Which AVR, and is it a good practice to eliminate the harmonic winding?

[dieselfuelonly]

My ST generator has the old rheostat on it and while I'm removing the doghouse and redoing everything I figured I might upgrade it to an AVR.  I see the that Georgia Generator has a few different options, one that appears to use one of the two 120v legs to do the regulation, and the other that uses the voltage across both legs to do the regulation.  Any reason to go with one over the other?  I'm guessing that the 120v regulator could be used with the generator in both 120v or 240v configurations, whereas the 240v regulator would require the series wiring at 240v to function.  The advantage I see to the 120v regulator is that it can be used to bypass the harmonic winding for excitation, where it appears the 240v one can't.  Opinions, alternative regulators, etc?


Here are the wiring diagrams from GG:

240v regulator:
https://www.thegenstore.com/index_htm_files/AVR230-Silver-Bottom-1-or-3-phase.pdf

120v regulator using harmonic winding for excitation:
https://www.thegenstore.com/index_htm_files/AVR115-Black-wiring.pdf

120v regulator bypassing harmonic winding for excitation:
https://www.thegenstore.com/index_htm_files/AVR115-Black-wiring-to-remove-the-harmonic-excitation-winding.pdf

Thanks

[playdiesel]

All depends on how clean your power demands are and also the individual head.  I bought an ST15 and ST5 from Tom when he was selling a lot of them and quality was high.  Both of mine are well regulated running on the harmonic.  The 15 runs the entire house when the grid is down and we live same as normal, computers, TVs all run perfectly. Other ST users have had difficulty due to poor wave forms and regulation. Hopefully one of the electrically smart people will answer your ither questions.

[BruceM]

Hey DieselfuelOnly,
Good question, but some more info is needed, though I can make some general suggesti9ns.  You didn't mention the engine, rpm, or ST head size, or whether your setup was 120, 230, or split phase 120/240. Those do make a difference.

Waveform distortions vary dramatically between different ST heads, since manufacturers have varied over the years.  For many applications and specific gen heads that may not be important. Voltage regulation via harmonic with a rheostat or adjustable resistor can be acceptable for many applications too.  Harmonic regulation will in fact handle large, low power factor loads much better than the typical affordable AVR, which does NOT regulate to true RMS volrtage. 

For AVR configuration, many ST3 and  ST5 heads will groan and make vibration noises if imbalanced by running spilt phase loads, or putting an AVR on one leg of a split phase setup.  I suggest avoiding that.  In fact I suggest avoiding split phase setups for these heads entirely.  The 230/240V rated unit sold by CGG in the past will minimize flicker best if installed with the mains only as excitation.  That will eliminate the "harmonic hump" distortion that is pretty band on some ST3 and ST5 heads.  The reason to use the mains is this:  the harmonic wiinding puts out more volts when there is a load, and on my ST3 I have measured the peak voltage at over 400V for starting a 1.5 HP equivalent motor.  It is just 4 skinny spikes per rotation, however, so when compression stroke comes and rpm slows down, so does the harmonic output.  If you put an AVR  using the Harmonic winding as excitation, it can only limit the output of the harmonic.  So for ST heads with properly wound harmonic windings, you are only limiting and already limiited output, and the mains voltage will sag.  If you use the mains for excitation, especially with the 240V AVR, you have some serious head room for the AVR to work with and it will partially correct the voltage sag of the compression stroke.  Either approach will improve the voltage rise during the power stroke. 

For most utility uses like running pumps and motors and tools,  there's little need for precise voltage regulation or worry about Listerflicker.

I hope that helps explain things.  I've been inside 5 -ST-3's and designed my own AVR electronics. A long time ago I offered a flicker reducing AVR plan as open source, and have thus seen oscilliscope shots of a few ST-3's ST-5's and a couple ST-7's.  Some have acceptable waveform distortion, the worst have such bad ratcheting around the peaks that I wouldn't run them myself.  With lots of distortion, some mechanical timers will groan and get hot, and motors will get a bit hotter and be less efficient, power wise.  A good AVR set up will reduce flicker as much as SOM type flywheels, and I've also posted that comparison publicly.  AVR flicker reduction is reduced as the head gets larger, because of the lag of larger inductance in the larger excitation coils.

My neighbor is running a CGG ST-3 with a cheap China 230V AVR.   It works well, but he keeps a spare on hand as they typically fail every 2-3 years, though he does put on perhaps 600 hrs/year.  They are similar to the units CGG sells, he was out of 230V at the time.  We tried to use the 120V AVR CGG sent with it but it made the ST set up as 230VAC moan from the small load imbalance.   I have also not been impressed with the quality of newer ST heads-  harmonic windings are so wrong (very high voltage and no rheostat)  that they can't be used without an AVR, and we got a head with aluminum rotor windings that failed open in 6 months.

My ST-3 on a homebrew AVR has been very reliable.  If I've got to run something  with a nasty PF like a small MIG welder, I flip a switch to go back to harmonic.
Bruce






The huge variation explains why some think ST heads are fine, while others really suffer. 

[dieselfuelonly]

Hey DieselfuelOnly,
Good question, but some more info is needed, though I can make some general suggesti9ns.  You didn't mention the engine, rpm, or ST head size, or whether your setup was 120, 230, or split phase 120/240. Those do make a difference.

Waveform distortions vary dramatically between different ST heads, since manufacturers have varied over the years.  For many applications and specific gen heads that may not be important. Voltage regulation via harmonic with a rheostat or adjustable resistor can be acceptable for many applications too.  Harmonic regulation will in fact handle large, low power factor loads much better than the typical affordable AVR, which does NOT regulate to true RMS volrtage.  

For AVR configuration, many ST3 and  ST5 heads will groan and make vibration noises if imbalanced by running spilt phase loads, or putting an AVR on one leg of a split phase setup.  I suggest avoiding that.  In fact I suggest avoiding split phase setups for these heads entirely.  The 230/240V rated unit sold by CGG in the past will minimize flicker best if installed with the mains only as excitation.  That will eliminate the "harmonic hump" distortion that is pretty band on some ST3 and ST5 heads.  The reason to use the mains is this:  the harmonic wiinding puts out more volts when there is a load, and on my ST3 I have measured the peak voltage at over 400V for starting a 1.5 HP equivalent motor.  It is just 4 skinny spikes per rotation, however, so when compression stroke comes and rpm slows down, so does the harmonic output.  If you put an AVR  using the Harmonic winding as excitation, it can only limit the output of the harmonic.  So for ST heads with properly wound harmonic windings, you are only limiting and already limiited output, and the mains voltage will sag.  If you use the mains for excitation, especially with the 240V AVR, you have some serious head room for the AVR to work with and it will partially correct the voltage sag of the compression stroke.  Either approach will improve the voltage rise during the power stroke.  

For most utility uses like running pumps and motors and tools,  there's little need for precise voltage regulation or worry about Listerflicker.

I hope that helps explain things.  I've been inside 5 -ST-3's and designed my own AVR electronics. A long time ago I offered a flicker reducing AVR plan as open source, and have thus seen oscilliscope shots of a few ST-3's ST-5's and a couple ST-7's.  Some have acceptable waveform distortion, the worst have such bad ratcheting around the peaks that I wouldn't run them myself.  With lots of distortion, some mechanical timers will groan and get hot, and motors will get a bit hotter and be less efficient, power wise.  A good AVR set up will reduce flicker as much as SOM type flywheels, and I've also posted that comparison publicly.  AVR flicker reduction is reduced as the head gets larger, because of the lag of larger inductance in the larger excitation coils.

My neighbor is running a CGG ST-3 with a cheap China 230V AVR.   It works well, but he keeps a spare on hand as they typically fail every 2-3 years, though he does put on perhaps 600 hrs/year.  They are similar to the units CGG sells, he was out of 230V at the time.  We tried to use the 120V AVR CGG sent with it but it made the ST set up as 230VAC moan from the small load imbalance.   I have also not been impressed with the quality of newer ST heads-  harmonic windings are so wrong (very high voltage and no rheostat)  that they can't be used without an AVR, and we got a head with aluminum rotor windings that failed open in 6 months.

My ST-3 on a homebrew AVR has been very reliable.  If I've got to run something  with a nasty PF like a small MIG welder, I flip a switch to go back to harmonic.
Bruce






The huge variation explains why some think ST heads are fine, while others really suffer.  

Thanks for the info.  I have an older (manufactured in 98 or 99 I believe) ST-10 head.  It's extremely heavy compared to the new ones weighing around 400lbs+.  It'll be powered by a 2TNV70 Yanmar so I only expect around 5kw output, if even, out of it.  Direct drive @ 1800RPM.  I'd prefer to go split phase 120/240 as I do have some 240v loads.  Perhaps I should see how it performs with just the rheostat first.  If I start experiencing some bad groaning from unbalanced loads or issues with over/undervoltage on the legs I'll investigate further.  It seems the best setup is running parallel 120v and then using a center tap transformer to get 120/240v output, but that's cost I'd prefer not to take on if I don't need to.

[dieselfuelonly]

Can someone tell me what this component in the big red circle (believe it's an inductor?) is?  It had a heavier gauge black wire that ran from from I believe U4 to U2 and wrapped around the component itself twice.  As far as I remember it had no connections to itself other than the wire wrapping around it.  Is it supposed to be for the amp gauge?

[playdiesel]

I have seen those in electrical panels to operate the amp meters thus I think you have that right. Interesting rectfier, the OEM rectifiers are known weak spots. Mine both failed in a few hours and I replaced them with metal case bridge diodes that must be used in conjunction with. a heat sink. If you are using a metal junction box it works fir that just drill a hole and mount the diode to it.

These are couple bucks each on ebay thus cheap to have some spares

[playdiesel]

pic is pretty bad sorry, let's try it again. I think mine are 5010 . As I remember added ampacity doesn't hurt operation and will add life. Bruce will know for sure.

[mobile_bob]

the unit in the big circle looks to me like a CT (current transformer)

it should have only one wire going through it, however there might be two if they are off the different windings
it measures amps going through those wires, and feeds that to the amp meter.

i have no idea how accurate it is, but the system could be verified by using a clamp on meter, to see if they generally
agree with each other.

bob g

[BruceM]

Thanks for filling us in on your engine and genhead details.  That's a whole other story; it's nice that you can ignore Listerflicker.  Using the harmonic for excitation should be fine in this case.  Just remember you can always use the mains for excitation later if the harmonic should fail.  As for harmonic hump distortion I have no experience with the St-10.  If you have or can borrow even the cheapest of ocilloscopes to get us a picture of your output waveform, that will give us a good idea of what should be done.  There are some very cheap, handheld digital o'scopes out there perfectly adequate for much troubleshooting/checkout work.  Single trace, even 1 MHz bandwidth will do the job; I use one for most of my own design and troubleshooting, except where bench testing requires my PICOscope.  My dual trace analog/CRT scope just gathers dust.

I concur fully with your decision to experiment to see if this ST-10 will handle split phase, to avoid transformer cost and losses if possible.  You might also see if the stock harmonic with rheostat is adequate for your range of loads on this setup.  It may very well be, and if so, I'd stick with that over the known to be less than durable cheap AVR's from China or CGG.  If you need the AVR for your loads,  get a spare if buying the China cheap one (same as CGG).  About 3 years of regular daily use is all you should expect.  On Ebay or elsewhere, you're looking for ST-10 brush generator head compatibility.  

One other issue is the regulation of the two legs of 120V when using an AVR.  If your head can handle split phase, mismatched 120V loads, just be aware that regulation per leg may get a bit sloppy.  I think the 230V regulator is best, which works similarly to the Harmonic regulation; the average of the two 120V legs will be right, at least.  If using a 120V AVR that is regulating based on one lightly loaded leg, and you put a very large load on the other, you could have a motor burnout low voltage situation on one leg. With ST heads and India's Listeroids/Clones, it's best to assume nothing and test/inspect everything  Both are very useful, affordable gear for very handy skeptics.

+1 on both current transformer and bridge diode replacement with metal bodied, modern one.  Some thermal grease or silicone caulk in a pinch on the connection to some metal sheet surface is advised.  I like to bolt through sheet, into a finned aluminum heatsink where practical, both with a thin coat of thermal grease.

[mobile_bob]

back when i was doing a lot of testing, i set up a test cell to measure every parameter i could think of
and had a load bank to measure in 500 watt intervals up to 8kwatts (i was using an st7.5)

about 90% of the time, when the loads were not balanced, that "sonofabeeatch" would scream like an air raid siren
and i had to use both in ear and outer ear protection, it was horrible.

it wasn't until i did a little research and found out that these heads are wound for "primarily used for" 230-240volt loads which balances the loads across the stator.

i had assumed wrongly that the stator was split and each set of windings for 120volts were 180 degree's apart on the stator, which they are not. they are 90 degree's apart and as such when  you load one 120volt leg and leave the other either unloaded or partially loaded, the rotor is pulled off center, and really doesn't help matters at all.

i don't know, but suspect had i gone in and split the poles and reconnected them so that each set was 180 degree's opposed, things might have been much quieter?

all i know for sure is, there is no way on earth a neighbor would want to live anywhere within a half mile of this bugger under those conditions.

the odd thing was, about 10% of the time, it would run dead quiet?

if however it was due to loose stator laminations, it likely would end in premature stator failure, due to wear on the winding insulation.

always figured that before this unit could be placed into service, i would be faced with a complete teardown of the s195 and the st7.5, to clean, inspect, and correct faults as found.

i am fairly convinced the designs are good, and the manufacturing is also pretty good, however some of the assembly might be questionable.  certainly would want to check them both out if i were to want to depend on them solely for survival.

bob g

[BruceM]

I know the Stamford heads have zero problems with split phase, imbalanced load operation, need no work on bearings, diode bridge or voltage regulation.  Gary at DES recommended them many years ago and all reports have been the same; no problems at alll.  Alas, they don't have a small one and the price was over my budget.

[glort]

about 90% of the time, when the loads were not balanced, that "sonofabeeatch" would scream like an air raid siren
and i had to use both in ear and outer ear protection, it was horrible.

I don't know what the differences are and I know the US System is a lot different to the native 240V Systems but I have played a lot with Imbalanced IMAGS and found no problem at all.
I have read a lot about heat build up but in an IMAG it makes no difference other than you can only get a lower output.  I put this down to the fact the machine is working a lot less and has in fact extra heat sinking capacity and cooling.  Never noticed any untoward noise but Imags tend to sing a little through the caps anyway but it would be nothing that would disturb anyone at any distance and certainly  wouldn't register in compared to the engine noise.

I haven't played with any 3 Phase gen heads like this and have often wondered if all the cautions about making sure the loads are balanced are just more over pedantic best practice hype or if it can really make a difference on small output units. I'm sure unbalancing a 100Kw+ unit may cause other problems but for something 5Kw or so. may not be that bad.... Maybe.
Could also depend on wether it was a 110. 220V Unit or a 415 Type where each leg is 220 and the configuration is different.

I notice a lot of 3 phase generators are sold here that have 220V taken off each leg and while they are supposed to be balanced, I would guarantee 99% of the time they are not and only are running off a single leg when used in single phase.  Maybe you can't pull 100% single phase power off that leg all day long but I'd be pretty sure you could pull something decent or these things would be letting the magic smoke out on building sites everywhere.

There are Phase controllers that can create 3 phase from single ( in our power system anyway) and can combine 3 phase to Higher current single. 
The latter would be ideal for load balancing 3 phase heads. 

[playdiesel]

Both of my ST heads are set up 110/220 so they may be connected to the main if need be. The ST15 is actually hard wired. From I have read I must have lucked out and got good ones. Neither of mine suffer from the excessive noise although the 15 is belted to a 1115 and it would have to make a LOT of racket to be heard over that engine. The ST5 is run by a 6/1 and if there is no load on one set of windings and a heavy load on the other it will groan slightly.  
I suppose this is known to all but noise aside my reason to balance loads when wired 110/220 is the ampacity at 110 is doubled with a totally balanced load vs using only one side.  100% balance isn't achievable in a home setting but the closer to balanced tbe loads are the. better, IMO

I would have had Stamford heads if they were reasonably priced when I needed them. Most are Chinese now, not so when I purchased mine and a Stamford was about 10X whatvthe ST cost me.

[BruceM]

The variable issue with the smaller ST heads and imbalance vibration is clearly related to the utter lack of tolerances and variation of rotors, stators, and windings used, though I have yet to see any ST3 can can be operated split phase.  As I've reported before, with 3 ST-3s in pieces on my bench, no two were alike in stator length.  Centering of the rotor is highly variable and that's rather critical.  In short, they are made much like Listeroids/Clones...buyer beware.  It will be interesting to see if this particular ST-10 can be operated split single phase with moderate load imbalance.  I hope so!

Otherwise, the obvious solution is to use a transformer to generate the 230 or 120V voltage you need the least of.  Antekinc.com is my favorite toroidal transformer supplier; they are typically about 93% efficient and can handle big surge loads (motor start) without a problem.  They are conservatively rated so you needn't allow lots of head room.  Though not ideal, I run 3-800W step down transformers in parallel for my shop 120VAC.   At the time I just couldn't find an affordable 2400W toroid, and I wanted to avoid the inefficiency and large stray magnetic field of a conventional transformer. No problems at 12 years, couple thousand hours of generator/inverter use of those transformers.  I use 2- Antek 1000 W transformers in my 5 step sine inverter and found they are using very efficiency, high inrush current cores; when I switched to them, I had to add a soft start software as I kept blowing fuses and transistors and couldn't figure out why.  I also had to add an inrush current limiting power resistor with timeout bypass relay to to the shop step down transformers for the inverter's sake.  The inverter easily handles my well pump start surge of over 4000W, yet would blow fuses trying to start that 2400W step down transformer, so be aware of the potentially huge inrush currents of high efficiency core toroidal transformers.  It sure surprised me and had me scratching my head until I measured the peak current and then did some technical reading. 

The ST-3 never had any issues starting the 2400W step down transformers, so I had been blissfully unaware of this issue.  Some toroidal transformers use core material designed to minimize this problem, and in my inverter development I started with a 120V output, and had unknowingly used some surplus 1100W transformers that were of that type.  When I first switched to new Antek transformers, for which I had rewound the secondaries, I had an "educational experience" with loud bang and smoke coming out of my electronics! By starting them on alternating polarity, very short pulses, gradually increasing in width, they could be essentially degaussed from their last operation and started up with no problem.  (Most electronic transformer soft starters use this method.)