Common batteries are free energy sources

[mondrasek]

I realized I would likely not be able to balance the switching to 50-50 since the solenoid in the relay was assisted by the spring that closes the normally closed set of contacts.  So the original arrangement fights the spring while switching from +9 V --> -9 V, but the spring is helping on the switch back the other way.  I had to modify the relay.

First thing was to eliminate the spring.  But this spring also holds the switch pivot point together.  I glued a tiny aluminum tube to the contact arm as close to the original pivot as I could manage and drilled a matching pin hole through both sides of the relay's clear plastic cover.  I now have a straight pin going through the relay cover and through the tube, creating a new hinge.  So the spring can be removed without everything falling apart.

After various trials of adding a magnet to the metal back of the contact plate, and moving the solenoid to different distances away, I scrapped the entire idea of using the original coil with it's metal core.  I couldn't balance the attraction of the new magnet to this core anyway, so I drilled out the back of the solenoid where the core was attached and removed it.  Now the solenoid has an air core.  After assembling everything I pushed a few small neo magnets down the coil from the back until they attached to the contact switch plate.  The result is a relay that pushes and pulls the contacts via it's reversing magnetic fields as it switches from +9 V --> -9 V and back.  It is very stable compared to the original relay and the output wave form is as perfectly balance 50-50 as I can tell on my o-scope.  I'm letting it "burn in" in the garage right now and will then charge the (4) x 9 V NiCds back to max over night before trying any run down tests.

I can see chatter and arcing very minimally on the wave form as the contacts open (I assume), so I'd like an answer to the questions about adding caps to suppress this behavior if anyone has that knowledge.  I have no idea how to select the best caps for the task if it would even work and/or be beneficial.

Thanks,

M.

[allcanadian]

@mondrasek
If you could post a circuit diagram of what you have done that would help, concerning your relay, if you add capacitance across the relay you have basically closed the circuit to some extent and this will dampen the on/off rise/fall times. An easy way to correct your relay is to solder a small piece of copper wire to the top of the relay. This adds mass thus momentum to the part of the relay that is moving so the contacts move fully from one set of contacts to the other, this slows the relay a bit but does not dampen the sharp impulses that are required in some circuits.You can also stretch your spring a bit to help extend the contact motion.

[mondrasek]

@allcanadian,

Thanks.  This is really frustrating for me since I have noone I know that I can discuss this stuff with in person or by phone.

I've attached the (admittedly poor quality) scan of the copy of what I drew up at work on Friday.  I can do better later, but it's what I used to wire the relay.

I've already been using your sugestion for adding mass to the relay switch arm to try and get more stable switching from the start ( I'm an ME so that part was easy).  I've been attaching a Radio Shack alligator clip to the tab on the top of the arm where the spring used to hook.  That's how I have run this for 90% of the time.  It runs right at about 50 Hz.  Without the clip it goes much faster, I think at one cycle every .4 msec, if I remember right.  But at that speed it is more prone to stopping.  It's more out of control.

I can see the arcing across the contacts.  I assume that can't be good (another loss).  But I'm not sure what to do about or with it.

Ultimately I would like to make a new solenoid to try and tune this and get bemf output for additional work.  I have no idea how to calculate the optimum solenoid design for that, but I have more magnet wire arriving tomorrow.  Hopefully I can get more help from the forums for that idea as well.

Thanks again,

M.

[mondrasek]

Man, I'm a lowsy photog!  But here you can kinda see the wave form for the switching voltage.  You'll have to take my word that is going from +9 V to -9 V.  The scope is an Iwatsu SS-5802 I borrowed from work.

I curious about the decay ramp on each peak.  It appears about the mid point of the peaks regardless of what frequency the added weights to the contact arm have caused.  Is this the point where the contacts have opened and the cap is the only remaining voltage?

M.

[allcanadian]

@mondrasek

I curious about the decay ramp on each peak.  It appears about the mid point of the peaks regardless of what frequency the added weights to the contact arm have caused.  Is this the point where the contacts have opened and the cap is the only remaining voltage?

I would attribute the decay ramp to capacitance, energy is being stored before the voltage is allowed to change and capacitance will do this, the larger the capacitance the longer the decay ramp. The contact arcing can be minimized by even a very small capacitance which is preferable. I see your setup is a typical series==parallel  parallel==series setup,  one thing to consider with this setup is the type of load you wish to power. If the load is inductive (coils, motors, solenoids) then the contacts will arc wildly without a capacitance across the contacts. If you can imagine an electric current as having momentum then when you open the contacts this momentum is applied instantly across the contacts gap and could be thousands of volts. You can connect a capacitor across the load if you plan on running inductive loads, a better way is to connect a full wave rectifier where your load is in the diagram--then run your inductive load from the (+)(-) connections on the rectifier, you can then put a steering diode across the load to route the inductive discharge back through the load.