48V charging via ST generator and solar charge controller (no ST mod)

[mbryner]

Some of you may be following our discussion on this forum about mod'ing an ST head for 48 volt battery charging.  ( http://www.microcogen.info/index.php?topic=460.60 )   The discussion is centered on separating the 4 coils of the ST into 60 volts each, and feeding into a charge controller.

Since I'd like to retain the ST head 240/120 volt output for emergency use and keep it as versatile as possible, I am looking into other options.  I'll summarize below:

My equipment:
    ** JKson 6/1 Listeroid, propane canister muffler, no heat recovery yet.   I have a 120 volt heat rope a la Jens for heating WMO on the injector line.
    ** Powersolutions 7.5 kw ST head w/ doghouse already removed (price same as 5 kw from Joel Koch at purchase time, so I upsized)
    ** Outback 3600 watt inverters x 2, in stacked config for 240 VAC
    ** 48 Volt 5600 AH battery bank, 16 x L16 type batteries (actually currently 8 of them, because the house isn't built yet, but it will be upsized to 16 in a few months)
    ** Outback MX80 charge controller with input from 18 x 190 watt panels (3000+ watts solar total)
    ** Grundfos SQ series 240 V well pump located 320 ft deep and 800 feet from powerhouse and homesite (reason for 240 V).

With a listeroid 6/1 as the limiting factor, sizing for 2400-3000 watt charging input is appropriate.   In the current situation, when the generator is running, the system charges via the charging circuits on the Outback inverters (basically running in reverse) and the power output to the house site, powerhouse, and everywhere else is just dirty listeroid power.   We're all familiar with a listeroid's lax governor.  If I start a saw or air compressor, the frequency drops and the fluorescent tube lights flicker terribly. 

*** Here's the proposal: ***
1.  Take the 240 volt output from the ST head, transform it down to a reasonable level.  This step is skipped if you don't need 240 somewhere in your system.
2.  Rectify with appropriate sized bridge rectifier.
3.  Filter out ripple with inductive filter (choke) and large capacitor
4.  Feed into the solar charge controller.

Many people on this forum already have a solar charge controller, and the limiting factor in this case is the rated open circuit voltage (Voc) of your charge controller.   If you can breaker and keep the inputs from the generator (120 V rectified DC) and DC in from the solar arrays separate (diode protected from backfeed), you can use the same charge controller.   When it's dark, charge using generator.  When the sun is shining, charge via solar.   It could even have an automatic transfer switch, but not necessary considering the manual start of the 'roid anyway and diode protection.   And there could still be regular 120 V AC plugs next to the 'roid for emergency use.   I like.... 

BruceM and Mobile_bob have given great ideas.
Toroidal transformers would be the most efficient at >90%.
Here's BruceM's schematic for 120 V DC into MX80 charge controller with overvoltage protection:

http://www.microcogen.info/index.php?action=dlattach;topic=460.0;attach=1032

Midnite Solar is going to come out with a new "Classic" charge controller that can handle high voltage input, but it will cost ~$700.   The transformers above are more cost effective.  The Morningstar Tristar-60 is a good price, $200, and can handle up to 150 V DC input and 60 Amps output @ 48 volts.

[mike90045]

>  ** Outback 3600 watt inverters x 2, in stacked config for 240 VAC

Dang.  If you didn't have those on hand, the Xantrex XW6048 could take your genset 240V and charge with that.

[mbryner]

Oh, the Outback inverters do take the 240 genset voltage, one hot lead to each inverter, and charge with it.   But you still get the 'roid power to everything when charging instead of clean inverter power.   Therein lies the problem.    Don't get me wrong:  I have 200 hours on the 'roid so far and I'm very happy with it, except no electric start (my wife can't start it).    But it would be nice to eliminate the 'roid's fluctuating AC power except for unusual use and have it supply constant amps for charging instead.

[BruceM]

Thanks for starting the new thread Marcus, and you did a great job on the summary.

Pity your existing AC charge capability insists on routing AC through.  I'd contact Outback and see if there is a possible mod to change this.  An AVR and governor fix might also be in order. The AVR solves most of the problem, it will keep voltage up even though frequency varies.

Re: starters-  I wonder if there are enough guys interested that Dave (XYZ'er) might consider making a batch of 12V friction starter kits?  His design looks like winner, and except for outside sets (rain, frost) friction starters are simple and reliable.  

One error in your summary, which was mine, caught by Mike: The PWM Tristar-60 by Morningstar is the $200 unit, it can only handle 125VDC max voltage input.  Their MPPT units are also called Tristar-60, but these are over $500.  They can handle 150VDC input.  

The Outback MX-60 and MX-80 are rated for 145VDC max input, for operational use.

There will be more toroidal transformers up on ebay, we can find something else for your filter choke, but I'd snatch the other two right away if you're going this route.    If we find something else suitable for the step down transformer, then we can use them for chokes, still, that gives us some flexibility.  

[mobile_bob]

i agree with Bruce, you better grab those toroids while you can,
to buy new ones is pretty cost prohibitive in my opinion, last ones i checked on
were priced as if they were made of unobtanium by tesla himself and stolen from a museum somewhere.

ok, maybe not that bad, but pretty high cost for good ones

bob g

[mbryner]

OK, I bought all 3 of the 1200 watt $50 ones.   Having a duplicate for a spare seems prudent.   Let me know if you recommend another purchase!   thanks

Marcus

[BruceM]

Ok, Marcus.  Good call on the spare, I'd have done the same.

You can order toroidal transformers from several US suppliers, custom to your spec or from standard designs, but they are a bit pricey.  Antec is a good supplier for smaller ones, his test specifications have always been accurate.  He sells on ebay.  I have at times just bought bargains for the cores, which I then rewind the stripped cores by hand for chokes, or rewind secondaries to get what I need as a transformer.

Ebay item 360217315683  (x 3), with all six of the dual secondaries in parallel, would do the job for the filter choke. This one is $39 plus $10 shipping.  Good deal for a 1000 watt core.

[Jedon]

the Outback inverters do take the 240 genset voltage, one hot lead to each inverter, and charge with it.   But you still get the 'roid power to everything when charging instead of clean inverter power.

That's the problem I have with the Xantrex SW+5548's, not only do they pass the power through, they don't charge evenly so sometimes one inverter will draw 20A from one side and none from the other which even with the SR2 causes voltage issues and flickering.

[BruceM]

Jedon, that sounds a lot like a hookup (neutral)  problem between the two units, rather than a Xantrex design problem.  I'd  check the hookup compared to the manufacturer's diagram for this configuration very carefully.

[Jedon]

I studied the hookup instructions for quite a while and think I have it set up right, it doesn't always draw unevenly, only after it finishes the first bulk charging state, so when I first turn it on the SR2 gives 25A evenly to each inverter, then when the inverters don't need the whole 50A anymore the primary inverter takes over the charging. The hookup is pretty simple, the inverters are grounded, the neutrals are bridged and one hot to each inverter.

[mobile_bob]

one inverter charger after the first bulk charge is slow to sample the battery voltage,
this allows the faster to pick up the voltage because it is charging, and the second being slower
to sample see's the batteries at the higher voltage and thinks they are charged, so it drops back to
float.

thats my theory  

there maybe a setup to where you can program a specific timed bulk, timed absorption and then drop to float
instead of allowing the units to sample and decide for themselves what the batteries needs are.

having never studied the inverter/charger theory of operation, because i don't plan on incorporating this type
of dual purpose component, i can only imagine this is the problem, and it is likely there are ways to correct for this
issue.

you might want to double check the manuals and see if they need to be programmed differently when connected in tandem
than they are programmed for singular use.

there might also be a programmable delay from sample to begin charging, if so you might want to extend the time to mid range
so that both units have time to sample and then wait to initiate startup. that way it wouldn't make any difference which could sample
faster, because it would have to wait several seconds to initiate startup, giving the other added time to catch up.

i might be wrong, but i bet there is something along these lines that is at the bottom of the problem
bob g

[Jedon]

One thing I have noticed is I'm pretty sure one of the units has a busted battery temp sensor, perhaps then one is sensing the battery voltage "actual" ( higher voltage right now since it's cold ) and one is sensing compensated voltage ( lower right now ).

Link to diagram on how to hook two up:
http://www.xantrex.com/web/id/1075/docserve.aspx

Even if this is solved it will still flicker with the listeroid of course.

[BruceM]

I think you've solved your own problem, Jedon.  One unit thinks the battery is done first because of the broken temperature probe.

I looked at the diagram, and it does look like there's no way to have an AC battery charging source without it also feeding the AC.  That screws the  Xantrex inverter/charger combo benefits for non-3600 rpm gensets.  Their designers must not leave the lab much.     

[mbryner]

It's actually heartening to hear someone else have the same problem, jedon.   My setup is wired similar to yours.   Here is Outback's wiring diagram which is precisely how my setup is wired:
http://www.outbackpower.com/pdf/wiring_diagrams/4-7-2kw-120-240vac.pdf

And mine also has problems charging to completion:  this afternoon, when I got to my land, the TriMetric read 80% battery charge.   Combo raining & snowing.  Turned on the 6/1, started charging through the Outback inverters and ran a halogen shop light to have some lumens while making a counter.   Must have been pulling 3000+ watts:  2000 W charging and 1200 W halogen light.   Frequency initially dropping to 53 Hz.   When we left this evening after running about 3 hours, I check the TriMetric again and it read 85% charged, no current going to the batteries, and voltage of 52+ V.   (Should have higher charged voltage.)

Marcus

[BruceM]

That's gotta be pretty annoying when you have paid for "first class" gear like Outback or Xantrex, and then find that very basic operational stuff (like charging and operating with non-3600 rpm generators) doesn't work right! 

I just assumed you could just unbox and hook up these higher end systems and all would work well.  Apparently, NOT.

Marcus's "backdoor AC-DC charge via solar charge controller" fix is going to be popular, I think; it fixes the non-3600 rpm generator problem, and also the crappy charger issue.

I guess it makes me feel a little bit better about not being able to tolerate being around the commercial gear (EMI and my shot to hell brain) but it sure is a PITA to have to design and build everything from scratch.