Jump to content

JustinSchoeman

Members
  • Posts

    446
  • Joined

  • Last visited

  • Days Won

    4

Everything posted by JustinSchoeman

  1. This reduces the thermal load on the inverter, but does not alter the peak load at all. So depending on what is limiting the inverter's max output, it may not help.
  2. Above ~3.4V both Ri and Ui shoot up, so once a pack reaches this point (~90$ SOC), pretty much all charge goes to the others. In this regard, parallel lithium batteries are excellent Marxists: "From each according to his ability, to each according to his needs" The big problem is that you can't parallel current delivery capability. If you demand 200A from 2 parallel 100A packs, the freshest one will try to deliver most of the current. So one might try to do 150A while the other only does 50A. This results in first the one, then the other BMS tripping, and you end up with nothing.
  3. The threshold voltage of the MOSFETS is around 4V. The moment it is below that the device is off and, you don't care any more. So if you end the fast discharge at 0.7V or 0.3V makes no difference - the device is already off in both cases. So manufacturers choose the cheapest fast signal diode that can handle the peak currents.
  4. A typical diode current-voltage plot looks something like this: Try to raise the voltage much above 0.7V, and the current goes pretty much to infinity. Say Vgs=15V, then without the diode, the resistor would conduct 1.5A. But if you follow the diode curve to 15V, you would end up with millions of Amps. But we also know that in parallel circuits, the current follows the path of least resistance, so it will follow the path that wants millions of Amps (diode), rather than the one that wants 1.5A (resistor). In reality, we won't get millions of Amps though. The diode will conduct as much current as it can (until the smoke comes out) to limit the voltage to 0.7V. So, you can view a forward biased diode as a fixed 0.7V drop. The resistor will conduct 0.7/10 = 70mA, and the diode will conduct whatever additional current is required to keep the voltage to 0.7V.
  5. Nope. It depends on the size (current capacity) of the device, not the actual current. If you want to switch on a transistor with a continuous drain current rating of 100A, it would require about 10x more current than to turn on a transistor with a continuous drain current rating of 10A (at the same switching frequency). When that 100A transistor is on, it can conduct 10A or 100A, but it would have taken the same current to turn it on. Once it is on, it requires no current to keep it on.
  6. The gates of both FETs and IGBTs are capacitors (sort of - actually non-linear compound capacitors, but it is an accurate enough simplification for this discussion). So, with a gate capacitance of C and a turn on voltage of V, we require a charge of Q = C * V to turn the device on. In truth, while we say the devices are voltage controlled, they are actually charge controlled. Because of the above equation, they are largely interchangeable. But it is the charge on the gate capacitor which provides the electrons in the channel/base of the device. And it is also the number of electrons (and therefore charge) that limit the charge carrying capability of the device. So the more current you need to carry, the bigger the charge required. But we also have Q = I * t. So for a specific turn on/off time 't' we require I = Q / t. So the required gate current is proportional to the gate charge, which is proportional to the current carrying capability. This is a little over-simplified, but it should give a good, basic understanding of the device parameters.
  7. Just make sure you have all the planning permissions in place, or they will give you a raft of fines rather than a disconnection.
  8. Good point - that time constant is not even enough to make a dent on the soft-start cap charge. The off time may just be a side effect of the need to discharge the cap. I see one other side effect. Adding R108 reduces the SHDN voltage from ~12V to ~5V (relative to -12V). Is it possible that they use alternative parts for Q60/Q61? If they use one with a low threshold and low Vgsmax, then adding this RC pair could make a difference.
  9. I think something else may be wrong here. Worst case, the SG3525 outputs will shut down 0.5us after SD pin exceeds 1V. R113/C57 already provides a delay function with an initial rise of 0.17V/us. So as long as the threshold voltage for Q60/Q61 is above ~1.2V, they will not turn on at all until long after the SG3525 is shut down. C15/R108 seems to have 2 functions: (1) make sure the device is NOT shut down at power up, and (2) increase the ramp time/delay even more. Not sure why Q60/61 are there. They seem to only speed up the final decay of 1 Vbe drop. If it was other side the capacitor (directly on the coil), it would be a very effective final clamp. The DC blocking capacitor is quite a bit bigger than for the IGBT drive, which implies more current, and also more energy in the LC tank which would need to be damped at shutdown, which could explain why a clamp is included.
  10. Are you sure those diode-capacitor pairs between pin 8 and 7 are not between 8 and 5? As shown there, they will kill the switching speed and likely cook the drivers?
  11. The fuse is primarily to protect a single string when multiple strings are in parallel (i.e. in a combiner box), where the sum of the currents from other strings can exceed the current limits of the bypass diodes in a bad string. This is also why fuses are recommended, as the fault current could be a reverse current. So you would need non-polarised breakers in a combiner. AFAIK only Australia has a regulation requiring fuses instead of breakers. I can't see any local regulation requiring a fuse instead of a breaker.
  12. I get the feeling he is only seeing ground path inductance in those plots. They look pretty much like what you would expect if measuring over a small inductor in series with the source.
  13. It should be trivial to add, but I don't currently have a need for it myself, or much time to work on it. Not sure if anybody else wants to join the effort and add more modules for other comms protocols?
  14. D32 bypasses the gate resistor for discharge (turn off) - so turn off is faster and less damped. You need to manage dI/dt during turn on, which is why you have R91. Don't have the same problem during turn-off, so bypass R91 with D32 for turn off.
  15. Ah. I remember the 'good old days'... You generally aim for a 10ns rise/fall time on the gate. So this signal has strong frequency components in the 100MHz - 1GHz range, and at these frequencies every track, wire and lead is a significant inductor and capacitor. So you end up with dozens of resonant frequencies which you need to damp, and PCB design becomes critical. Then you add a 10pF scope probe and add a whole new set of potentially resonances. But how do you tell the difference between 'real' transients and those introduced by the scope probe itself? I remember this actually being one of the hardest parts when designing a switching regulator. I have never had to debug someone else's design - but if I had to, I would just check component values and connections. If everything checks out I would assume that the actual drive waveform is correct and I am just seeing measurement induced transients.
  16. That is pretty much it. There is enough energy in the entire cycle to rapidly charge the gate capacitor. But not enough instantaneous power. So, D37, R47 and C112 form a filtered power supply. Q52 is a class B amplifier (current amplifier) running from this power supply to supplement the transformer power during the switch on phase. The other components are a mirror for the switch off phase. The lower currents on the HV side mean that you have to move much less charge to turn the devices on and off, and you can get away with driving directly from the transformer.
  17. You are measuring the element temperature more than the water temperature. What you are seeing there is the heat transfer from the element to the water. Rather measure on the outflow pipe, as close as possible to the geyser wall. My Kwikot 150L (without geyser blanket) loses ~0.5°C per hour at 55°C.
  18. Remember that currents on the IGBT side are 1/8 of those on the MOSFET side. So expect similarly lower gate charge (and probably a slightly more relaxed switching time requirement). Apparently (for the devices in use) this is enough of a difference to allow for direct drive of the IGBT, but require a totem pole drive for the MOSFET.
  19. You also need to take power factor into account. The power factor of a capacitive dropper is roughly equal to the output voltage to the input votage. So, if you halve the power, the power factor will be around 0.7. Inverters are designed and rated for a PF of 1, and generally derate from there. So, while you halve the heating capacity, you only reduce the effective inverter load by 30%. So the win may not be as big as you expect.
  20. You could potentially claim it is essential computer equipment and power it from an unprotected red plug.
  21. Interesting design. Since fuse response time is substantially faster than breakers, the first thing that happens on a surge is the fuse blows and removes the SPDs from the circuit...
  22. Negative. There MUST be an overcurrent protection device before the SPD. The SPDs job is to crowbar the lines and trip the overcurrent protection. The SPD cannot dissipate enough power to provide any real protection on its own.
  23. From your IEC-60269-6-2010 document, which defines the gPV utilization category: "1.1 Scope and object These supplementary requirements apply to fuse-links for protecting PV strings and PV arrays in equipment for circuits of nominal voltages up to 1 500 V d.c." gPV is specifically for protection of solar panels. The question at hand is with regards to battery fuses: DC rated but with gG utilization category.
  24. Fortunately, SANS10142 only requires fuses for conductor protection (not device protection), so gG should meet the legal requirement. I have never seen gS or gR anywhere in South Africa (technically the outside facing part of the battery is the BMS, so gR would be more appropriate - gS would be for LA batteries). For DC fuses, GC Solar is showing stock. Lite-Glo is not showing stock, but if you pop them an email, they will generally source whatever you need in a day or two.
  25. The fuse holder is rated for 440V DC. The fuse *might* have a DC rating, but you will need to ask the supplier for a datasheet to confirm (you need a password to get tech docs directly from Onesto).
×
×
  • Create New...