Author Topic: Inverter Build Project.  (Read 5743 times)

0 Members and 1 Guest are viewing this topic.

Offline welshman

  • Full Member
  • ***
  • Posts: 136
  • Karma: +6/-1
Inverter Build Project.
« on: September 15, 2020, 06:21:32 pm »
As a backup inverter I plan on utilizing this casing and transformer.

I've been brainstorming lately on how to build a power board / mosfet circuit and i think i've come up with something that is new and should be interesting to see.

over the coming weeks i'll be putting this together and documenting it's build.

this is what im starting with.


Offline Pete

  • Sr. Member
  • ****
  • Posts: 385
  • Karma: +19/-1
  • Mount Barrow Tasmania
Re: Inverter Build Project.
« Reply #1 on: September 16, 2020, 05:43:43 pm »
I am interested in how much Idle current that type of transformer draws. Old welding transformers like that are pretty easy to come by these days with everyone going for inverter welders. They may work out well for inverter builds. Just wondering about the looseness of the magnetic circuit and if they have horrendous idle currents.
Good luck
Pete

Offline ClockmanFrance

  • Full Member
  • ***
  • Posts: 171
  • Karma: +11/-0
Re: Inverter Build Project.
« Reply #2 on: September 18, 2020, 02:24:41 pm »
I have had a APC UPS 5000va, that would self start and would give about 3.7kw of 240vac from my 1100ah 48v battery bank.

The problem was that type of sheet laminate EW transformer was eating 200 watts from my precious batteries just running.

Now days i use a toroid wound core, and with suitable Ferrite Choke my new OzInverter Inverter now only uses about 30 watts, so far more efficient using a toroid core.

However, toroid cores and the Inverter power board need to soft start.

Offline welshman

  • Full Member
  • ***
  • Posts: 136
  • Karma: +6/-1
Re: Inverter Build Project.
« Reply #3 on: September 26, 2020, 08:03:18 am »
ill let you all know what i get up to and how it performs.. i've ordered all the bits needed and drawn up the pbd in eagle and had it printed by pcbway. just waiting for everything to arrive. im going to try to take a few lessons learnt with the other inverter and put them into practice with this "backup" device.

the mosfet board design i've gone with is quite a bit different to the normal.  i've made sure that each pwm to the mosfets has exactly the same length of signal wire so the timings of the chips should be in sync a bit better.

the actually mosfet bridge will be something im sure you will like to look at, like i said its a radical design, fingers crossed that it makes sense in the end.


i've ripped up and inspected the relay on two powerjack control boards now and it looks like they are the culprit when blowing up mosfets. my figuring is, due to the dirty contacts on the relay, when the generator stops and slows down the pj is supposed to open the relay to hand the mains back to the inverter. but if the relay doesnt break fully and is noisy. the generator as it slows down is going to get its decaying sine imprinted slightly on the inverters own sine. iv'e bought a 100amp solid state relay to replace the standard one. the state of the contacts on the standard relay is bad enough to be causing a problem on its own. maybe it's why it took many years of switching on and off between inverter and charge multiple times times a day before it blew the first set of mosfets. who knows?


Offline Pete

  • Sr. Member
  • ****
  • Posts: 385
  • Karma: +19/-1
  • Mount Barrow Tasmania
Re: Inverter Build Project.
« Reply #4 on: September 26, 2020, 06:04:56 pm »
Sounds like a good plan to swap the relays with a Solid State Relay. There may be a problem with time lag in the relays too.
Just wondering if you are going to put some delay in the SSR relay switching times to match the mechanical relays or whether you think the faster switching will not upset the mosfets more.
Look forward to reading about the build
Pete

Offline rossw

  • Senior Moderator
  • Hero Member
  • *******
  • Posts: 879
  • Karma: +35/-0
  • Grumpy-old-Unix-Admin
Re: Inverter Build Project.
« Reply #5 on: September 27, 2020, 05:03:01 am »
iv'e bought a 100amp solid state relay to replace the standard one.

Just a cautionary word.
Check which type of SSR you get. Some have all sorts of electronics to stop them creating switching transients, and they may not turn off when you think they will - so do your homework on their suitability before you put them in line!

Offline welshman

  • Full Member
  • ***
  • Posts: 136
  • Karma: +6/-1
Re: Inverter Build Project.
« Reply #6 on: September 27, 2020, 05:38:11 am »
Sounds like a good plan to swap the relays with a Solid State Relay. There may be a problem with time lag in the relays too.
Just wondering if you are going to put some delay in the SSR relay switching times to match the mechanical relays or whether you think the faster switching will not upset the mosfets more.
Look forward to reading about the build
Pete

iv'e bought a 100amp solid state relay to replace the standard one.

Just a cautionary word.
Check which type of SSR you get. Some have all sorts of electronics to stop them creating switching transients, and they may not turn off when you think they will - so do your homework on their suitability before you put them in line!

this is what im using

some of the specs

<10ms on/off time
7.5ma 12v trigger

now there is something else that's interesting about the solid state relay. it only turns on and off at the zero cross point of the sine wave, to avoid surge, rfi/efi. could this be a solution to the charging problem?

8756-0

Offline rossw

  • Senior Moderator
  • Hero Member
  • *******
  • Posts: 879
  • Karma: +35/-0
  • Grumpy-old-Unix-Admin
Re: Inverter Build Project.
« Reply #7 on: September 28, 2020, 01:19:36 am »
now there is something else that's interesting about the solid state relay. it only turns on and off at the zero cross point of the sine wave, to avoid surge, rfi/efi. could this be a solution to the charging problem?

It's exactly this that I'm referring to.
I've seen some of these relays refuse to turn off at all, depending on the nature of their "zero-crossing" detector.

For a giggle, lets hypothetically ask a couple of questions?
Is the zero-crossing based on volts, or current?

If you have a reactive load, current will lead or lag volts, so when volts=0, current will be non-zero.

One would assume that in order to minimise RFI/EMI, it would be switching at 0 *CURRENT*, but most of the ZCD circuits I've seen are zero *VOLTS*.
Depending on the technology in the SSR (and I've seen it happen) - with certain scenarios of load, they simply won't turn off.

Offline Pete

  • Sr. Member
  • ****
  • Posts: 385
  • Karma: +19/-1
  • Mount Barrow Tasmania
Re: Inverter Build Project.
« Reply #8 on: September 28, 2020, 01:59:46 am »
As you say Ross rarely are loads purely resistive where Volts and Current are in phase.
In an Inductive load the Current will lag the voltage.
I a Capacitive  load Current will lead the voltage.
So depending on load the SSR may turn off early or late.
That is one reason I suggested a time lag be built in.
Maybe in some situations a very fast reaction to the mains dropping of is necessary, such as in running a computer with no UPS but otherwise what is a few seconds. Smoke coming from electronics is rarely as easy to fix as the movies show, where they just hit the equipment with a fire extinguisher and it all comes on with no problems. ( learnt that from Lost in Space when I was a kid) never seen it work in real life.
Pete

Offline rossw

  • Senior Moderator
  • Hero Member
  • *******
  • Posts: 879
  • Karma: +35/-0
  • Grumpy-old-Unix-Admin
Re: Inverter Build Project.
« Reply #9 on: September 28, 2020, 02:12:39 am »
As you say Ross rarely are loads purely resistive where Volts and Current are in phase.

It gets (potentially) much worse in the situation of switching a generator, and this is one that may slip notice for some - if you have two non-phase-locked signals (such as a free-running inverter and a free-running generator), and lets say they are both "nominally" 240V (in this country that would be the norm), now you have a signal with a peak-to-peak voltage of 340V, and another peak-to-peak also of 340V, and when they're completely out of phase, your relay COULD see the full whack across its input/output. (Depends how things are connected, obviously, but in my case it does apply). You're going to want that silicon to have plenty of headroom!

Quote
That is one reason I suggested a time lag be built in.
Maybe in some situations a very fast reaction to the mains dropping of is necessary, such as in running a computer with no UPS but otherwise what is a few seconds.

Lots of electronics stuff won't like more than a few cycles of outage.
Lots of things with compressors will absolutely hate "short duration" outages - half a second, anything up to potentially a few minutes - many will try to re-start a motor with full head pressure, stall and burn out!

Quote
Smoke coming from electronics is rarely as easy to fix as the movies show, where they just hit the equipment with a fire extinguisher and it all comes on with no problems.

Ahh yes, movies are a world unto themselves. Any relation to real life is purely co-incidental :)

Offline Pete

  • Sr. Member
  • ****
  • Posts: 385
  • Karma: +19/-1
  • Mount Barrow Tasmania
Re: Inverter Build Project.
« Reply #10 on: September 28, 2020, 09:11:32 pm »
Any clues Ross on why some inverters can handle load switching easily and others not.
I remember reading that PowerJack inverters can have problems and blow mosfets when switching from stand alone to mains inputs, and that PowerStar W7s were a bit more robust and handled it well. ( I think it was Oz that said that)
I am guessing that the transformer type may have had something to do with it.
Pete

Offline rossw

  • Senior Moderator
  • Hero Member
  • *******
  • Posts: 879
  • Karma: +35/-0
  • Grumpy-old-Unix-Admin
Re: Inverter Build Project.
« Reply #11 on: September 28, 2020, 11:18:25 pm »
Any clues Ross on why some inverters can handle load switching easily and others not.
I remember reading that PowerJack inverters can have problems and blow mosfets when switching from stand alone to mains inputs, and that PowerStar W7s were a bit more robust and handled it well. ( I think it was Oz that said that)
I am guessing that the transformer type may have had something to do with it.
Pete

Not really, because I haven't delved into their bowels to find how everyone is doing it.
I have an expensive commercial unit specifically built and marketed to be "generator interactive", they go to great lengths to do the transition "gently" - when the inverter senses input on the generator input, it monitors the genset voltage and frequency and adapts its own output to match the same voltage, frequency and phase. Only then does it pull in the contactor. Once that is done, the inverter then GRADUALLY changes its output to transition the load across to the genset, and then GRADUALLY starts taking as much power as it is permitted to by the configuration (generator capacity less instantaneous load = available capacity) to charge the batteries.

It does the same thing in reverse when isolating from the generator.

I've never heard of any of this type of inverter suffering the kind of problems some have experienced with the chinese inverters, but I suspect that is largely by the design and perhaps partly by the choice of components.

As to why some inexpensive inverters survive and others don't, I could only guess (and probably be wrong!)

Offline noneyabussiness

  • Full Member
  • ***
  • Posts: 155
  • Karma: +7/-0
  • I think it works....
Re: Inverter Build Project.
« Reply #12 on: September 29, 2020, 11:46:36 am »
Oz explained it the best somewhere here, toroidal transformers are electrically " stiff " and as ross said to switch over cleanly they need to match the phase and frequency and voltage etc. before changing over, PJ's don't as far as im aware.  There is a version that has a series resistor accross a relay setup to " soften " the change. The W7's did it ok because they used a E/I laminated core, which has enough " give " electrically that the mosfets etc. could survive the change over..

If you do a search,  OZ and Madness, by memory go into depth about it...

Offline welshman

  • Full Member
  • ***
  • Posts: 136
  • Karma: +6/-1
Re: Inverter Build Project.
« Reply #13 on: October 04, 2020, 10:20:48 am »
This seems to be a very nice transformer..

it even has the ability to move the coils. there is a piece of aluminium, with some grub screws that is bent by the screws to fix the coils in place.

what would be the best option here? leave them as they are or move them closer?

It weighs 46.5KG.

Offline Pete

  • Sr. Member
  • ****
  • Posts: 385
  • Karma: +19/-1
  • Mount Barrow Tasmania
Re: Inverter Build Project.
« Reply #14 on: October 04, 2020, 03:49:17 pm »
Most of the coupling between the coils will be via the magnetic circuit so I am not sure about whether moving the coils will make much difference.
If it were mine I would put some power onto it and measure the idle current, then move the coils and see if it makes any difference.
The main problem with that sort of transformer will be how much idle current it draws.
Pete