 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. |
Forum Index : Electronics : Measuring system currents
| Author | Message | ||||
| davef Guru  Joined: 14/05/2006 Location: New ZealandPosts: 499 |
My off-grid system has two charging sources and one load (pure sinewave inverter). I have made several attempts at measuring input and output power levels and always end up with about 10% more output power than input power! Considering battery losses and inverter efficiencies I would expect about 80% overall. One charging source is a rectified 3 phase AC from the microhydro through a LTC6101 hi-side current sense IC. The other charging source is the output from a MPPT solar controller through another LTC6101 sensor. Whereas, the waveform from the microhydro is a constant DC level with some 3 phase ripple on it, the output from the solar controller appears to be pulses of DC into the battery. Now, the input current waveform for the inverter was a bit of a surprise as well. At lowish currents it looks like a 100Hz sinewave going from 0 to about twice the average current. At higher loads this "sinewave" becomes distorted. The inverter current is also measured by a LTC6101 and a 100A shunt. I have all these sensors outputs feeding buffers followed by a RC network (1K series and 1000uF to ground) ... originally put in to filter out any AC components. I am now wondering if this is the right way to get a reasonable current measurement. I was hoping that the RC network would average any waveform shape to the equivalent average DC current. Would appreciate any direction on doing current measurements under pulsating DC conditions. |
||||
Downwind Guru Joined: 09/09/2009 Location: AustraliaPosts: 2333 |
Are your shunts low side or high side shunts? Low side shunts are known to cause problems as it can cause a difference to ground and cause ground loop problems, as any resistance in the ground line can cause problems to refference to ground, as almost all electronic measurements work to a ground reference. As for your RC filter, remember with AC the filter will charge and discharge with the AC wave, so perhaps a series blocking diode might help with a better average output. The 100Hz on the inverter seem s a bit strange and i can only think you are somehow reading a double reading, as you would expect to see a 50Hz ripple on the inverter current. Placing a AC tong meter on the battey leads to the inverter will give a amp reading, even though its a so called DC supply, the ripple is strong enough to use a AC meter. Your information is a bit patchy so its hard to understand the actual problem. Pete. Sometimes it just works |
||||
| davef Guru Joined: 14/05/2006 Location: New ZealandPosts: 499 |
Pete, Hi-side side shunts. How the RC filter responds to input waveform I think is where things are coming adrift. Further reading suggests one should sample the waveform and do some maths on in to get a true RMS, which would be equivalent to a "pure DC heating value". Maybe, the RC filter is averaging the waveform. The resultant value would also depend on the sink and source capability of the buffer driving it. Average on a pure sine wave is .637 compared with .707 RMS. See here This would explain most of the difference. I'll re-measure the inverter current waveform, but I am sure I saw a 10ms period. Thanks for your response. Dave |
||||
| Downwind Guru Joined: 09/09/2009 Location: AustraliaPosts: 2333 |
As i have no idea of what system you are using to monitor the data from the sensors, so its a little hard to comment on what method to suggest to improve your readings. One suggestion rather than trying to compensate in software for a AC signal would be to convert the AC signal to DC, where it can be buffered and smoothed to give a steady response, its how i would tackle the problem. Working with AC and trying to compensate for what part of the wave is tested for a result is a can of worms in my view, as it all comes down to timing, and if your timing is not spot on then the results can vary. Much simpler to use a DC method where timing is not so important, as the data is always available, albeit there can be a small lag time, but far less hit and miss than trying to work with AC signals in my view. Pete. Sometimes it just works |
||||
| davef Guru Joined: 14/05/2006 Location: New ZealandPosts: 499 |
Pete, Just heading over to the bach to do some more measurements. Got the oscilloscope with me to do a more careful investigation. There are no AC signals, just highly pulsating DC ones! Which are filtered in a RC network. So, there is fairly constant DC being presented to ADC inputs. I will report back in a few days. Thanks for helping. Dave |
||||
| davef Guru Joined: 14/05/2006 Location: New ZealandPosts: 499 |
Results: The input current to the inverter is pulsating DC at a 100Hz rate. I have been reassured that a simple RC filter will average any pulsating DC waveform to the average value ... if the values are chosen appropriately. I have now got the solar MPPT controller back in the system rather than diverting excess power to dump loads. Utilisation is now 70% or so, which is more in line with expectations. Rather than 3 separate sensors I am going to look at bi-directional hi-side current sensor just on the positive lead of the battery. This is the most important parameter to measure anyway. Correction: When the the solar controller is in boost mode it is a reasonably clean DC level. Maybe, under float conditions it might be more of a pulsed output. |
||||
Don B Senior Member  Joined: 27/09/2008 Location: AustraliaPosts: 190 |
Hi Dave F, Trying to compare power in circuits with anything other than steady state DC has its problems. Even pure sine wave AC needs to take account of the power factor (the angle between the voltage and the current). What you need are true RMS (root mean square) values. Some multi-meters are designed to give true RMS readings, but even these have limitations. I suggest a little reading in this area will help you to make more sense of it. Regards Don B |
||||
| davef Guru Joined: 14/05/2006 Location: New ZealandPosts: 499 |
Don, I didn't make the connection that power factor was relevant. As far as the inverter power is concerned I can see that I do need to think again. Would the same have to be done on the charging sources, before the power is dumped into the battery? The inverter input current, over most power levels, is a 100Hz fairly clean sine wave going from zero to twice the average value. A higher powers I was seeing some distortion on the current waveform so I thought of trying to do some real-time RMS calculations on the current and voltage at discrete time intervals. Not for power factor reasons, but because I realised that voltage applied to the inverter was slightly dependent on the current it drew. I think I could take 10-20 samples each cycle. Would measuring the current and voltage at discrete time intervals give me more correct power estimates? Or does the maths get complicated? Thank you for pointing out a significant issue. Dave |
||||
| davef Guru Joined: 14/05/2006 Location: New ZealandPosts: 499 |
Don, On second thought ... if the supply voltage to the inverter did NOT change over the 100Hz cycle caused by the inverter then I don't see why power factor is relevant. However, when sucking a kilowatt or two out of my old batteries the voltage can drop from 28V down to 22-23V. With this relatively large change in voltage you then need to consider power factor. Correct? Off to Wikipedia Dave |
||||
| davef Guru Joined: 14/05/2006 Location: New ZealandPosts: 499 |
Further testing and observing the waveforms with a dual-channel 'scope show that there is about 35-40 degrees of phase shift between the current and voltage. With a 1KW load the ripple on the battery voltage is 1.5V p/p on a mean value of 26V. As well as more reading looks like it is time to brush on up on me maths. |
||||
| Don B Senior Member Joined: 27/09/2008 Location: AustraliaPosts: 190 |
Hi again Davef Sorry to have taken so long to get back to you. As you may, by now, realize if you have done any delving, trying to get an accurate power reading when both the Voltage E and Current I are continuously varying is a difficult exercise, usually requiring some specialized electronics. If you are dealing with DC only, then you need to be constantly calculating and summing (ie integrating) the instantaneous values of E times I. If there is any AC involved, then you need to be constantly calculating and summing E times I times cos (ie cosine of) phi, where phi is the angle between the voltage and current waveforms. As you suggest, your filtering elements are probably modifying your readings and introducing discrepancies. The only other thought that might help is that, irrespective of the wave-form, the heating effect of the current when passed through a (non-inductive) resistor (ie the temperature rise) will be the same for identical current flows irrespective of wave-form. This principle was used for measuring current in HF ammeters, which were really just a thermocouple powered temperature gauge calibrated in amps. Keep delving. Regards Don B |
||||
| davef Guru Joined: 14/05/2006 Location: New ZealandPosts: 499 |
Don, That's fine. Gave me a chance to do a bit of research and to perform some more measurements. With a 1KW load the ripple on 25V is about 1.5V p/p. This is about +/-3%. By measuring the average DC voltage and average current to my inverter (with a PF not equal to 1) I am really measuring apparent power. And apparent power will be higher than true power in a system with PF not equal to 1. As a rough guess I am over-estimating my input power by a "small percentage". My daily untilisation is never above 70%. As the inverter efficiency is about 90%, I conclude that battery efficiency is about 80%. I initially thought this was quite low, but possibly it is indicative of lead-acid batteries at the end of their service life. Currently looking at importing a set of LiFePo4. Cheers, Dave |
||||
powerednut Senior Member Joined: 09/12/2009 Location: AustraliaPosts: 221 |
Cool trick, thanks for the tip Don. |
||||
 Print this page |
