Announcement

Collapse
No announcement yet.

SMA ShadeFix ?

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

  • SMA ShadeFix ?

    Wondering what people think of this "ShadeFix" feature that SMA seems to claim eliminates the need for panel-level optimization:
    https://www.sma.de/en/shadefix.html

    It seems a bit implausible to me. Let's say two panels in a string are shaded, so they can't pass much current; the silly little video (at the link) seems to imply that the inverter takes those two panels out of the current path. How it does that is a mystery to me.

    In practice, it doesn't seem to be working for me. My grid-tie system has two strings of 8 panels each. Right now, one string is in full sunlight, and one string has two of the panels about half-shaded. The second string is producing about half the power of the first string.

    Is it possible I need to update the firmware ? Right now I have firmware version 3.1.25.R.


  • #2
    Read about bypass diodes. With 2 panels shaded, the other 6 will push there current
    through them, the string voltage at that current would be only 6/8 (3/4) of 8 unshaded.
    That works until you connect a string of 8 unshaded panels in parallel, at full voltage.
    They will be, inefficiently, forced to the same voltage. One possibility is the shaded
    section will slide up the panel V-I curve to match the full voltage, but at that operating
    point it has very little current capability.

    Another possibility is the unshaded string will drop to 3/4 of its MPP voltage, supplying
    slightly more current (same curve). The shaded panels will put out normal current at 3/4
    voltage, so your system delivers about 3/4 power. The operating point found by an MPPT
    control could be anywhere between these 2 extremes.

    Optimizers were developed to improve efficiency in such situations. Bruce Roe
    Bruce Roe

    Comment


    • #3
      The best you can hope for with a string inverter without panel level optimizers is that the MPP-seeking algorithm scans the range of possible voltage and current combinations widely enough to find the absolute maximum instead of converging one one of the lower local maxima.
      SunnyBoy 3000 US, 18 BP Solar 175B panels.

      Comment


      • #4
        Originally posted by bcroe View Post
        With 2 panels shaded, the other 6 will push there current
        through them, the string voltage at that current would be only 6/8 (3/4) of 8 unshaded.
        That works until you connect a string of 8 unshaded panels in parallel, at full voltage.
        They will be, inefficiently, forced to the same voltage.
        My inverter has two MPPT's, so the 2/8 shaded string, and fully-unshaded string, are on separate ones.


        Comment


        • #5
          Originally posted by RShackleford View Post
          Wondering what people think of this "ShadeFix" feature that SMA seems to claim eliminates the need for panel-level optimization:
          https://www.sma.de/en/shadefix.html

          It seems a bit implausible to me. Let's say two panels in a string are shaded, so they can't pass much current; the silly little video (at the link) seems to imply that the inverter takes those two panels out of the current path. How it does that is a mystery to me.

          In practice, it doesn't seem to be working for me. My grid-tie system has two strings of 8 panels each. Right now, one string is in full sunlight, and one string has two of the panels about half-shaded. The second string is producing about half the power of the first string.

          Is it possible I need to update the firmware ? Right now I have firmware version 3.1.25.R.
          'ShadeFix' is just 'clever' marketing for what has always been true since bypass diodes became a thing if you don't have parallel strings on a MPPT.

          As long as each string is able to operate independently 'String Level MPPT' (And all string in a panel are roughly the same with the same orientation) there is virtually no benefit to module level optimization. SMA just finally got around to confirming the obvious with science.

          Do each of your strings produce the same power in full sun? You may have failed bypass diodes in the underperforming string.

          Comment


          • #6
            Originally posted by nwdiver View Post
            'ShadeFix' is just 'clever' marketing for what has always been true since bypass diodes became a thing if you don't have parallel strings on a MPPT.

            As long as each string is able to operate independently 'String Level MPPT' (And all string in a panel are roughly the same with the same orientation) there is virtually no benefit to module level optimization. SMA just finally got around to confirming the obvious with science.

            Do each of your strings produce the same power in full sun? You may have failed bypass diodes in the underperforming string.
            Each of my two strings is 8 panels, and all 16 panels are identical 270 watt panels. Each string is on a separate MPPT input to my 3.8kW SunnyBoy inverter. All 16 panels are in a single fixed ground-mount array (so all have the same azimuth and elevation).

            I believe both string produce about the same when in full sun, but I'll check.

            P.S. Thanks for the simplest/clearest explanation I've seen of this so far. I was under the impression I'd benefit from module-level optimization, and got a pretty good deal on 16 TS4-R-O devices. I was confused about whether that'd be compatible with "secure power" (the thing that gives you up to 2000 watts of power when the grid is down and sun is shining), but decided it was. Then a tech engineer at SMA told me no, it wasn't - you can't use secure power if you have optimizers. I complained that I'd been told you could, and asked if they'd cover my expenses, and he said they would. Fortunately Northern Arizona Wind & Sun not only gave me a full refund for the TS4's, they even gave me a prepaid return label, so shout-out to them. So I was only out the original shipping charge, $45 or so; SMA said they'd refund that, and then they reneged saying their "legal team" forbade it. So perhaps they make a good product, but I think their ethics are questionable. Also, there tech info is lame: at this point, I honestly don't know whether or not "secure power" is compatible with module-level optimizers, or not, but as you say, there's not much to gain from them. Their manual sucks too; it gave me zero help in connecting my inverter to the "Sunny portal" thing, but a patient tech support guy walked me through it, he was very helpful. As a kicker, the online version of the lame manual repeatedly crashes "Preview" (the app you use to view pdf's on a Mac). Oh well.



            Comment


            • #7
              Originally posted by RShackleford View Post
              Each of my two strings is 8 panels, and all 16 panels are identical 270 watt panels. Each string is on a separate MPPT input to my 3.8kW SunnyBoy inverter. All 16 panels are in a single fixed ground-mount array (so all have the same azimuth and elevation).

              I believe both string produce about the same when in full sun, but I'll check.

              P.S. Thanks for the simplest/clearest explanation I've seen of this so far. I was under the impression I'd benefit from module-level optimization, and got a pretty good deal on 16 TS4-R-O devices. I was confused about whether that'd be compatible with "secure power" (the thing that gives you up to 2000 watts of power when the grid is down and sun is shining), but decided it was. Then a tech engineer at SMA told me no, it wasn't - you can't use secure power if you have optimizers. I complained that I'd been told you could, and asked if they'd cover my expenses, and he said they would. Fortunately Northern Arizona Wind & Sun not only gave me a full refund for the TS4's, they even gave me a prepaid return label, so shout-out to them. So I was only out the original shipping charge, $45 or so; SMA said they'd refund that, and then they reneged saying their "legal team" forbade it. So perhaps they make a good product, but I think their ethics are questionable. Also, there tech info is lame: at this point, I honestly don't know whether or not "secure power" is compatible with module-level optimizers, or not, but as you say, there's not much to gain from them. Their manual sucks too; it gave me zero help in connecting my inverter to the "Sunny portal" thing, but a patient tech support guy walked me through it, he was very helpful. As a kicker, the online version of the lame manual repeatedly crashes "Preview" (the app you use to view pdf's on a Mac). Oh well.


              SMA recently published a work around for using SPS with TS4-S and Os. But no, there's really no benefit to module level optimization if all panels in a string have the same orientation. Even if there's some shade. As each cell group is shaded the bypass diode is activated and 'deletes' the shaded section from the string. The same effect would occur with optimizers unless all cell groups in a panel are shaded in which case the production is barely worth harvesting. This can cause issues if you have parallel strings since now you'd have two strings in parallel that are effectively different lengths. That's why string level MPPT is so beneficial.

              Comment


              • #8
                Originally posted by RShackleford
                Sorry about this, was trying to get the screenshot to look right and this stupid system flagged my post.

                Anyhow, screenshot would show that string B has similar voltage as string A but about 1/4 the current.
                Same orientation?

                SMA inverters 'clip' one string at a time. If the inverter is saturated string A should have lower current.
                Last edited by nwdiver; 09-11-2020, 01:42 PM.

                Comment


                • #9
                  Originally posted by nwdiver View Post

                  Same orientation?

                  SMA inverters 'clip' one string at a time. If the inverter is saturated string A should have lower current.
                  Yeah, all 16 panels are same azimuth/elevation.
                  Inverter is 3.8kW so I don't think it's saturated.

                  I'll just redo that post ...


                  Comment


                  • #10
                    Originally posted by nwdiver View Post

                    SMA recently published a work around for using SPS with TS4-S and Os. But no, there's really no benefit to module level optimization if all panels in a string have the same orientation. Even if there's some shade. As each cell group is shaded the bypass diode is activated and 'deletes' the shaded section from the string. The same effect would occur with optimizers unless all cell groups in a panel are shaded in which case the production is barely worth harvesting. This can cause issues if you have parallel strings since now you'd have two strings in parallel that are effectively different lengths. That's why string level MPPT is so beneficial.
                    Hmm, so may somethijng isn't right. At this moment, one string ("string A") is producing almost 2000 watts DC, and string B only about 500 watts. String A is completely sunny (mayb slightly shaded when this screenshot was taken) and string B has two panels less than half-shaded.

                    Screen Shot 2020-09-11 at 1.26.04 PM.png

                    If the bypass diodes were working correctly, seems like string B would be showing about 3/4 the voltage as string A but about the same current.

                    Comment


                    • #11
                      Originally posted by RShackleford View Post

                      Hmm, so may somethijng isn't right. At this moment, one string ("string A") is producing almost 2000 watts DC, and string B only about 500 watts. String A is completely sunny (mayb slightly shaded when this screenshot was taken) and string B has two panels less than half-shaded.

                      Screen Shot 2020-09-11 at 1.26.04 PM.png

                      If the bypass diodes were working correctly, seems like string B would be showing about 3/4 the voltage as string A but about the same current.
                      Yeah... looks like something is wrong. I would swap the strings. That would tell you if it's a problem with the panels or the inverter.

                      Another problem might be that your voltage is too low; But the MPPT is supposed to work to 100v.
                      Last edited by nwdiver; 09-11-2020, 02:08 PM.

                      Comment


                      • #12
                        Originally posted by nwdiver View Post

                        Yeah... looks like something is wrong. I would swap the strings. That would tell you if it's a problem with the panels or the inverter.

                        Another problem might be that your voltage is too low; But the MPPT is supposed to work to 100v.
                        Yeah, I never expect voltage much higher, since only 8 panels in strings and NOCT operating voltage is 29v. Thought of doing one string, but with STC open-circuit voltage of 38.5v, I'd be playing with fire (literally ) since max DC input of SunnyBoy is 600v. Looking at efficiency curve here (albeit for 6kW model) it looks like I only lose about 2% by having my DC voltages down in the 200s.

                        Screen Shot 2020-09-12 at 1.45.28 PM.png


                        But yeah, something wrong. Won't get a chance to mess with it for awhile, but I'll check it out for sure.

                        Comment


                        • #13
                          I guess I should start a new thread about this, but, is it just me, or does the "Sunny Portal" (the page at suunyportal-dot-com) give a lot less information than one can get by connecting to the inverter's wifi network (the one whose name is "SMA#" where '#' is the inverter's serial#) ? Primarily, the former often gives pretty stale figures for the current power and today's generated watt-hours, and does not provide info about the individual strings.

                          Comment

                          Working...
                          X