So , inverter is powered from 12v battery as i understand, and how inverter's case is comnected? Is it grounded to car 12v negative, metal chasis, just wire to negative, or just floating (isolated electrically)? Metal case should shield from variuos rf noise transmission, emitting, but it must be grounded, sometimes just wire to chasis is enough.
Please bear in mind that the Amy Alice filter was very carefully designed for SMPS noise - I'm not sure your inverter noise has the same frequency/amplitude profile. It's probably a good idea to ask in the dedicated thread with the pictures of what you're seeing on the oscilloscope, just to confirm if it will work for you.
12.8V lifepo4 battery isolated from engine battery. If we state ground as car chassis, nothing has modified the original car circuit. It's a standalone setup with a galvanic isolated dc dc that let alternator to recharge lifepo4.So , inverter is powered from 12v battery as i understand, and how inverter's case is comnected? Is it grounded to car 12v negative, metal chasis, just wire to negative, or just floating (isolated electrically)? Metal case should shield from variuos rf noise transmission, emitting, but it must be grounded, sometimes just wire to chasis is enough.
Inverter chassis is normally EARTHED to the outlet AC socket. I have then connected it to other chassis (MPPT), surge protectors, lighting protector, ecc... and then a 10mmq wire go out of the van connected to a zinc coated braid that touch the asphalt.
I'm afraid to link chassis (engine battery negative) to EARTH. What do you all think about?
Inverter chasis may not have internally connected to negative 12v , so you should measure and check how it measures. If many safety things are connected to inverter case, case also must be grounded to something, to car chasis. Inverter's chasis can't be floating with outlet grounds connected to it, but isolated from car chasis i think.
In fact I've made earth (isolated from ground) touching asphalt.Inverter's chasis can't be floating with outlet grounds connected to it, but isolated from car chasis i think.
Ground = is for GND of 12V engine battery, car chassis, etc... as manufacturer designed it.
Earth = is for 230V AC and I made a steel braid scratching the asphalt. When stationary on grass, that earth cable will be insert in the earth with a copper stick.
Since ground is isolated from earth by wheels, maybe I can put a device between ground and earth, a sort of resistor to let the 2 path be at a diff potential? I dont know..
If you wanna use isolated power input for inverter , lifepo4 pack, charged from ra separate source, like alternator, that's ok, but common ground should be present i think. Can you check with oscilloscope also , what noise you have at inverters case , in respect to 12v inputs negative? Maybe ac output wires both are floating, and no capacitors are present , to lock potential at ac rf, to reduce variuos leakages. In fact, few days ago had similar problem to yours, I was experimenting with power supply prototype and electronic load was connected, and i was unable to control it - some buttons not worked, when power supply was turned on. Problem was i missed a capacitors between transformer's primary gnd (-300v) and output , when i added it , buttons works. So what i learned , is that output in variuos transformer based converters cant be fully isolated, it must be rf coupled to input , to return intertwining lcapacitive leakage caused currents again to input. So check noise in variuos places, in respect to grounds , cases , input .
The 3rd image (5 microsecond) seems the more realistic. From that I suppose is 50kHz frequency. It is above 20kHz audio, and presupposes 1000x the freq of AC 50Hz.
If I understand right, inverter freq divided by AC freq gives the n° of steps used to make a period: 1000 steps.
Dived it in 4, to obtain steps from 0 to V peak, and obtain 250 steps.
Is my analysis right?
If it right, I need a filter for 50kHz.
Third image looks like burst packets , maybe your inverter operating at low load and therefore in shorts periods. Just guessing. One idea what to try. Connect oscilloscope probe ground(aligator clip) to inverters case and connect nothing to signal pin, and by moving probe , locate which wire emits maximum noise. I'm just guessing you need to add capacitors from 12v negative to one of 220v output wires, for testing, to stop reduce emitting that noise, but that needs fk be verified first. I'm using 4700p blue capacitors, used in smps from line wires to ground. For safety, connect two or three in series. Turn off inverter first offcourse and remove dc input fuse before , to be safe.
- Home
- Amplifiers
- Power Supplies
- 5V DC ripple analysis method