I’m wondering if it is possible to use the positive half of the voltage signal to power the emontx and at the same time use the negative half to recover the complete voltage signal for voltage reference purpose.
in this way the driven current from the transformer doesn’t going to disturb the negative half of voltage signal so it remains clean for power measurements purposes.
I assumed in this proposal, that the positive and negative halves of the voltage signal are always identical and contain redundant information about amplitude and phase.
this improvement needs some modifications in emonTx three phase firmware.
I can well envisage situations where that assumption is not justified.
Before you embark on this, have you carefully evaluated the effect that powering the emonTx has on the voltage wave, and compared it to other distortions that are present?
If accuracy is your principal aim, then you should use a separate 5 V d.c. power supply to supply the emonTx, and probably not use the standard a.c. adapter, but build something like GIVMon or a capacitive “transformer” (search the Archive, and note the health warnings).
This does not apply only to the 3-phase sketch, but to any emonTx that uses the a.c. adapter as the power supply.
Thank you for your reply ,I read about GIVMon and it’s interesting 
I was trying to get rid of Dc power supply so installation become easier
That in fact was the whole purpose in adding the half-wave rectified power supply circuit to the emonTx V3.2. Its predecessor, the emonTx V2, always needed a separate 5 V supply.
As is always the case with engineering in the real world, some compromise is always necessary. Evaluation of the size of the error introduced by the current pulse showed that, in most cases, it was likely to be small compared with other sources of error in the system.
In fact I am managing to add Wifi module to emonTx and I think that error will increase so I am trying not to use Dc power supply , in this intention I have another idea I would like to share
what about using a kind of controlled power gate at the input of power supply circuit “transistor” ,so we can shut it down when sampling process taking place and reactivate it between sampling durations
we can use big capacitor or replace linear regulator with step down converter so that reduces current driven from capacitor and insure power stability for emonTX
Is this idea worth to spend time for search and implementation ?
You cannot power a Wi-Fi module via the a.c. adapter. The current required (upto 250 mA) is far too great. The circuit is designed to limit the current so as not to distort the wave shape too much, and the value it is limited to is barely enough for the RFM69CW module.
I think your ‘gating’ idea has been looked at before (search the Archive) but I don’t know whether it was ever implemented. If an emonTx with the WiFi adapter becomes main-stream, then clearly the developers would wish to investigate all options including that. But a note of caution: I believe one long-term objective is to replace the ‘discrete sample’ sketches with continuous monitoring. If that happens, the gating approach cannot be used.
And if you were to go back to your original proposal, you’d need to carefully consider the effect of a large d.c. component on the a.c. adapter’s operation.