Author Topic: Inverter induced spiking of dc bus  (Read 2874 times)

0 Members and 1 Guest are viewing this topic.

dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Inverter induced spiking of dc bus
« on: March 06, 2017, 07:04:23 PM »
I wondered if adding lots of super caps to my Toyota would make it start or run better. I already had a 53F/12v boost pack attached to the 1kw sine inverter I have sitting in front of the passenger's seat. The caps allowed me to start a window a/c when the inverter was in my van. The inverter is normally turned off but still wired through a 175 amp fusible link so the boost caps on it are always in parallel with the car battery and therefore provide a starting boost. There is a high amp twist disconnect at the battery to disable the setup if something goes wrong.

The price of caps has come down some. I recently got 20F/12v for $28 from eBay. 20F/12v is a 6-pack of 120F 2.7v in series with a balancing and overvoltage circuit. This one is mounted with short wires directly to the battery and has a 30 amp fuse on it. If I ever blow the fuse it means the car battery failed to provide the needed amps and the caps had to.

The car runs smoother with the second boost cap added and seems to start a little easier. The battery is 3 1/2 years old, which is almost twice the expected life down here in south Florida. The heat kills batteries.

I put a scope directly on the battery and switched to a/c coupling and watched the effects of rpm and inverter loading on the smoothness of the car system voltage. The inverter puts some nasty narrow spikes into the system even with no load. With a small load (a 35 watt box fan) it got much worse. I don't know enough about inverter operation to know why.
Idle, inverter off

Idle, with inverter on but no load

Idle inverter running 35 watt desktop fan

SparWeb

  • Global Moderator
  • Super Hero Member Plus
  • *****
  • Posts: 5452
  • Country: ca
    • Wind Turbine Project Field Notes
Re: Inverter induced spiking of dc bus
« Reply #1 on: March 06, 2017, 11:13:05 PM »
Most of this is ripple from the alternator.  What's the spacing of the scope divisions?  I think I can guess but...
How about a reading with the car at idle, leave the inverter off, and then turn the headlights on.
If the spikes remain, then they are not being filtered out by the car's voltage regulator.
If the spikes go away, then it's the inverter you are using. 
Also, what does the ripple look like with the caps disconnected?

Sorry, I'm not actually answering your question.  I'm still working out if your car's electrical system is OK. 
Until I figure that out, no point in guessing what the capacitors and inverter are doing.
No one believes the theory except the one who developed it. Everyone believes the experiment except the one who ran it.
System spec: 135w BP multicrystalline panels, Xantrex C40, DIY 10ft (3m) diameter wind turbine, Tri-Star TS60, 800AH x 24V AGM Battery, Xantrex SW4024
www.sparweb.ca

dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Re: Inverter induced spiking of dc bus
« Reply #2 on: March 07, 2017, 12:22:10 PM »
The top pic is the alternating current component in the system with the inverter off. If I switch to dc on the probe, it looks smooth, but rises and falls with loads like light and air conditioning. The voltage also fades a little bit when the car warms up.
The peak voltage is 14.3 but drops below 14.2 when everything is warm.

The narrow spikes only show up when the inverter is powered on. The words directly above the pix are the description.

dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Re: Inverter induced spiking of dc bus
« Reply #3 on: March 07, 2017, 06:54:20 PM »
Here are some new pix. In order. Engine off, dc. Engine on, ac component (obviously none, since the engine is shut off) and last a close up of the spikes that randomly occur when the inverter is on with no load.





joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: Inverter induced spiking of dc bus
« Reply #4 on: March 07, 2017, 06:55:15 PM »
those spikes are just noise. its not possible to tell where they are coming from without a lot of effort, perhaps equipment you can't afford, and the voltage ripple is directly proportional to the inductance between the three capacitors in your system, the capacitance and the resistance of those three capacitors, the 60 or 50hz frequency of your inverter, and the variable frequency of the alternator. by capacitors i mean your car battery, the super capacitors you have connected, and the capacitors inside the inverter.

the super capacitors should not affect the engine at all. if you have enough voltage ripple in your system that your engine's ecu can't hardly function, that's a very serious and legitamate possibility but unlikely to be the case. if that was the case your car battery would be unlikely to start the engine.

if the inverter puts out so much rfi and emi that your ecu is having trouble, also possibility, but unlikely that capacitors on the 12v dc side of things could stop it.

keep in mind that the car battery's impedance is relatively high in the neighborhood of 13 to 14 volts. the voltage ripple you saw might be entirely the voltage ripple generated by the alternator. there is litterally nothing to see here.

if the inverter is cross conducting you could see the kind of noise you're dealing with, is that 10 or so cycles in 200ns? could be cross conduction. this kind of noise is also generated by the ignition system and your car's ecu should be able to handle it
« Last Edit: March 07, 2017, 06:59:59 PM by joestue »
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.

dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Re: Inverter induced spiking of dc bus
« Reply #5 on: March 07, 2017, 07:52:42 PM »
The big voltage spikes ONLY occur when the inverter is on. They do not occur otherwise. The noise is probably from the alternator. The noise exists even if I disconnect the 2 added capacitor arrays.

The noise is millivolts so I don't expect any computer problems. I have noticed that the overall system voltage slowly sags to below 12.6v if I idle long enough in gear with the lights on. Turning on the a/c causes the idle solenoid to kick the idle up a bit and if the a/c fan is set to low 1 or 2, the system voltage never collapses. The original alternator couldn't charge at idle, either. I manually adjusted the throttle cable to force the idle up a little so I don't have to put the car in neutral if I'm stuck in traffic for a long time. I also added LED tail lights to reduce the load when sitting still in park with the brakes. The lack of charging at idle is a continuing nuisance, but the only other fix I can think of is to use a smaller alternator pulley, and I don't think that's physically possible. Sitting at idle in traffic with the a/c on would eventually overheat my Ford van, but doesn't bother the Toyota.

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: Inverter induced spiking of dc bus
« Reply #6 on: March 07, 2017, 08:34:55 PM »
not having time to try and decrypt what is going on, i will tell you those spikes are of too high a frequency content to affect anything. no, they are not from the alternator.

as you say its still there if you disconnect the capacitors. yep, sounds like its from the inverter. i would speculate that its cross conduction, its not actually on the dc bus but its being picked up by the test leads you have connected to the dc bus.
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.

dnix71

  • SuperHero Member
  • ******
  • Posts: 2513
Re: Inverter induced spiking of dc bus
« Reply #7 on: March 07, 2017, 09:42:10 PM »
I'm rather fond of that inverter. It ran a window a/c in an aging van that wasn't worth fixing the built-in a/c in and will run  my digital control washing machine. If we go off the grid for weeks again after a hurricane, I can keep my clothes clean.  I have also cooked beans and rice with it on the fly, as my commute is at least 45 minutes each way. A crock pot on the front seat and some soup and I have a hot meal ready when I reach home or work.

If someone could come up with a microwave oven that didn't need gigawatts of surge to energize the coil that would be great. This stupid long commute since the move has cost us some nice people. The head of HR just announced she was retiring early because of the stress the long ride to the other building was causing her.

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: Inverter induced spiking of dc bus
« Reply #8 on: March 09, 2017, 06:27:28 PM »
There is another source of noise of such freq content. Paralleled mosfets that dont turn on at the same time, the leakage inductance between the fets themselves (say ten nanohenries) resonates with the gate and gate to drain capacitance. You can find tiny ferrite beads on the drain of the fets in some power supplies. They act like saturateable reactors and it stops the resonance.
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.