You have me stumped. Please post pictures and Schematics of the working version once it is seen.

 

I'm simply lost for words here because this sounds like a perpetual motion machine.

 

I'm trying to look up the sensor lab, this is what I got...

http://www.radioshack.com/search/index.jsp?kwCatId=&kw=28-278&origkw=28-278&sr=1

Could this be what you're talking about?

http://www.radioshack.com/search/index.jsp?kwCatId=&kw=sensor%20lab&origkw=sensor%20lab&sr=1

Something that surprises me, why haven't you shown any schematics of what you've tried? It would give us a starting point, if nothing else.

 

Yes the 2nd one is what i have to work with I have nothing close to working to show the closest Ive got is the tank circuit on this site

 

OK, I looked for a manual for it online, couldn't find one. As a suggestion draw something up, and we'll try to help you tweak it. Go with a Colpit's or Armstrong Oscillator (both are on Wikipedia).

The picture gives very little to work with as to component count.

 

Look at www.falstad.com/circuit/index.html I tryed it with a 10 cap & a 100 ind 2 res @ 2.2 ea & 2 LEDs cant get sine wave right plus tryed other size cap & ind

the best ive got is 10 cap 1h ind & 2.2 res & 2 LEDs come close to work

 

The site's animation is misleading. Some of the energy in the system gets converted to heat in the resistor, and so the oscillations damp out and cease fairly rapidly. There is no steady-state operation. Your setup might produce a flash or two, depending on how they are in circuit.

 

I know I will lose energy from the two resistors & LEDs & some from the cap & ind I need the right #s to get to run the longest & what order to put the compontants

 

I think I would build a colpitts or an armstrong oscillator, stick two LEDs in series with the capacitors and call it done. Put two electrolitics back to back in series, which their positive leads connected and you have a large non polarized capacitor. This might help you get the frequency down to visible ranges.