synthesizing 24/36/48V inverters out of 12V inverters

[jack11]

Say I have two identical off-the-shelf 12Vdc-to-120Vac inverters, and want to synthesize one 24Vdc-to-ac inverter out of them.
“Identical” is the key word here, these two inverters must have the same topology, component tolerances, and internal timing  delays.
The benefits of doing this are obvious, if you already have 12v inverters on hand and want to build 24/36/48v systems without using dc-dc converters.
Not knowing much about inverter design, I am sure I've gone badly wrong with my reasoning below, but where, can someone point out?

I connect the two 12v inverters in series on the dc side, and tie them to a 24v dc source.
I connect the inverters' ac outputs in parallel (L1 to L1 and L2 to L2).
Perhaps some external isolation/blocking circuit would also be needed at the inverters' ac outputs, or not because this may interfere with the ac feedback, see below?
I tie the inverters' on/off switches together to ensure identical turn-on transient timing.

On the dc input side, the mosfet collector currents should be the same in both inverters, as these inverters see the same ac load, and operate with the same pwm controls and timing?
However, each inverter's H-bridge would see 12vdc across it, and not 24vdc?

On the ac output side, the voltage and current phasors from both inverters should add in-phase.
The two inverters see a common ac load, so the feedback from common L1 and L2 to the pwm controls in both inverters should be the same, ensuring mosfet control synchronization in both inverters.
This cross-feedback may also cause any phase errors between ac outputs from both inverters to be driven to 0, ensuring ac outputs synchronization (because each inverter sees the other inverter's ac output, and its own).

If this scheme flops, are there any other schemes possible to accomplish this goal?

[dnix71]

Don't bother. It's cheaper and safer to just buy a real 24v inverter. Trucks in the US have 24v systems, so 24v inverters are common and cheap.

[jack11]

have you checked the prices lately, the 24v inverters of decent quality in 2-3kW range are running about $1000/apiece, on ebay/amazon/etc. I found some that are a little cheaper, but they have limitations I can't live with, or the manufacturer's reputation is crappy.
The 12V inverters are much cheaper in general, and easier to get (plus I happen to have a whole case of them).

and what about the 36/48V inverters, even more expensive and harder to get, and not commonly used in trucks. I hear some new controllers will take pv array voltages up to 600Vdc, and will be able to charge battery packs of 100Vdc or so, these systems will need inverters too.

I am simply wondering if a common 12v inverter may be used as a building block to make whatever else dc-input inverter you'd want (in multiples of 12 volts).

[boB]

My first statement comes as an afterthought to what I posted below...  If you want to go to
all this trouble to series inverters, why not just get a large 12V inverter and use more 12V
batteries in parallel OR talk about paralleling multiple 12V inverters on their secondary and
DC primary ? Assuming you could synchronize and parallel them.  That way you would at least have
a common negative.

You might be able to take multiple 12V inverters and connect them across 12V taps of a 24V,36V, or 48V battery bank,
but unless they can be synchronized, you won't get them to do what you want here.

Even if the can be synchronized, (cheap inverters do not do that), they will normally need to have
a common negative for that synchronizing signal unless you do some reverse engineering
and make an isolated synch comm gadget.

Just connecting 12V inverters of the same kind together will not guarantee that they share voltage
drops but placing them across 12V sections will help.

Having any non-balanced discharge of each 12V section is not good for battery longevity either so
you would have to make sure they share the load well, OR, make sure those 12V sections
get charged good.

Inverters do not automatically synchronize their AC outputs just because they are connected
together.

My conclusion is that this is not a good idea.  Especially with El-Cheap-O inverters. Cheap
inverters break too easy (usually).

If you have a lot of extra time and money, try it.  Let us know how it goes.

boB

[OperaHouse]

First off the outputs would have to be electrically isolated.  Thise that are are already expensive inverters.  There are inverters out there trat you can sync but there control lines are not floating.  If there was a market for such a thibg it would exist. I do lots of fun things with inverters, but this is something I would never consider

[dnix71]

http://www.ebay.com/itm/Samlex-PST-200S-24A-2000-Watt-Pure-Sine-Wave-Inverter-24V-/390552968466?pt=US_Power_Inverters_&hash=item5aeec55d12

Pure sine inverter, 24v, 2kw $750 delivered

$900 for a Wagan Elite pure sine inverter

http://www.ebay.com/itm/Wagan-Tech-2205-Elite-2000-Watt-Pure-Sine-Wave-Power-Inverter-/390401633759?pt=LH_DefaultDomain_0&hash=item5ae5c02ddf#shId

Aims (almost new, with transfer switch) for $875 shipped

http://www.ebay.com/itm/Aims-24V-3000W-Pure-Digital-Sine-Wave-Inverter-and-3-way-crossover-switch-/221212928216?pt=US_Power_Inverters_&hash=item33815168d8

http://www.ebay.com/itm/NEW-Go-Power-GP-SW1500-Watt-Pure-Sine-Wave-Inverter-24-V-Remote-Included-/360619775965?pt=LH_DefaultDomain_0&hash=item53f69d17dd GoPower sine inverter 1500/2000 watts $650 shipped.

All those prices seem reasonable considering they are sine output. You can get mod-sine inverters for less than 1/2 that, but I wouldn't want them. It's stupid to try and save a little money playing around with high voltage and currents. If you mess up what you damage could be a lot more expensive than what you thought you would save.

[Bruce S]

There is also the sync'ing of the 120Vac , being that the el-cheap-o 12Vdc inverters are most likely Mod-sine that may not be very easy at all.
On the 120Vac side you will need to be able to make sure the phasing is dead on, otherwise the Volts could slip between more than 120Vac and lower.
While it can be done, it may end up costing you more than purchasing one that is built for the voltages you're wanting to use from a battery bank.
In all of the inverters I've owned and been luck enough to look inside of, they all take the battery bank voltage step it up to 120Vdc then using magic smoke  step that to 120Vac .
Actually depending on MSW of PSW they will step it to either 144Vac and 170Vac 144Vac is the MSW side so they can account for losses. PSW uses the 170Vac for true RMS values.
You could in fact rewire a battery bank for 120Vdc, find the point where inverters have the DC volts stepped up and attach your 120Vdc string there.
NOTE: THIS CAN BE DANGEROUS! so don't do it unless you understand electronics and have safety equipment !

Perhaps putting some of your extra 12V inverters up for sale, then using the proceeds to purchase higher DC Voltage inverters would save time?

 

[XeonPony]

I am going for the less tactful methods the other have used and cut to the point : The idea is idiotic at best, for one simple reason, SAFETY unless you plan to spend thousands in failure mode testing and fail safe systems it isn't worth it in any way,  Burning to death really screws your week up!

The other posters covered all the technical reasons thoroughly, Penny wise pound foolish is the saying that comes to mind here, you save a quid and costs you a fortune as your insurance will not cover the fire damage (And that is if you're lucky it will only be fire).

The proper tool for the proper job.

I run a 12v system and I plan to switch to 24V, I never draw more then 60Amps on the 12v side (Usually less then 45A). the proper safe solution?  Get a 24v to 12V dc-dc converter make its rated out put is 60A cont this gives me a nice safety factor for surge loads on the 12v side.  Buy a proper 24V inverter with some stacking capability so I can expand on it's out put if I want.  Safe clean and reliable system with no issues to worry about fire or explosion.

[joestue]

sounds like something i would try to do lol.

not going to work with MSW inverters, can't sync them, if you add the outputs together with transformers... they may even run at different frequencies, even a tiny bit adds up and then your lightbulbs have a .01Hz ripple to them, works great for all of 5 minutes...
also, you would need a 12v battery for each inverter absorb the differences in current draw between the inverters.

if you want to screw around with these crack pot ideas... do what i PM'd you about. take the ~300 volt bus out of any msw inverter and feed it into a 3 HP VFD, set it for 240vac 60Hz and use an isolation transformer wired as an auto tx to get your 120/240.
you'll need a reactor and a motor run capacitor to get a nice sine wave, and the downside is you'll be eating a lot of no load losses in the VFD.
but it will only cost you the cheap MSW inverter plus ~100$ for the VFD, and you'll get 3 phases at that.

[jack11]

Thanks all for good feedback (except for personal bashing).
In fact, I plan to set up a test in a controlled environment and take some measurements, when I have time. I expect nothing, as the mosfet gate syncing problem will probably cause the dc-side circuit to be open, on average.
And remember, this is just a mental exercise, so there is no need get excited or resort to bashing. If you can simulate this thing in your mind then let us know your TECHNICAL thoughts only. I've seen stupider issues submitted for discussions.

I suspected the sync would be a problem. Slight deviations due to component tolerance or timing would throw this thing off, even though it might work for a while. Stackable inverters have solved this problem at the ac waveform, perhaps by going all the way back to the bridge control, but that increases the complexity and cost and defeats the original purpose of having simple generic building blocks.

Plus there may be a more fundamental problem, even if perfect sync could be made. I don't believe it would be possible for the first inverter to pass the dc current to the second one, except through the ac load, because of the way the H-bridge cross-fires its mosfets.

Too bad it can't be done, but you have to admit it would be nice to have a generic inverter building block to use on whatever system you please.

[Bruce S]

The bashing comes more out of concern than trying to be personal about it.
The voltages involved can, will and do reach dangerous both inside and outside of inverters.
Even the DC caps and stay energized for a long time without proper de-energizing, and those can be lethal.
As a mental exercise it'll be a good one for sure.
Keep us up to date on how you're progressing.
Stay Safe;
 

[zvizdic]

This post  grabbed my attention.   
I am  on a project  of converting 12v input  xpower inverter to 48v.
Starting with 300W  xpower inverter if it works going bigar .
Switching frequencies 42kh ferrite core primary 4+4 8strands of #24 wire rewound to 16+16 2strands #24 wire replacing mosfet of higher voltage and input power caps .
I am going to add a choke filter on input no choke on 12v version .
Any input would bee appreciated.

[jack11]

I have not had the time to do anything with this yet, but I think it's possible to re-use some of the 12v inverter's sections.

Once I get around to messing with this, my line of attack is going to be to change the dc-dc converter to produce the same dc-bus voltage for the ac-generator (H-bridge) out of a different dc-input voltage. This way the ac-generator should not see any difference from the original configuration, at least in theory, unless the dc-dc converter and the ac-generator are coupled in some way???