diy solar

diy solar

Auto Transformer based all-in-ones (how do they handle ground?)

I’ve tested this rewired under a 5000W load on one leg and it handled it like a champ:
521A3D34-45C0-4361-8A53-2DFA8A1C7C7C.jpeg
Inside the “ST-5000” power adapter is a toroidal autotransformer with 0/110/200/220/240v taps, I wired the 0/110/220 through a 3-pole 30 amp breaker. I already have a high-frequency split-phase inverter, I added it to balance loads, but if it trips then the split-phase inverter faults if there is too much load on one leg anyways.

I’m doing zero-export to CT and it balances enough loads so that one side of the inverter doesn’t max out occasionally and I catch more of my usage.

$100 on eBay.
 
You are allowed to wire 240 volt sockets and have 240v appliances. It is just not normal to have 240V small appliances. I see nothing that would prevent you from putting a US 240V plug on a European tea kettle.
Careful with that; the switch and safeties on the European version would be single pole so a ground fault could cause a dangerous situation.
 
Also, if the European tea kettle case is metallic.
And is bonded to what would be the European neutral. (Now being connected to US L2) you could have 120v on the case. Sit it by the sink, and touch both at the same time. And you could be in for a dangerous shock.
 
I also think it's pretty stupid that I'm "not allowed" to have 240v appliances like kettles.
I recently learned a bit about the history around this.

When power grids were first being installed, 120V was about as high of voltage that could be used on a light bulb without it burning out too fast. 120V grids were being installed throughout the world.... but the US was way ahead of the rest of the world. As the tungsten bulb was perfected, 240V became acceptable for lightbulbs and people started advocating for 240V grids. In Europe, they weren't as far along with 120 volts so they were able to transition. In fact, some power companies paid customers to switch to 240V bulbs so they could change the grid. However, by the time this happened, there were too many 120V installations in the US so the US never made the change.
 
I recently learned a bit about the history around this.

When power grids were first being installed, 120V was about as high of voltage that could be used on a light bulb without it burning out too fast. 120V grids were being installed throughout the world.... but the US was way ahead of the rest of the world. As the tungsten bulb was perfected, 240V became acceptable for lightbulbs and people started advocating for 240V grids. In Europe, they weren't as far along with 120 volts so they were able to transition. In fact, some power companies paid customers to switch to 240V bulbs so they could change the grid. However, by the time this happened, there were too many 120V installations in the US so the US never made the change.
Yeah, it’s one of the many technological leapfrogs that has happened to technologies developed and rolled out in the U.S. first.
 
Off-grid, if auto-transformer is what defines neutral, you don't want auto-transformer to disconnect by breaker tripping without also disconnecting 240V power to panel. Otherwise, your hair dryer on one phase and small gadget on other phase divide voltage unevenly.

On grid, auto-transformer may or may not help with anything, and may or may not try to rebalance grid. If it has 2-pole breaker sized for its rating and neutral has 2x the ampacity it should be OK. My concern would be it running too hot but not tripping breaker, could damage insulation over time. If it tries too hard (due to very low resistance like that toroid I linked), it could trip breaker. If that is possible, I'd rather isolate auto-transformer while on grid.

I think it is OK if you have zero net export, with equal kW exported on one leg and imported on the other. Don't think the utility is paying attention.
A isolation transformer would eliminate that, so only true net export or import goes through the wires.

I've been browsing a bit. You can get 25kVA, 240/480 primary 120/240 secondary used for around $400 plus shipping.
I'm stalking your old posts to learn things. I had planned to pick one up for a while once I came to realize that "480x240-240/120" was a standard type of transformer a while back, but before I read this post I picked up a 25kVA up a couple of months ago for $300 plus a couple of hundred freight to my local hub https://www.ebay.com/itm/353971576700 where they put it in the back of my truck for me. I'm going to experiment with it and eventually use it with a few of these: http://www.tbbpower.com/index/allProducts/detail?bigClassProducts=Inverter&Charger&products=Kinergier Pro in a container to go in place of my grid connection with that transformer to make neutral. Keeping the Deye 8k hybrid inverter + new smallish battery + PV on my house, and cycle the new smallish batteries into the container once a year for much lower C rates and build up a large battery for long term storage.

I want to do some lab experiments with it wired as an autotransformer and as an isolation transformer.

72072C75-E8AD-442F-B1EE-5CAB717480EC_1_105_c.jpeg
I put it on castor wheels ;-P. Edit: caster
 
Last edited:
I plan to get one too, one of these days, for a related application.
Meanwhile, I have a smaller 9kVA toroid (with only one 240V winding.)

It'll be interesting to see how these things behave regarding inrush. That is kind of random, depends not only where in the waveform power is applied, but also which polarity compared to last half-cycle when it was turned off (and where in the waveform at turn-off too.
For the grid, it can trip breakers. For an inverter, might shut it down. Maybe different if present at power-up vs. switched in. I've wired one 9kVA through a breaker my single 5kW Sunny Boy, hasn't been a problem.

My clamp meter inrush feature didn't catch transformer inrush, and my CT attenuated it. A higher frequency response clamp probe does read it.

I used a 40A clamp probe at work, and when that clipped a colleague's i2000. I've since ordered one for myself.

https://www.fluke.com/en-us/product/accessories/current-clamps/fluke-i2000-flex

 
I'm stalking your old posts to learn things. I had planned to pick one up for a while once I came to realize that "480x240-240/120" was a standard type of transformer a while back, but before I read this post I picked up a 25kVA up a couple of months ago for $300 plus a couple of hundred freight to my local hub https://www.ebay.com/itm/353971576700 where they put it in the back of my truck for me. I'm going to experiment with it and eventually use it with a few of these: http://www.tbbpower.com/index/allProducts/detail?bigClassProducts=Inverter&Charger&products=Kinergier Pro in a container to go in place of my grid connection with that transformer to make neutral. Keeping the Deye 8k hybrid inverter + new smallish battery + PV on my house, and cycle the new smallish batteries into the container once a year for much lower C rates and build up a large battery for long term storage.

I want to do some lab experiments with it wired as an autotransformer and as an isolation transformer.

View attachment 93236
I put it on castor wheels ;-P. Edit: caster

How is that transformer behaving?
I'm now encountering issues using a similar but 3-phase transformer for step-up rather than its intended step-down.
Basically, driving secondary (which matches my voltage) draws excessive current but driving primary seems OK.

 
Trust me, I'm a Professional ;)

I have this funny and annoying habit of correcting even the "experts". Because I think things through, rather than just parroting what I was taught in school. The common knowledge is often correct only in special cases.

Yes, if you have 240V source, auto-transformer, two 120V 100A loads distributed across the two phases, their neutral currents flow from one to the other and auto-transformer neutral carries zero. Windings of auto-transformer carry zero.

But I've figured out when neutral carries 200A. This turns out to be a reason for the "120% rule", although the reasonings of NEC committee that I read would indicate they were unaware. Maybe one of them understood, but it wasn't mentioned.

Consider a utility pole transformer sending 120/240VAC to a house over wires appropriate for 100A (e.g 4 awg, rated 95A in a bundle of 3 current carrying conductors but I think round up to 100A breaker is allowed.)

In the house, a 100A 2-pole main breaker and 100A busbars in the panel.

The 120% rule would let me put a 20A PV breaker at far end of panel, but I'm going to outsmart it and put 100A PV breaker there. After all, driving other end of busbar means busbar doesn't get 100A + 100A = 200A anywhere because current is flowing in opposite directions to feed loads. All points in the busbar carry 100A or less, dropping to zero somewhere in between.

Stick breakers on busbar between 100A main and 100A PV breaker, with a total of 200A worth of 120V loads. That's 24kW on one phase "L1".
Let's say that's ten, 20A single-pole breakers.
Note that without PV, if more than 100A loads are active, the 100A main breaker will heat up and trip.


Feed 100A worth of PV (24kW of PV) into the 2-pole PV breaker.
That's 100A into L1, 100A into L2. Each branch circuit breaker draws off 20A from L1, So busbar carries 100A, 80A, 60A, 40A, 20A, 0A as you move along it.
Neutral bar connected to white wires returning from branch circuits carries 20A, 40A, 60A, 80A, 100A.

The PV current on L1 flows into loads, supplying 100A and 12kW, which is enough for half the loads.

With no more current available from GT PV inverter, the remaining five loads draw their current from main breaker on L1. That point on busbar between 20A load breaker #5 (counting from PV end) and breaker #6 is carrying 0A. Closer to main breaker, it carries 20A, 40A, 60A, 80A, 100A. That's 100A coming from main breaker. In addition to 100A from PV breaker.

Neutral bar has ten wires from ten branch circuits, and is fed with one 4 awg wire from utility.
We've counted up to 100A so far, at connection from circuit fed by breaker #5. Keep counting: 120A, 140A, 160A, 180A, 200A.
You now have 200A in the neutral busbar.
200A in the 4 awg utility drop cable, which causes 4x the heating that 100A would.

PV breaker is also feeding 100A into L2. No loads connected, so 100A through 4 awg L2 wire back to utility transformer.
In the utility transformer, power delivered to L2 couples by magic, I mean by magnetic, to the other winding.
This is how 24kW 100A 240V from GT PV inverter interacts with utility pole transformer (could have been your own auto-transformer) to become 24kW 200A 120V in the breaker panel busbar and certain wires. Zero power drawn from grid, from primary side of that transformer.
L1 and L2 each carry 100A at 120V (from ground), 240V from each other. Neutral from transformer carries 200A.

This does not occur with an isolation transformer getting say 24kW 100A 240V on primary. It can't deliver more than 100A to "neutral" center-tapped 120/240V secondary. A 100A 2-pole breaker protecting L1 and L2 is sufficient to protect N as well.

When an auto-transformer is used (or an isolation transformer backfed as I described above), Neutral can carry 2x the current of either L1 or L2, basically the sum rather than the difference between them.
That is OK if wires are heavy enough. Either run 4 wires from the two 120V windings of the auto-transformer to breaker panel, or join the to wires together and have a single neutral wire with ampacity twice what a 2-pole breaker is rated for. Breaker panel busbar must also be rated for that 2x current. If other sources are present (e.g. grid operating in parallel), available current from those sources must be included in the analysis, for an even higher ampacity neutral requirement.

Alternatively, you could use a 3-pole breaker to connect the autotransformer, limiting current from neutral. This could apply in your off-grid setup with auto-transformer, but with limited current from inverter you probably just want to use heavy enough wires.

In the case where utility transformer acts as auto-transformer for your GT PV (or backfeed from battery inverter), 3-pole breaker including neutral probably isn't an option. So obey the 120% rule and be less stressed, knowing neutral wire would only be moderately overloaded.
 
Last edited:
I’ve tested this rewired under a 5000W load on one leg and it handled it like a champ:
View attachment 93001
Inside the “ST-5000” power adapter is a toroidal autotransformer with 0/110/200/220/240v taps, I wired the 0/110/220 through a 3-pole 30 amp breaker. I already have a high-frequency split-phase inverter, I added it to balance loads, but if it trips then the split-phase inverter faults if there is too much load on one leg anyways.

I’m doing zero-export to CT and it balances enough loads so that one side of the inverter doesn’t max out occasionally and I catch more of my usage.

$100 on eBay.
Nice clean set up. I've got the same auto-transformer. Did you by chance monitor the transformer temperature during the 5kw 120vac pull ?
 
I’ve tested this rewired under a 5000W load on one leg and it handled it like a champ:
...
Inside the “ST-5000” power adapter is a toroidal autotransformer with 0/110/200/220/240v taps, I wired the 0/110/220 through a 3-pole 30 amp breaker. I already have a high-frequency split-phase inverter, I added it to balance loads, but if it trips then the split-phase inverter faults if there is too much load on one leg anyways.

Clamp ammeter display is off.
If 5kW load was partly handled by split-phase inverter and partly rebalanced by auto-transformer, how much of the (120V or 240V) current did auto-transformer carry?

Are you able to measure transformer current waveforms?
Determine reactive power, VA of apparent power vs. W dissipated in transformer?

I have not measured any intended for use as auto-transformers (but we have some at work, should try.) The isolation transformers I've been working with are kind of designed on the edge - at spec operating voltage they go far enough into saturation that the current waveform looks like a Hershey's Kiss rather than a sine wave. That puts additional demand on the inverter.

No-load losses are also something most of the transformers I've seen aren't particularly optimized for. Rather cost, I would say. Of course, they were meant for grid use.
 
Nice clean set up. I've got the same auto-transformer. Did you by chance monitor the transformer temperature during the 5kw 120vac pull ?
Not exactly, I pointed my temp gun at it a couple of times and it registered slightly warmer than atmospheric, like ~5F if memory serves me — and that was after running for some time.
 
Clamp ammeter display is off.
If 5kW load was partly handled by split-phase inverter and partly rebalanced by auto-transformer, how much of the (120V or 240V) current did auto-transformer carry?

Are you able to measure transformer current waveforms?
Determine reactive power, VA of apparent power vs. W dissipated in transformer?

I have not measured any intended for use as auto-transformers (but we have some at work, should try.) The isolation transformers I've been working with are kind of designed on the edge - at spec operating voltage they go far enough into saturation that the current waveform looks like a Hershey's Kiss rather than a sine wave. That puts additional demand on the inverter.

No-load losses are also something most of the transformers I've seen aren't particularly optimized for. Rather cost, I would say. Of course, they were meant for grid use.
When I tested it I I ran a bunch of loads through the transformer as it came. I took 240v and then ran about ~4.5kw of resistive load off the 120v for almost an hour.

When it was on the wall in my house setup it would end up balancing about half the current with the inverter handling the other half when in off-grid mode. When competing with the distribution transformer on the grid it was much less. I accomplished more by moving my breakers around to balance each leg.

I’m using it with my portable solar I made with a 240v inverter now.

I would like to do some waveform measurements but I need to get some instruments for it.
 
balancing about half the current

So 5kW 120V 40A load, about 20A from inverter neutral, 30A from one inverter phase, 10A from other inverter phase.
Auto-transformer drew 10A on each phase and delivered 20A on neutral.

There are some Rigol scopes in the $300 ~ $400 range. Or higher end brands used. Current transformers can convert these lower frequency currents to voltage (but won't track millisecond inrush surge.) Need voltage probes good for at least line peak, although those aren't rated for the transients that can come through a power line. Perhaps DIY another 10x attenuator to protect them.

It is handy to have math functions in the scope for FFT, multiplying waveforms, and so on to compute VA vs. Watts. Not sure how much is in the lower end scopes. Deeper memory so not just 512 or so points across the screen.
 
When I tested it I I ran a bunch of loads through the transformer as it came. I took 240v and then ran about ~4.5kw of resistive load off the 120v for almost an hour.

When it was on the wall in my house setup it would end up balancing about half the current with the inverter handling the other half when in off-grid mode. When competing with the distribution transformer on the grid it was much less. I accomplished more by moving my breakers around to balance each leg.

I’m using it with my portable solar I made with a 240v inverter now.

I would like to do some waveform measurements but I need to get some instruments for it.
what is the 3-pole circuit breaker value ? and what awg are the wirings ? couldn't tell if your neutral has larger gauge. I've thought that we've decided that it's a bad idea to put a breaker on the auto-transformer neutral ?
 
what is the 3-pole circuit breaker value ? and what awg are the wirings ? couldn't tell if your neutral has larger gauge. I've thought that we've decided that it's a bad idea to put a breaker on the auto-transformer neutral ?
30A 3-pole, 10awg. It was with a split-phase inverter that has a central neutral, so if the breaker trips it just falls back to the inverter to supply the power to both legs, and if one of them is overload, then it just shuts down. It’s different than when transforming a pure 240v supply.
 
Claimed 5kVA rating at 120V would need 40A on neutral (and 20A on each of L1, L2)
10 awg of course can carry 40A, we just aren't supposed to do that. At least not in house wiring, might be allowed inside "appliances".

Is 10 awg center-tap built in? Or do you have access to a terminal directly fed by both halves of the windings?

As extra balancing, yes that should be safe. I plan to do the same thing, load sharing and creating second phase if slaves are sleeping.
 
Claimed 5kVA rating at 120V would need 40A on neutral (and 20A on each of L1, L2)
10 awg of course can carry 40A, we just aren't supposed to do that. At least not in house wiring, might be allowed inside "appliances".

Is 10 awg center-tap built in? Or do you have access to a terminal directly fed by both halves of the windings?

As extra balancing, yes that should be safe. I plan to do the same thing, load sharing and creating second phase if slaves are sleeping.
It is a single winding toroidal with the two ends and loops marked for 110v, 200v, and 220v. So I tested it first (with 24vac out of a small transformer), and applying 240v across the end wire (that I called 0v) and the 220v loop, gave me exactly half (120v) out of the 110v loop.

I crimped the copper wire to the aluminum windings with some stainless crimps I had. Even though this is 90C wire and can go up to 40A I figured limiting every connection to 30A was a good safety margin (even that that limits the difference to ~3.6kw). It's now part of a portable solar generator project for a camper trailer and should never get near that anyways.
 
Claimed 5kVA rating at 120V would need 40A on neutral (and 20A on each of L1, L2)
10 awg of course can carry 40A, we just aren't supposed to do that. At least not in house wiring, might be allowed inside "appliances".

Is 10 awg center-tap built in? Or do you have access to a terminal directly fed by both halves of the windings?

As extra balancing, yes that should be safe. I plan to do the same thing, load sharing and creating second phase if slaves are sleeping.
here's a pic of the inside before mod, the ST5000 is on the left, right is ST2000. Zoom in to see the 3-connection to the windings

RXzdzNILgdc3VZ8qrG0_QBgvr9RWYsHV-Iy7HYKX18hBxTeDuuGS6SzISej6yBn3AGSwpxC_JtMnykoy92F4IT91bBmuqN49p9FPuqYaTM8QI_3dDPAA4e5yApXg90bGRW0mEW1KqRAebghTdJ3_O-EOyJTfzcAP0suLiH7W5Pxl0MJOkTrCwgMaKIt34-Joc68K3M0k_L0-29i_aNCsT6z_8PwtY9DhfyGWwSPM0tk1YZ-lYpqwdRTNJHSH2uppliNMJ1cGBgC5mL3guYNkU3zjdKX9o3A9O82TEHtbwxwDTCHkX2UhNvZEAjFJBlkjkzm3MfgFfiykSbnXzBqoxYlotugoOCC_QIagQl3vgWxzr6IOMaAKz3b3ZfJ4rNonZt00FhsN4qTB2nznMHsm9yEyL1arQ3aUVS_1_ODZQqH21Wize8e1HhdYEXXhvEOYAnMFZB7mqYHoeg-9OiUdmFzOa9cRcG5BxohgJhu3oKZOYMPPUygvA2swtsCi40pPyK-89psx8xPFx1k1cQ8NAVJ1e5JDcJRSYSU_xdWRhJhzL0_JTdL5quz7HwbiIkW1aWRUHM8jmn-MbSJjaq8HewqIt_33moaOMZBxlOyIlrWe7UlFxt8lJC6v9TynMrEufiIZ41cyJz-pYg4VhsJEsvfNhELDxh8fPazI8TlecXtYx-LTgoyBiPKHRO15dFFU6mGjc8Aj5AGLALFRn1NjHClD8JiERZpmlQPJBSHf0uZ29PV1b4p2Ple2en-g3RJL2VBkPBMvrDLDQELJOeXYjYCQyjy47rthCKeCerVK3DwnkFtXjVpkgK1RqbQy9GZoyE1k4mmNSwmKoUMhGpwp6wjU_e0a9BVmJZrM0fwkmjhhMvELMvB9yJ3xIY1nuepUpyt71aqTSguzWEA4zZ66lwc7fUdduv2tgmSdUBvCY_AbAJqDfA=w927-h695-no
 
Back
Top