Jump to content


  • Posts

  • Joined

  • Last visited

  • Days Won


BritishRacingGreen last won the day on October 10 2023

BritishRacingGreen had the most liked content!

About BritishRacingGreen

Profile Information

  • Location
  • Interests
    Embedded hardware/software , Linux , ARM CPU , railway signaling and solar.

Recent Profile Visitors

3,897 profile views

BritishRacingGreen's Achievements

  1. Unfortunately not. The grid input, the inverter output and the load output becomes tied together once the inverter has synchronized to the grid. The only possible solution you may find is to feed an external 48v battery charger from the gen, and connect the charger's dc output to the existing battery connection. That will set you back some money though.
  2. That's very interesting , I was under the impression that the SA had an ad-hoc OS based on Erlang Virtual Machine. That was an assumption because I remember that the OS was inaccessible at time I checked it out 3 years ago. I had a conversation with the Pierre the founder of SA , and he told me SA is developed using Elixir programming language and framework. Because Elixir needs Erlang VM as a host , I assumed they made a proprietary OS with BuildRoot and Erlang VM . But after reading your post I discovered that Erlang VM can also be installed on plain Debian as a package. So that's probably the route they took. My main interest at the time was whether I could host a standalone Linux app alongside SA on the same PI. Do you think that's possible.?
  3. Hi P1000 , you were absolutely correct that the switching losses are lower on 4668 than 4468. I have two identical 2.5kW dc-dc converters on the bench that I am exciting manually , one set is standard fitted with 8 x 4468 on the battery side and one set fitted with 8 x 4668 (repaired). The +12V driver side of the 4468 converter shows about 400mA total , that of course includes the 4x igbts on the high side and the system supply as well . The 4668 side draws only about 120mA . I assume that is the price we pay for the fact that the 4468 has such a low Rds ? I have integrated the repaired module back into the InfiniSolar inverter and tested battery mode operation under no load . In the next couple of days I am going to put the system under incremental load increase , and monitor. I am considering mounting two 18B20 temperature sensors on the heat sinks in order to compare heat dissipation under various loads. Thank you for your and @Coulomb's valuable info on the 4468 vs 4668 provided , that's already nearly a year ago !
  4. Thank you Justin for your valuable feedback and the links provided.
  5. I want to bring you up to speed with what I know about Electric Vehicles (EV). I know nothing. You are now up to speed That's why I found the attached video of EV Inverter Teardown so interesting, and like to share this. Now I know that @Coulomb is amongst other, an EV guru . So I have a couple of questions after watching the video. By the way I thought the young gentleman that presented the video did a nice job and seems quite knowledgeable too. The choice of 96vDC battery input , is that standard for small vehicles ? Is this voltage higher for bigger vehicles? The inverter is shown with a large heat sink . Is cooling typically forced by electric fan or the natural flow of air thru the car, or both ? He mentions that the inverter is 20kW . That's a lot. Now I assume that because there is no dc-dc conversion that the output is also 96V peak. Correct? There's no output filtering on the phase outputs . I assume the inherent dampening (inductance) of the motor does the trick ? How is speed controlled ? Frequency or duty cycle? That little inverter must be quite efficient switching 20kW with really not a lot of mosfets ! EDIT : So the power terminals for these inverters are typically 2 x battery inputs and 3 x phase outputs. Is the digital interface of these modules typically standard , is there an industry standard already? Would be nice to know if one can interchange inverters from different manufacturers.
  6. Is the status of this thread solved or unsolved? If solved, kindly share with us so we can equip ourselves to help other members with similiar problems.
  7. https://www.deltaww.com/en-US/products/PLC-Programmable-Logic-Controllers/252
  8. The OP does not give details regarding the configuration of the PV strings and the type and model of inverters. @Scorp007 and @TaliaB has pretty much covered important issues here. The cost of such a selector system will be rather high , say between ZAR10k - ZAR15k . Which is already the cost of half the panels at stake. I would have just added the the other half and installed panels statically. However the OP may have the situation where roof space is just not available and hence its not an option. If so , I assume that the OP has done the math in order to make sure that the net result may be not close to stealing from Peter to pay Paul . if such a selector then becomes a viable option , then a number of issues will be important to address. And these issues are not trivial. As @TaliaB pointed out , DC contactors are expensive. And further to this , the roving solar panel string needs to be fully double cut , both its negative and positive leg must be switched. It is imperative in a transformer-less inverter design , that the MPPT(s) of and inverter should be in isolation to MPPT(s) of other inverters. Neglecting it can affect serious damage to both inverters , mainly because the Live and neutrals are not isolated from PV- and PV+ . And using a dc contactor with a changeover contact set is not good enough , unless they are specifically rated as break-before-make. So in practice you can find yourself using up to 4 dc contactors to perform this function. And on top of this you need a contactor with an auxiliary contact so you can read and prove that a contactor has been de-energized , before an attempt is made to energize the alternate contactor. Furthermore the selector will require that the 10 panels be configured such that it will be in parallel with the static panels , so that when the string is removed by contactor the inverter MPPT will not see an open circuit. This is true for both solar panel clusters. You will then need an industrial Programmable Logic Controller (PLC) , example the DELTA DVP range . The PLC must be fitted with a real time clock (RTC) . The PLC will be responsible for the timekeeping as well as the safety cycle involved to switch a string out of circuit and in to circuit. It will also be recommended that the plc be fitted with a small control panel , with a 3 way selector switch (FORCE A - AUTO - FORCE B ) , plus indications to show to which cluster the string is switched into. *** The Delta PLC as indicated above has standard Modbus R485 , which may be used to control the selector remotely as well , using algorithms that are bases on power demand criteria etc. The budgetary quote of 10k-12k excludes installation , cables,wiring , 24VDC power supplies , enclosures , testing , commisioning costs.
  9. Hi 007 , the context of that statement was when installing. The main reason is that if you inadvertently touch screw driver to earth while connecting PV when inverter is on , then you risk damaging the DC-AC bridge. Regarding the context of this thread, I don't think removing the PV will damage the inverter , but if that PV is under load then the abrupt removal will ripple thru the power chain and cause all sorts of metastable conditions , I think which should be avoided.
  10. This is a nasty one and its got me head scratching. I can relate to F58 error , but that's typically when the DC-AC bridge has one IGBT not firing . In that case only half of the full bridge is working and the RMS output of the AC is half of rated voltage. But in your case the inverter starts correctly when DC is supplied without grid. And then when grid is applied , the overload occurs . I can only think of one failure mode which would explain such behavior, and that is if the grid / safety relay contact is somehow shorted or its coil control circuit is stuck in energized state . When working correctly this relay eventually closes and connects the grid to the inverter output , but only after the inverter has synchronized to the grid. This failure could explain the overload as the grid is prematurely tied to the internal inverter output before proper synchronization has taken place. It also explain why the inverter is working fine from dc sources when grid is not attached. This is a long shot , I could be wrong, and I would like to prompt @Coulomb to maybe shed light on this issue. One way of checking my theory is to physically remove the grid from the mains input terminals and then switch on the inverter via battery only. Then check on the LCD display whether both the Mains INPUT and Inverter OUTPUT shows the same voltage. In any case please refrain for the time being from connecting grid to the system. It could eventually result in the main board being damaged. EDIT : My theory could be wrong as I have just realized that the inverter should check for live voltage on its output before attempting to generate its own supply . But I am not sure if the software is doing this test .
  11. Sorry my response is somewhat late . My argument only centered around the post of someone stating that earth carries the current for all homes back to neutral. So I am not questioning your particular arguments/request . In fact I am totally apposed to permanent neutral bonds that installers are introducing at the inverter output side . Not only is it illegal , but dangerous as it introduces earth loop impedances between the new bond and the first encountered upstream bond, when the system is in on-grid mode. As @TaliaB points out it is actually specified in no uncertain terms in the new SANS specs. Notwithstanding , the general law has been specified ages ago as per the snippet below.
  12. I can appreciate your reasoning , but there is one flaw in real life with high/medium voltage electrical reticulation , and that is you can never bargain on a perfect floating system. A perfect floating system will isolate you from the ac potential , and you will never get hurt. However in real life there is always unwanted small leakages from the carrier cables to earth . So now you touch one cable with your body that's earthed and that small leakage impedances now constitutes a return path to the other cable , and the results are not good.
  13. No , Earth does not normally carry load currents , only fault currents. Even if the substation neutral is not earthed , will the neutral current return via the neutral cabling. That means the only way that the live current from another home can return via your permanent earth bond , is when the that house's live has a leakage to earth . That fault current then tunnels thru earth and return via your bond. That is typically small fault currents , and remember that that house's main ELD detects leakage to earth and trips . So the probability of this scenario causes physical damage such as explosion is ruled out. And as a bonus , your own main ELD will also trip , because it detects larger return current thru your neutral than went thru your own live . EDIT: ok this statement is only applicable if your inverter grid input is wired from the main ELD output. And i know that probably 90 per cent of installations the inverter is connected to raw incoming mains.
  14. well done , makes sense as that resistor is the gain / attenuation component , and the failure led to the opamp driving max output.
  15. By output I assume you mean the output of the opamp . The +12V is indeed causing the full swing in reading. I think the cause is now limited to the opamp being faulty or there may be dry joints in the resistor strings. you will have to check for continuity. If you are able to short circuit the + and - differentials inputs to the opamp , the output should go to zero volts. if not its probable the opamp that needs replacement.
  • Create New...