Simple to build isolation transformer that consumes less power than it gives out

[JouleSeeker]

The isolation trafos that I ordered arrived, see photo and schematic below.
I did some initial measurements with one trafo using a 40W incandescent bulb as a load.  I put the bulb in my light box so I could also keep track of light output in lux.

Input from mains, 122V:
  46W input,  87.6V out at 36.2W, eff = 79%  (so-so), 2050 lux
  Turn light off (open secondary, no load):  2.8W idle

Input from variac, 100V:
   33.8W input,  70.2V out at 26.1W, eff = 77%  (so-so), lux 800
   Turn light off (open secondary, no load):  2.0W idle

Looking at the schematic, the input goes to the red wires in this case;
for the output I used the red/blk and green/blk wires.

Next,  I did some re-wiring to use both available coils as primary; thus connecting blk and red/blk also connecting yellow and green/blk -- to put these as the primary coils in parallel.

Then the red wires become the output.

The efficiency improves using both coils as primary:

Input from variac, 100V:
   52.9W input,  98.9V out at 44.2W, eff =84%  (better), lux 3480
Vout/Vin = 0.989, nearly unity.

Input from variac, 110V:
   60.6W input,  108.3V out at 50.4W, eff = 83%  (good), lux 4810
   Turn light off (open secondary, no load):  2.6W idle

I will soon proceed to build Jack's 2-trafo design.

[gyulasun]

Hi baroutologos,

I also disagree with using a diode in series with a mains transformer primary coil, the simply reason is the diode does a half wave rectification and the created DC current in the coil creates significant copper loss AND may bias the core towards saturation (as you mentioned): this may reduce the self inductance of the primary coil too, increasing input AC current.

Regarding your measurements on transformers, it is indeed strange and I also found big differences in measured no load input current and the calculated input current which came from measured transformer parameters. 

The problem may come partly from the digital LC meter: it does not use 50 or 60 Hz test frequency for L measurements, my own LC meter (Maxwell DMM MX-25 304 old type) uses about 200 Hz in the some hundred milliHenry and Henry ranges and it is doubtful how the different mains transformer cores perform at such a "high" frequency, most cores may lose permeability to some percent but some other cores may lose even half of their '50 Hz' permeability.  This means that a 50 Hz test circuit should be used  for measuring the transformer coils.  I repeat this frequency difference does not fully explain the situation.

However, here is a link which may shed some more light onto this problem:
http://sound.westhost.com/articles/xfmr-dc.htm 

It deals with DC component appearing across the AC line. The DC component can occur for instance from (say) the next door neigbour's hair dryer which uses a series diode in its control switch to adjust heat power, see Figure 1 and Table 1 and also Figure 2.  There are some other 'nasty' appliances that can 'distort' the pure AC voltage by adding a DC shift to the line.

I wonder if you could use a decent 1:1 isolation mains transformer and drive from it the primary of your 200VA trafo and check the idle current, would you find similar 20-30 mA idle current (assuming the secondary output of the 1:1 transformer gives also 220V and not 210V or other different output).

On your example 2: partly the output impedance of the variac plays a role also in the resultant current and it is in series with your secondary trafo coil of course, reducing the current but the difference you measured is rather big and it cannot fully explain the phenomena.

Will ponder on this and if I have some further thoughts, will return.

rgds,  Gyula

PS:  In series resistor - inductance circuits like a transformer coil represents the coil's DC resistance is also to be considered, for a 200VA trafo the primary coil may have 10 -20 Ohm DC resistance, you surely know the formula for such cases:
( http://www.sweethaven02.com/ModElec/DcAc/acee/equ0704.gif )from this link: http://www.sweethaven.com/sweethaven/ModElec/acee/frm0702.htm  (I know that in case the inductive reactance is much higher than copper resistance, then the latter does not count much.)

[gyulasun]

Hi Steve,

Normally a decent 1:1 mains isolating transformer has a single primary and a single secondary coil and they are labeled to identify for the user which is which.  The reason is that in the secondary coil the number of turns should be a few percent higher wrt that of the primary to "compensate" for all the losses at a given output power range, this means that the DC ohmic resistance of the secondary should be higher than that of the primary coil.  And if you happen to operate this 1:1 transformer backwards, naturally the original primary coil which is now the secondary will give way less output than used in the designed direction.  In the few percent higher secondary coil turns the approximate efficiency of the transformer manifests (i.e. assuming a 96% efficiency for a decent 1:1 mains transformer (above 100VA power level) the number of turns for its secondary coil is increased by at least 4% wrt the primary turns).

So I suggest to measure the DC resistance of all the 3 coils. Probably the way you connected the two coils in parallel is good but in this present situation where you measured 84%, try to use this backwards too to see how efficiency changes if it does at all.

rgds,  Gyula

[T-1000]

The isolation trafos that I ordered arrived, see photo and schematic below.
I did some initial measurements with one trafo using a 40W incandescent bulb as a load.  I put the bulb in my light box so I could also keep track of light output in lux.

Thanks for a try.

The whole meaning of this experiment is to force condition where you make oscillating resonant circuit between 2 transformers + capacitor and when it happens, it should start pushing power back to the source. So your ampmeter will show least consumption as possible of both trafos. If phase is 180 degrees obviously the ampmeter will show maximum consumption of power rated on first transformer where no load is attached. The power taking out of resonant circuit without killing resonance is another matter and should be governed by opposite than Lenz law.  And the losses of transformers should be compensated by primary power source. When you get resonance with correct phase shift you only neeed support resonant circuit running while current in second trafo between capacitor and windings is on maximum level.

Cheers!

[NerzhDishual]

Hi Baroutologos,

Very, very interesting experiments and remarks.
Actually, IMO,  your LC meter should be right!
What do you mean by idle current? (= With no load?)
Could you give some simple schematic?
I have a variac, a LC meter and some trafos and I would be pleased to
reproduce your measurements.
-----------------
I was in touch with a (French) guy who claims to get 'OU' out of trafos.
He uses 230 V/60 HZ AC grid current, tri-phases (often hand wound) trafos
and also 'off the shelf' 'normal' (2 or 3 phases) ones.
BTW: he sometimes also uses diodes.

Unfortunately, he was very obscure and refused to perform some measurements.
IMO, this person have some huge communicating problems (kinda schizophrenia issues).

So, I gave up, but I'm still believing that he is really into something and that
it remains some hidden secrets in trafos...
-------------------
BTW, thanks also to Prof S. Jones for his indefatigable experiments/measurements.
For my part, I'm more a 'contemplative' individual.
Anyway, I have ordered  two 65VA (2*115/230 volts) trafos...

Very Best,
Jean