Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.
InPhase Senior Member Joined: 15/12/2020 Location: United StatesPosts: 178
Posted: 01:40am 14 Nov 2023
Copy link to clipboard
Print this post
As long as the wire is sized to carry enough current, the voltage drop under load will not imbalance the secondary too much. The inverter will try to keep the secondary at 240 volts, and since any 120 volt leg is between the end of the winding and the middle, it stays pretty balanced. If you start loading down one 120 volt leg, then the 240 will sag also, causing the regulation to compensate.
JohnTaves Newbie Joined: 01/11/2023 Location: United StatesPosts: 19
Posted: 05:57pm 13 Dec 2023
Copy link to clipboard
Print this post
Ampinvt sent me a new board. It still is oscillating. It seems to be oscillating with a lower amplitude than before, but still large relative to the load. I have not been able to sample the current at the high rate of 4.87ms that I did before, so I cannot prove it is a multiple of 60hz, but I think that is a safe assumption.
I figure that the big capacitors are effectively a low pass filter to eliminate the 120hz noise from the inverter. Is there a significantly different frequency response between lead acid and lithium batteries?
I am wondering if it is possible that Ampinvt has tested this inverter only with lead acid and because lead acid might not have as high a frequency response as LiFePo, they did not see the problem with this design.
No, scratch that. If lead acid cannot keep up with the 120hz demand, then the amplitude of the output would not be sufficient. I mean to say that caps are not just for filtering, but also to provide the current at that frequency.
I think I answered my own question, but figured I would post it anyway.
JohnTaves Newbie Joined: 01/11/2023 Location: United StatesPosts: 19
Posted: 01:22am 07 Jan 2024
Copy link to clipboard
Print this post
I am wondering if this current oscillation matters.
I used this inverter for several weeks where my ammeter (an ina226 chip with an esp32) was sampling and averaging so it wouldn't see the oscillations. I swapped that previous design for a Reidon SSD, which is a digital shunt. At around that time I think the board fried. The replacement board seems a little better, but still I am seeing large amplitude changes with the current reading. My clamp meter shows a steady value.
1) Is this normal? I am wondering if others don't see this because their ammeters are sampling and averaging and my digital shunt is just sampling.
2) Is there an obvious inductor in the design of inverters that filters/smooths out these current oscillations? Can I spot if this inverter has this?
3) The Reidon SSDs are a new product, and 4 of 4 have gone bad on me. Is it possible it is reacting poorly to noise? My ina226 design had a fluctuating V reading, but no fluctuations with the current reading. I assumed it was noise, and solved that with some caps and some averaging. I figured that it was reasonable to have noise on the 1m of wire from the shunt to the ina226 on the voltage gnd/V twisted pair, but less likely on the twisted pair for the voltage drop across the shunt, because the shunt pair is just differential, whereas the V pair could have ground complications. This new Reidon SSD is a self contained thing. It has no lengthly wires to pick up noise. Is it logical to assume that the current measurement is unlikely to be garbage/noise? I should not see large swings in the current, right? I mean the ripple current will be at most some single digit % of the base current, right?
Thanks for your thoughts.
jt
phil99 Guru Joined: 11/02/2018 Location: AustraliaPosts: 2155
Posted: 04:47am 07 Jan 2024
Copy link to clipboard
Print this post
No, it should be a lot, ripple is normal. With no filter caps on the DC input to the inverter the 120Hz P-P ripple will be 200% of the average value. With very large caps close to the MOSFETs you might get it below 50%.
Averaging is used because a lot of ripple is normal. Edited 2024-01-07 15:01 by phil99
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1685
Posted: 11:05pm 07 Jan 2024
Copy link to clipboard
Print this post
Yes, when reading a shunt with NO averaging (or processing) of the shunt voltage, you will see positive going excursions from each half cycle of the AC output generation showing on the inverter DC input current.
A DSO will display a 100Hz or 120Hz positive going sinewave on the DC input.
Without averaging or filtering, you are choosing to display the Peak DC input current waveform, you can filter some of the wideband HF noise and other low frequency distortion /mixing products but the current oscillation will always be there unless you turn the inverter off.
Unlikely, but depending on the shunt design more likely over current. However in the past, I have had standard little integrated resistive 50A hall effect shunts handle 300A current pulses and not fail.
BTW: I have head of random failures of this type of device if the Shunt sensor is plugged into the powered Control head "before" the shunt is connected into the DC line.
Yes all ADC input readings obviously need filtering or additional filtering depending on the design of the display unit. There is so much HF noise radiating inside and around the Inverter feed cables that shunt output voltage, even with balanced twisted pair output can have a problem with noise induced into the shunt and not filtered correctly, especially when it's connected into the DC supply to the inverter. Even lower cost fully isolated Hall clamp sensors can have problems.
There should not be large visible swings in current unless the AC Load current is also varying at the same rate.
It also depends on what you mean by seeing "large swings in current" and what display device you are seeing this on?
If it's not a DSO, are the large current swings real (following AC load) or just random digit cycling at each display update due to a serious noise or instability issue.
Any possibility of getting a DSO capture of the DC input? Apology if this has been suggested previously.
FYI: Three 200A DALY BMS units I have will display incorrect swinging DC Currents when the Inverter is running.
The low cost JUNKTEC units I'm using for each battery bank have digital displays with resistive Shunts connected right next to each BMS.
The Juntecks are 100% stable and the current display is also 100% accurate, no matter how high or low the inverter load is.
The J-Tecks appear to have very low frequency <15 Hz active filtering and amplified low impedance outputs - they are completely immune to inverter noise and MPPT regulator noise. Obviously they will never be able show narrow peak current pulses either. . Edited 2024-01-08 11:11 by KeepISIt's all too hard. Mike.
JohnTaves Newbie Joined: 01/11/2023 Location: United StatesPosts: 19
Posted: 07:24pm 09 Jan 2024
Copy link to clipboard
Print this post
OK, thanks, I will conclude the inverter is fine. It does have large caps next to the MOSFETs. I am getting maybe peak current of 80% of the load. Maybe it is going from 0 to 2x.
I thought that the caps were there to smooth this out, or rather to provide the current for the output peaks and absorb the current while crossing zero. However, I suspect my thinking is too simplistic. The battery, depending on its chemistry, might also be capable of handling that. Are the capacitors there in case the battery cannot handle that frequency?
Is it correct to say that a LiFePo4 battery will have lower inductance than a lead acid, and therefore the capacitors are needed for the lead acid and not so much for the LiFePo4? Couldn't there be some resonant frequency response problem with the caps and the battery?
I thinking that my fears of 2x the current causing problems with my circuit breakers and wires is unfounded, because the issue is heat. With an oscillating current the heat is going to be proportional to the average current, not the peak.
Thanks for your help.
jt
JohnTaves Newbie Joined: 01/11/2023 Location: United StatesPosts: 19
Posted: 07:42pm 09 Jan 2024
Copy link to clipboard
Print this post
These two statements, above, seem contradictory.
I have a digital shunt (Reidon SSD). I think it is measuring the current with some hall effect, because when I take it apart I don't see the resistors between the two sides of it. It is sampling the current and I display that momentary current.
I am thinking that your two seemingly contradictory quotes above are saying that the current is going to oscillate and the peaks are OK from nearly 0 current to 2x the output AC load current, but generally an ammeter is not going to return that detail, so I should not see those swings.
I don't have a DSO, however, the Reidon SSD is providing me with instantaneous digital (D) amperage readings so all I need to do is store (S) those values, and display them and voila, I have a DSO.
-------------------- In summary, I am concluding that I will program the Reidon to average the readings over say 500ms (>> 16ms) thus ignoring the oscillations that I mistakenly assumed would have been smoothed out by the inverter.
I appreciate all your help.
jt
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1685
Posted: 10:21pm 09 Jan 2024
Copy link to clipboard
Print this post
What you are saying is correct.
Yes they do, and yes, I assumed that any standard DC display, would by design, not respond much to a 100Hz sinewave current peaks in a DC line.
My DALY BMS shows current swinging between -8A and +15A when measuring a -5A current, it's caused by slow data update of the BMS as it grabs two noise affected samples at separate points in time. The software / firmware makes the meter display swing, and numeric display cleanly switch (sample hold), between those two incorrect random points. Obviously it's a garbled mix of noise that a BMS should not respond to, unless they thought the BMS would never be connected to an AC Inverter It's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1685
Posted: 11:22pm 09 Jan 2024
Copy link to clipboard
Print this post
I've seen some sensors ("some") with two heavy connections to a substrate and a resistive / hall element inside (looks a bit like a printed circuit track), high over current and the internal element blows like a fuse.
Obviously transient response depends on the type of processing, internal and/or external included for noise reduction, a few have link selected time constant for the desired bandwidth response of the filter.It's all too hard. Mike.
Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.
Print this post
