ST power useless, Please assist me in improving the output!

[rmchambers]

Chris,
   I've gone the inverter + battery method in preparation for having supplemental power from a generator (very near future) and I have to admit that the computers and other network equipment don't mind the modified sine wave power from the Trace DR inverter.  You don't need to drop all that cash on a pure sinewave inverter unless you've got something that just can't stomach modified sine wave.

At the moment I have 4 trojan T105's and am going to get another 4 trojan T125's probably next week.  I'm in the process of building my whole backup system so it's in a state of flux at the moment.

My thought is that when the power goes out, the inverter takes over running the critical stuff until I can get outside and fire up the generator, let it warm up, then go back inside and throw the transfer switch.  The house runs on the generator and the inverter can start recharging the batteries.

Something to think about is the Trace takes the tops of the sine waves to charge the batteries, if the ST isn't putting out the high points of the sine waves then the batteries won't recharge properly.

The setup below ran my computer and networking gear, wireless and 2 IP phones as well as the 20 gallon fishtank pump for 12 hours the last time the power was out for an extended period.  Depending on your load you don't need much, but the more batteries you have the more you can store of course.

Robert

http://i164.photobucket.com/albums/u14/rmchambers/TraceDR3624.jpg
http://i164.photobucket.com/albums/u14/rmchambers/Fuseswitch.jpg
http://i164.photobucket.com/albums/u14/rmchambers/Batterybank.jpg

[Doug]

Nice to see the plexy cover on those fuses.....

I have a little discomfort witht he idea of mounting any electrical device on wood because of fire....

It used to be a code requirment to put a sheet of dry wall between a house sevice panel and combustable wall.

I like a bonded steel plate with dry wall or since the new regualtions came into effect saying you don't need the dry wall any more a 1 inch aiur gap at least.


Not critisising your job Rob its very neetand cleanly wired. I just have a lot of fears mostly based on bad practices of the past.

I like your battery back up inverter set up. I'm sold on it and want one now too.

Doug

[ronmar]

The capacitor I used is 1500 uFarad, 50V electrolitic.  1500uF is a little small for the nearly 2 amps of current draw.  If I recall my theory, at least 2000uF(or more) per amp of draw should be used for adequate filtering.   It is physically about the size of a "D" cell battery.  It was something I had setting in a junk box of old salvaged parts.  I will have to look around to see if I have any more as 3 in parallel would be more proper for 2 amps.  It is large enough to clean up the humps on the upslope/downslope a bit though which is why I was experimenting with it.

http://listerengine.com/smf/index.php?topic=634.30

The harmonic winding output resembles nothing like a sine wave.  It is a series of faster rise time spikes, 2 positive and 2 negative.  I can't output a pic from the digital scope I am using, but the above link is to another post on this forum that I know you have seen as you posted the next comment after the o-scope picture(next to last post on page 3).  The oscope pic courtesy of Bill Rogers, is a little hard to see, but is of 2 waveforms superimposed over each other.  One is the AC sine wave output from the stator windings.  The horizontal line across the middle has spikes that occur before and after the peaks of the sine. That spiked waveform is the "Z" winding output.  Because the peaks of the output of the "Z" winding are so far apart,  the ripple when rectified is terrible.  You need a big cap to take the edge off of it.  That thread discusses the harmonic winding causing distortions on the upslope/downslope and I am sure it does.  In my opinion, I don't think those distortions are really a problem for any load you might power.  It is the poor shape(flat top and multiple peaks) at the peaks of the sine that will cause the most problems with running more sophisticated electrical loads.

I am almost done working over my fuel system.  I will try and measure the ripple I am seeing with my single cap when I get it running tonight.

[ronmar]

OK, I found the file utility that alows me to download pics from my O-scope.  I started up a Coppermine account and have just uploaded 8 waveform pics in no particular order.  I saved all these without graticule display lines so the waveform details would be easier to see. I listed relevent voltages and information in the descriptions.  Here is the link.

http://listerengine.com/coppermine/index.php?cat=10113

Doug, I don't have 50HZ Stator taps on my ST-5, only 60HZ.

Chris I measured about 5V PK-PK of ripple with that 1500uF cap in circuit.  The peaks of the unfiltered rectifyer output go to nearly 70V.  There are pics of both in the gallery.  I will try a 4500uF on Monday and post some more pics. 

Ron

[rmchambers]

Ron,
   I was expecting a lot worse than I saw from your O-scope traces.  The DC ought to be able to get cleaned up with a better bridge rectifier setup, I know there are better ones out there.  Before he retired my dad made them, huge honking ones for use in MRI machines which would have not even broke a sweat with the load an ST head would put on it.  Too bad he retired before I started on my power generation thing.

There is still a slight blip on the up and down swings even when excited from a steady DC source so there's something else at play here that most likely Doug can pinpoint.

It looked clean enough that most electrical devices wouldn't have an issue with it though.  Might be some noise coming through on a stereo or something but when the lights are off, having hum free stereo is the least of your problems.

Thanks for sharing, you know what they say, a picture is worth a thousand words.  I'll be interested to see what happens when you increase the capacitance, see if that sands off the edges a bit.

Robert

[ronmar]

A better(larger) bridge might help as I think those stock diodes are right at their limit for current capacity.  I have a couple of larger case mount diodes(solder terminal on one end, threaded stud on the other:) that I was thiking of trying.  A larger diode might not ring as bad when the P/N junction shuts off.  The cap raises the average output to the rotor so if I limit the current being drawn out of the "Z" winding to compensate, this should also reduce the load/flow thru the diodes.

My field is electronics and I have made my own power at many places around the world.  Unfortunatly that generation experience was as an operator and not as an engineer.  Internal generator design,  construction and winding is beyond my current knowledge.  That may be why I am really enjoying this project.  It is dredging up theory I havn't applied in a long time and pushing me into areas that I have never dealt with.  I also expected the peaks to be cleaner under pure DC which led me to internal design/construction.  I also wonder if that "Z" winding resonates with no load on it's leads.

I agree that most loads should be fine with this waveform, with the microwave being perhaps the biggest annoyance.  It would however be nice if I could continue to run a UPS or two with the generator as I do now on commercial power.  I was contemplating a generator control circuit for over/under volt and freq using an old UPS board as the sensor. That is probably what got me even looking into the waveform in the first place. This coupled with a few well placed temp, pressure and level switches and a solenoid to trip a spring loaded fuel rack lever would get me a automated protection package.  I had looked at commercial sensors but I am not sure how a production volt/freq relay would react to this signal and that would be an expensive experiment. 

Ron

[Doug]

I'm looking at those filtered wave forms and Ihave to say WOW thats not as bad as I expected....

With respect to the distortion on the leading and trailing edges of the peaks, this might be caused by differences in the magnetic field desity of the iron and air gap as the poles move over the rotor. In some places soem combination of poles face and coil group may be closer togother than others there for the combinations induce a slightly higher ( faster rise ) and slight lower ( faster drop ) of the wave form.

You also have to consider this is a concentric coil winding, that was chosen by the manufacturer for its speed in spinning coils and assembly, not over all efficiency and performance.

There are also going to be varriations in air gap magnetic field density caused by the slot and tooth ( differences in the permiability of air and Iron )  as the pole crosses the coil group in the stator ( this can creat sort of a step patern up one sode and down the other of wave form , but I don't realy see this here. In one machine I scoped back in college I was able to count the slot per pole of the machine ).

At this point I wonder of simply a line reactor at the head or maybe some capacitor filtering at the points of use for specific equipment that finds this power objectionable might be enough.

I'm kind of suprised this head doesn't have a 60 - 50 hz tap selection. I thought the all did......

Doug

[ronmar]

OK, with 5200uF of capacitance on the rectifyer output I am seeing about 300MV of ripple.  The output waveform from the stator windings looks almost exactly like that from pure DC.  It still has the multiple peaks and slight steps on the ascending and descending portions of the wave. just like it did without using the "Z" winding for excitation.  I believe, at least on my gen that as Doug suggested,  this is a function of how it's core is formed and how it is wound.  I havn't tried a UPS on this filtered output yet, but because the waveform still has those characteristics, I don't think it will work.

Ron

[Doug]

Try running it with an inductive load like a water pump. The motor ( would perfer to see a cap run motor ) will eat some of the harmonics and help a clean things up.

Doug

[DaveW]

  Just a thought in looking at those waveforms.  A few years back I had an Allmand 5K generator developing those "squiggles" on the sides of the waveform, and a humming when under near full load.  I thought maybe the bearings were weak so I tore it down.  I noticed the lacing was cracked on one end of the winding.  I double laced the windings with unwaxed flax, varnished the windings, and baked it at low heat for about 20 hours.  That cleaned up the waveform a lot.  Seems the windings were vibrating under load.

[Doug]

I don't want to sound too negative but its a big job rewinding things like this.

Doug

[ronmar]

I don't want to sound too negative but its a big job rewinding things like this.

Doug

Yes it is, and well beyond my current skill level.  But revarnishing or flooding with epoxy, especially while the thing is relatively new/clean sounds like a good idea.  The windings on my ST-5 seemed fairly tight though so I am not sure how much this would effect/improve the output waveform.  I can have it down to bare frame/stator in about 30 minutes but I would have to make up an oven for it. 

I have a 1/2 HP motor on the drill press I can plug in tonight to see how that effects the output, but I obviously wouldn't want to run it all the time just to clean up the power:)

Ron

[ronmar]

Well I have had some success this evening:)  I got a Tripp-Lite 700VA UPS to accept the ST-5 power and startup/go green.  With the 5200uF of capacitance on the rectifyer output and a pair of 1.3 Ohm auto ballast resistors in parallel on one leg from the "Z" winding to the diode rectifyer to get my voltage back into the 120's, I connected the UPS to the generator under no load.  I had 60.4HZ and 126VAC on the meters and the UPS would not startup/accept the power as good.  For giggles and grins, I switched on the 1.5KW heater I have been using for a loadbank.  The output dropped to 59HZ and 120VAC and the UPS immediatly accepted the power and went to green.  I ran it for about 30 minutes trying different load combinations.  It continued to accept the ST power as good AC regardless of switching on or off the heater, drill press, welder and hand grinder.  The UPS stayed up till I lifted the shutdown lever and the voltage dropped to around 105 ish where it switched into backup mode.

Doug, I tried the 1/2HP motor on my drill press with no appreciable change in waveform.  I got to thinking about it though and I think it only uses a cap for startup.  But I think I found something better.  I started looking at the input filter circuits on switching computer power supplies.  I had a CPS laying in the junk bin so I plugged it in while feeding the other load and UPS.  It does a pretty good job stripping off the multiple peaks.  I posted 2 new pics at the following link this evening.  One shows the output with 1.5KW load with the UPS happy.  The other shows the same output/load with the computer power supply switched on.  With the computer PS in circuit, the sine has a much more definite peak.  The UPS ran happily  throughout all this experimenting. 

http://listerengine.com/coppermine/displayimage.php?album=152&pos=1

Ron

[fattywagonman]

Hi Ron,
I've been playing with improving the sine on the inverter / motor drives I'm wanting to use to power my shop...  Since the wave is created with PWM the sine wave is really spikey .... Today I installed a couple toroidal coils on the output and this immediately cleaned up the sine... It almost looks like a grid quality sine... 

[Doug]

Funny I was thinking about sugesting  a reactor/cap combination to act as a band pass filter.....
Or just a reactor.....

Your on your game....