Wireless Mains Adapter-Frequency etc.?

[pwdrive]

Try altering the coil tap in favour of the drain ie: more turns for the drain winding.

Lawrence.

[Skywave]

The series-connected capacitor to the gate may have too little reactance - thus loading down the feedback path: a smaller capacity capacitor might help. Plus Lawrence's suggestion.

Al.

[Refugee]

Perhaps try a start up circuit.
A 100uf capacitor in series with a 47K to 100K resistor connected from the power supply rail to the gate.
This should give the gate a kick when power is applied and then as the capacitor charges it will float up and leave a little bit more gate loading. With 8.2K already there the effect of another 47K can be safely ignored.

[boiss]

OK Refugee,
I'll give it a whirl tomorrow. Today I tried adding more turns on the drain side and reducing the gate feedback cap to 6nF. Also tried increasing the Gate R to 220K in steps and then also reducing it to 47 Ohms. All to no avail. One thing I noticed which I don't understand is this. I connected a variable stepped resistor between +12V and the gate. With no resistor the gate voltage was 0.27V, adding the resistor starting at 100K and reducing it to 10K the gate voltage only rose to 0.5V, Why didn't it reach about 5 to 6V, does the gate take I ??
At present to start the cct it is necessary to apply a positive pulse to the gate, which means that your idea looks to be on the right track.
Mike

[Refugee]

Perhaps the gate has an ESD protection diode connected internally.

[boiss]

Perhaps the gate has an ESD protection diode connected internally.

Hmmm, It doesn't look like it! With the cct not oscillating (checked with the scope) Why should there be a + voltage on the gate when there is no connection to +Ve, this voltage is maintained with Rgs=1k Vgs=0.2V,with Rgs=100 ohm Vgs=50mV. Is this normal for Fets? I posted the datasheet for the IRF2807 a few pages back.
Mike

[Refugee]

I will have to power one up and see what happens.
I do have an HT power supply on the home brew list. I intend to include a cap reformer in it.
FETs in both constant currant and constant voltage will be inside it.

[boiss]

Going Loopy here.tried the mod with the cap and R between gate and +Ve, nada !!
Tried different values of C and R , Nada!!!
Eliminated the cap and zapped the gate with ever smaller Rs connected to +Ve.
Went down to 1K. No go!!!. With 1K permanently between +Ve and gate, the gate volts are +3V NO change on the drain. If I connect the small charged cap .22uF across the 0.22 gate cap the cct starts and runs perfectly..................................
Does anyone have any clue as to what is going on here ?????????
Mike

[pwdrive]

Looking at your latest circuit it shows the coil as 13 turns? not a lot of L considering the value of the capacitor, I don't know what the effective diameter of your coil is but as an example 1.5" diam. @ 13 turns would normally have a capacitor in the Pf range to form a tuned circuit and it's resonant frequency would be several Mhz. Try increasing the resonant frequency somewhat.

Lawrence.

[boiss]

Hi Lawrence,
I have tried up to 420KHz with the present 12T coil and with the others 12T, 18T and 30T and this is by far the best. It runs when I can get it to start at 30KHz, and will give me 2 x 2.5W at about 2cm separation which is more than I need.
Raising the frequency to about 80KHz only gives me about 50% of the power.
Is it possible that the Fet is damaged in some way but still works? Looking at the curves shows no current under 4.5V Gs is this correct? I always thought that Fets behaved like valves. This IRF2807 appears to sink 9ma (1K Resistor +Ve to gate)
for there to be only 3V on the gate and the device appears to still be off....
Mike
PS,
Some of the stuff Nick directed me to had single turn coils and ran at around 1MHz but were not practical for this purpose, and RFI ??
Just disconnected the FET, It measures 1.2K Gate to Source And Drain to Source measures about 1K.
Obviously it is stuffed but why It should still work beats me. I will try a few power BJTs