Well things never go according to plan... this is no exception.
The unit works very well as a stand alone unit, which was what was intended by the manufacturers of the EG8010 chips...... and we proved they were right, as a stand alone power unit up to a 4 or 5 or more kw it is stable and very useful.
Given more fets and bigger transformer, it gets to be very useful, and will easily run any off grid house and a lot more.
A lot of testing was done with local loads up to and well above 10kw peaks, and 5kw continuous for 20 mins or more etc etc.... unit performs very well with very good voltage stability.
Even as loads are added in the multi kw range, it is within a volt or so... and really perform well above expectations..... and it is about here the wheels fell off.
Wet weather has arrived, and with it a change to operability?
?? how so?
Local loads of up to 5kw or more still seemed fine, but if I connected the house to it now... even without flipping on the connection o/load... ie no load and only the earth/neutral connected... the voltage went haywire.. from 240 to 120 and a few places in between depending what and how you played with it... and here was why.
This is a problem. No ,matter how I tried, I could not get the voltage to be stable from idle unconnected to idle connected.
If I connected and THEN used the trimmer to get the 240v, it worked reasonable well with it's stability.... now down to 10v or so.... but stll I was not amused.
What went wrong, and why now and not before... well here it is.
This is the EG8010 circuit for the AC voltage sensing.
Notice the voltage divider formed from the 100k resistors, and then they ground one side , and sense the other..... I always wondered about the stability of this, but discounted my fears as it worked so well on the local loads... and in the dry, the house loads were only a little bit out, and once corrected, seemed fine.. but the wet weather changed all that...why?
Because the battery gets wet in my setup, and the panels get wet, as does everything else..... and this gives as a very high impedance to ground for the battery bank. Those folks with grounded neg banks would find this out from the get go.... mine are not.
So it appeared that the DC grounding stuffs up the AC voltage sensing.... so a change is required.
Luckily very simple, and Clockmans board has plenty of room at the end we are interested in... we simply get rid of the offending voltage divider and replace it with a small as we can get transformer, and this gives us a floating reference not bothered by anything else.. and suddenly it is very solid, regardless of the load situation, or the battery situation. As can be guessed, a portable battery on a protable installation would not couse these problems, but htis fixes that anyway.
It looks like this:
and this shows how i put the output directly where the 50k resistors went... before the bridge.
Now it performs without any bad habits.
.......oztules