The third iteration of the relay is up and running very well.  Wave form across the solenoid coil is very symmetric.  I cannot see any asymmetry but the traces are difficult to analyze on the o-scope right now.  The frequency of the switching is regulated by the length of the switch arm.  It is working like a seesaw now, with one side of the pivot having the switch contacts and being of a length very similar to the original relay.  The other side was left long (about 5 inches) and can be shortened to increase the frequency.  With the current length it is switching closer to 20 Hz and so while the o-scope traces look very symmetric, they may be off by smaller time scales than are noticeable now.

I started the relay running at about 6 PM last night.  It was still running as of 8 this morning when I shut it down.  I just let it run to mechanically "burn in".  The batteries were not fully charged when I started it and were deffinitely not charged equally.  I just fast charged them for a short duration to get them up to ~9 V to check out the new relay.  When it ran so well, I just let it run.  I checked one battery at that time and it was at 9.13 V (still not sure why these HV spike "conditioned" NiCds charge and run above their expected 8.4V rating).  It had dropped to 9.06 V later in the evening when I was shutting down the o-scope for the night.  I also disconnected the scope probe since I had started to notice a loading effect due to the probes on my last build (I think).  I didn't expect it to still be running in the morning!

I turned on the o-scope for a second just to see the wave form this morning while preparing for work.  The trace was still symmetric, but was only reading about 1/2 way to +9 and -9 V.  Also had a higher spike on the leading edge.  I did a quick check of the o-scope settings to make sure I hadn't left it on a different V/div setting and don't think it was.  I expected the batteries to have run down and think that is what I was seeing. 

So then I disconnected all the batteries so I could put them on the charger and hopefully equalize and top them off while I'm at work today.  But before I did I checked their voltages.  They read 8.91, 8.66, 8.75, and 8.58.  That was strange enough that I took a second to write it down.

So I appear to have a working 4-battery type current syphon that I can experiment with.  Next I need to figure out what kind of experiments to do.  Any suggestions?

M.

I took the time to mount the relay to it's own board rather than in the middle of my work bench so it would be portable and able to be moved out of the way.  I also mounted a DIN rail with terminals so I could wire it up better.  The conductors are all approximately the same length from the batteries to the relay with the exception of the four jumpers involved with the switching from series to parallel and back.  The batteries are all draining more equally now.  Before I would always have one that didn't seem to drain at all, one fully down, and two in between after running to a stop.

I was wondering what to do about selecting a frequency for the relay, as that can be adjusted higher as I trim the upper portion of the switch arm.  Currently it is running at about 28Hz.  While reading about circuit resonance and some of "The Tesla Project" thread I came to the conclusion that my set up is a neat LC tank circuit, being excited by an oscillating +9 V, -9 V square wave.  So I may be able to tune it to resonance or a harmonic by selecting the proper cap and adjusting the frequency of the mechanical switching.

So that leads me to needing to measure the inductance of my relay solenoid.  I understand this can be done experimentally with a function generator, but I don't think I can justify the cost right now and don't have access to one I can borrow for more than a day at a time (if that).  I'm interested in a cheap LCR meter and wanted some advice.  Looking on the net I have these three inexpensive models in mind and hope someone can tell me if they are decent or not:

http://www.amazon.com/dp/B000PHZ5T4?smid=A3IZHOEADOGAP0&tag=yahoo-ce-20&linkCode=asn

http://www.omnitronelectronics.net/phpstore/html/LCM1952-LCR-METER.html

http://www.amazon.com/dp/B000KAEQWS?smid=AMH4W1K8OCGMX&tag=yahoo-tools-mp-20&linkCode=asn

Thanks,

M.

Quote from: mondrasek on September 20, 2008, 08:06:38 AM
....
So that leads me to needing to measure the inductance of my relay solenoid.  I understand this can be done experimentally with a function generator, but I don't think I can justify the cost right now and don't have access to one I can borrow for more than a day at a time (if that).  I'm interested in a cheap LCR meter and wanted some advice.  Looking on the net I have these three inexpensive models in mind and hope someone can tell me if they are decent or not:

http://www.amazon.com/dp/B000PHZ5T4?smid=A3IZHOEADOGAP0&tag=yahoo-ce-20&linkCode=asn

http://www.omnitronelectronics.net/phpstore/html/LCM1952-LCR-METER.html

http://www.amazon.com/dp/B000KAEQWS?smid=AMH4W1K8OCGMX&tag=yahoo-tools-mp-20&linkCode=asn

Thanks,

M.

Hi Mondrasek,

I just sent you a PM for you PM but I may have offered a poor choice on the LCR meter so may I correct myself here on suggesting this type out of your three choices: http://www.omnitronelectronics.net/phpstore/html/LCM1952-LCR-METER.html   This has higher inductance range than the first I suggested and its other features also worth the 50 bucks.

Regards,  Gyula

Notice when you wish to use a relay coil as an inductance to make a resonant circuit, such coils has many turns of thin wire hence the DC resistance is significant and appears as high loss resistance embedded right inside the resonant circuit.  So do not expect a resonant LC circuit with highly selective properties from relay coils but with rather flat response at or near the resonant frequency. If this is not a drawback for an application of your choice than of course do not care but I thought worth mentioning this.

@Gyula,

Thanks for the input on the meter and info on solenoid resistance.  I played around with the solenoid and RLC sims on coilgun.info to estimate the inductance while waiting for my meter to arrive.  With that high of resistance the circuit would definitely be overdamped.

I also went for a different meter since Omnitron wanted ~$20 for shipping.  I found another meter with similar specs that also had the other DMM functionality.  With shipping it came in $5 cheaper.  I need to return the Fluke I borrowed to work anyway.

So I'm not sure if tuning the solenoid control circuit to resonance will do much.  The mechanical frequency needed for smooth operation might be more important.  But maybe I can get both synced up.  When I try to raise the frequency by shortening the top of the switch arm it becomes unstable and beats itself to death.  I quess this is due to the switch frequencies relation to the natural mechanical frequency.  It's tuned close to that natural mechanical frequency now at around 14Hz after adding more weight to the top of the switch arm (allegator clip again).  It sounds close to the same frequency whether excited by the batteies or if I rattle it just by hand without the bateries hooked up.

I also realized that trying to raise the capacitance to get a square wave like form is proably counterproductive.  Instead I want the contacts to make only long enough to generate a magnetic field strong enough to repel the switch arm to the other set of contacts.  The cap is there to keep the magnetic field in place across the gap in the switch.  The minimum capacitance to do so would be best.  If the cap is down to zero volts by the time the opposite set of contacts is made (or before), that would waste the least amount of current.  Actually it would be best if the cap could be removed all together, but I don't think that is possible with this set up.  Though I can see it in my mind with a bigger set up.  It becomes a magnetic pulse assisted pendulum.

I removed the thin wire from a spare solenoid and wrapped it with bifilar 34 gage just for some testing (while waiting on the LCR meter again).  It rings out at 28.6 Ohms per coil.  When in series with the existing circuit it has almost no noticable effect on the wave form (unlike the 1:1 phone transformer).  It also outputs great bemf spikes that charge a cap to ~86 V.  The 1:1 transformer would not go past 30-something V.  Interesting thing is once I ran the numbers for this solenoid on coilgun.info it appears to have almost exactly the same magnetic strength as the original coil.  Blind luck there!  I wanted to test it as the switch coil since I am not convinced that the second set of coils cannot be used to extract the bemf as was mentioned ealier.

M.

Hi!

Mondrasek
You did a very good job with the tesla switch!

Jesus