, 35V, 105 degree C capacitor for E6 on the repair.

            Comment


              #26
              Re: Help needed with fried SWMPS - RING RSC612/RESC612

              I noted a couple of interesting points on the boards. The earlier 2015 unit has signs of corrosion on some of the wire links on the top of the board, this is in the form of green verdigris and although still intact, it shows that moisture is getting into the unit, which is kept in the garage or placed under the bonnet of the car when operating.

              On the 2015 board, I noted around 10 joints that looked to be iffy from a dry joint perspective. I don't think any had got to failure yet, but there were clearly visible rings forming and the single sided nature of the board and very small amounts of solder gave virtually nothing to provide any strength on the joints. These were re-flowed, some did not want to initially comply possibly due to oxidation or the conformal coating. The 2020 board seemed to have more solder on the joints.

              As an observation, the unit has a retractable hook on the back of the case, which is at the top (mains side), so if its hung up on that hook and there is moisture around, it would go straight through the high voltage area of the board and could easily get to the the area where mandan1's had the problem, so this might potentially be the cause for his issue.

              The method of mounting also means that when the fan runs, its trying to suck hot air down and out, rather than pushing hot air up, which is an odd design decision.

              Comment


                #27
                Re: Help needed with fried SWMPS - RING RSC612/RESC612

                There is a typo in post #23, Z3 should read

                Z3 is a 1N4746A, 18V Zener

                Comment

                Working...
                X