Guru
Joined: 26/07/2007
Location: AustraliaPosts: 1686
| Posted: 12:37pm 03 Nov 2017 |
Damn you do some interesting stuff. Am following your exploits.
I assume it will sync to the mains, and this should be nice, but when you lose the mains ( switch off from mains to inverter ), I think you will need to do some interesting ...something.... to avoid sudden switching pulses hitting the torroid. We know that this can be fatal... wondering how you can implement that, or even off grid usage without the grid running in parallel ( using it for freq)
I will keep watching with much interest.
Thanks Poida.
On another note, I think the discussion around the VFB is mostly pointless apart from knowing when it actually looks at the VFB voltage as an academic exercise. It wants to see a voltage with which to compare so as to set a pulse width. It does not care about anything else but the amplitude it is given whenever it looks.
If we get too cute with pulse VFB trying to match the read points, it will be most likely unstable, and may well lead to total instability.
It is best to just rectify the output sample, and chose a 4uf or whatever you want to set the hysteresis you need for sensible operation.
I found that the best for me is in the 4uf with the resistors on the board ( somewhere here). This gives very fast control of the output voltage ( even 5kw dumps don't seem to bother the lights), and most importantly, very stable output.
Tinkers comment re: noise is very relevant, as any downward inflection at the read time will produce what may be a disproportionately wide pulse width... leading to unstable operation, and maybe fet damage for no good reason.
I don't think trying to synchronize to the read point is wise in any way. I think it is being over thought, for what is really a very simple VFB requirement.... just my thoughts on it.
........oztules
Village idiot...or... just another hack out of his depth