Blasphemy..... Solar power.

[starfire]

Starfire- most of the newer 48V inverters are running at 94-98% efficiency; this is "marketing speak", since it is only the inverter loss and not battery and wiring, but still, your figures were much lower than what I'm familiar with in the Magnum Energy line, and every other 3KW+, 48V, name brand inverter I've seen.  At idle (no load), these inverters do eat some power and efficiency then sucks but in fair sized home systems it's negligible. They have come a long way and your figures are fair for some old systems. 12V systems did suffer badly from wiring battery resistance issues, also.


Bruce, if you connect a DC wattmeter to input, AC wattmeter to the output of any inverter, over time, from days to weeks, you will see the real net efficiency of the inverter.  None will achieve that manufacturers rating.... ever.
The real life efficiency the average bloke will get is actually no where near that 94/97 they claim, 70/80 percent means its doing very well.
Inverters spend almost all their life running at a small part of their rated continuous loading, and can only achieve max efficiency  within a very small window.
Same with mechanical generating equipment where max efficiency is also achieved only near max output.

[BruceM]

I had the same impression myself, Starfire, but in looking at Magnum's MS-PAE series (48V) load vs efficiency chart, they have efficiency above 90% from 500 to 3500 watts.  Beyond that in the 80s.  It's pretty damned impressive.

See page 41 of the manual:

https://realgoods.com/downloads/dl/file/id/382/ms_pae_series_owner_s_manual.pdf

Their 24V model is  less efficient; 85-90% from 500 to 3500 watts.

I agree that these figures are high enough to raise credibility issues but my other experiences in working with Magnum engineers is that this an area where they are very proud of their work.  It would take very good toroidal magnetics design to achieve this. The low frequency design of the Magnum does help greatly with MOSFET switching losses.  That plus the newer MOFETs have amazingly reduced gate charge capacitance and incredibly low on resistance.

Below 500W the efficiency does drop precipitously; they don't show below 300 watts where it is only 70%...and likely for good cause(!).  So for low power use, even the best of 4KW inverters really do suck at conversion efficiency. (A smaller unit might be better at this.) This can't really be helped with a low frequency, transformer isolated design.  You'd almost have to have another unit with different magnetics for very low power, and switch between them as the load increased.

[BruceM]

My own 5 step sine inverter is loosely based on the early Trace SW series, low frequency design.  I've only tested it for efficiency (92%) at 500 watts so far, with two surplus 1100 watt toroidal transformers I had on hand which can easily handle a 1500 watt continuous load and 3500 watt peak (my design goal).

Just for kicks, I did some searching and found this efficiency curve from Trace. (see attachment)

No wonder the Trace SW series had such a good reputation...the low power performance is better than most made today.  This was a true low frequency design, not a hybrid.

[starfire]

Yes, agree, that particular example is very good, above a minimum load.
Using their figures.
In "search" mode, the standby current is 6 watts, or around 800mA at the nominal 48 volts.
Inverter overhead when running is around the 25 watts.... this is the voltage sense, mosfet drivers, gate current etc
At a 500 watt AC load .... a  large lighting load using CFLs for an instance , 500/25 = 20 or   around 80 percent efficiency, Pf =1, ie resistive load.
At 1500 watts , efficiency climbs  to 1500/25, or close to their 90 percent.
With inductive or capacitive loads..... most given the proliferation of SMPS in almost all electronic devices, even including CFL lighting, the actual figures will be less, but close.
A period of low energy use, when sleeping for example, will see the lower efficiency over a longer time, perhaps say 8 hours in a 24 hr period. Higher cyclic loads during the day tend to be intermittent, refrigeration, microwave oven, kettle etc,  therefore the higher efficiency periods occurs at random and for shorter periods.
So, the 500 watts drain over the 8 hour period will use  4 kilowatts, and a further 200 watts in "waste"
If the inverter then runs another 16 hours cycled between heavy and light at say a 50 percent duty cycle, another 8 hours can easily be spent running light, giving almost a half kilowatt in wasted consumption total over that 24 hours.
This is why an  efficiency rating over time  is far more informative in real life than an absolute  "snapshot".figure.
I have always advocated using many smaller inverters for this reason, redundancy increases reliability, each inverter can be scaled to that one circuit, a large inverter that uses 20 watts overhead or more can then spent time either heavily loaded or not at all, where we can enjoy the lower 6 watts standby  mentioned above.
 Smaller inverters generally have less phantom overhead than larger types.... some as low as 50mA
This is all academic to many, but when off grid entirely, each watt costs from the battery bank and assumes greater value than an inferior cheaper version  downloaded from the national grid.
A battery watt effectively requires another 5 to store it, given battery cycling depth issues, and to replace that watt takes almost an extra  one  given generator and chemical conversion losses at the bank.

[BruceM]

Starfire, I think you had a Sr. moment.  25watts is included in the efficiency of 90% at 500 watts, and is half of the 10% total loss shown.  It's 25w/500w=0.05 or 5%.  Your point is made though, night time low power loads will affect total efficiency negatively.

Night time phantom loads are a serious problem for newbies to off grid living.  GFCI's, arc detecting breakers, washing machines and other electronics that are actually on when off, all contribute, all night long.  One contributor here calculated his actual phantom load for a large off grid system and it exceeded my actual winter night time power use (incandescent lighting and computer/projector).

The Magnum inverters (and I expect most other off grid inverter designs) see all reactive power needs as real load.  I confirmed this characteristic on my own low speed design inverter; purely capacitive loads do show up directly in DC input current.  This means filter design to minimize both differential capacitance and capacitance to ground (leakage) is critical; every single uF is another 50 ma of current.  A typical surplus facility filter may seriously impact the battery bank SOC overnight.  Genisco and RFI corp have both now have designs specific for home inverters at my request; they draw just 15 watts  instead of the 60 or 100 watts of their standard facility filter.

I feel that all the power system should not be in the house for directly radiated EMI and conducted EMI.  It's easier to just remote them in an outbuilding with batteries and PV nearby.  Then only one AC line to the home need be filtered and shielded. Distance is cheap compared to high performance shielding for the directly radiated emission of the inverter and PV charge controller.  For the absolute minimum health insurance, one of the Schaeffer 2 stage common mode choke filters (about $75) from Digikey or the equivalent seems wise to me. (Pick one with good insertion loss data, designed for switching power supplies.) The independent evidence that EMI on home wiring is a needless and serious health stressor grows every year.  The trend towards crappy, high EMI design switching power supplies in every appliance, lightbulb, etc., is a bad one, as is electronic variable speed motors.  The "modern" home is an EMF horror show.  Even refrigerators now get switching power supplies and digital readouts. An embedded microcontroller is so cheap now that anything with timing and sequencing needs or much logic gets one.  Appliances have been exempt from even the FCC's lax emission standards.

I like 120VDC best for home power...no inverter losses at all, no PF losses.  Also, good for me, zero ELF magnetic or electric fields for resistive loads.

[starfire]

Bruce, its a pleasure to talk with you about this stuff, it seems you have far more experience than I do with the nuts and bolts of these.
I remember the very first inverter I made simply used a copper  commutator cobbled from water pipe, spun with a small 9 volt DC speed controlled tape deck motor, and four  brushes ratted from an old car generator, a mechanical bridge if you like.
This was a red necked  attempt at negating the transistor thermal runaway problems with Mullard OC35 germanium..... a fairly expensive component back then with abysmal specs looking back. and hard to parallel without  excessive losses in the required emitter resistors to force current sharing. The point 2 CE voltage drop was good though. I think they had an Ft something like 20Kc at 8 amps collector current, Hfe of maybe 10
Silicon and Mosfets came decades later
A bank of 10 x 12 volt batteries supplied the juice, and a hefty welder transformer primary winding used as a filter, protecting the brushes during overlap, and giving a fairly decent 230 volt sine.
The efficiency was pretty decent as well. Later, I remember using some special graphite/copper oxide brushes. These had very low axial conduction resistance, with higher resistance at 90 degrees to decrease intersegment losses... or so the advert reckoned. These were made by I think Radcliff in Australia, and did dramatically reduce sparking..

Sometimes I do wonder if this scheme is worth revisiting...... especially as a unmodified  Prius battery could slot straight in.
The voltage regulation was pretty good, with the output load  only ever being a very small part of the huge current reserves available, and it was SIMPLE, although it did require a starting proceedure for obvious reasons.
My recent DIY inverter uses IGBTs ,  ... project on hold until I can get these things to turn off faster. I have tried keeping them just above saturation, this causes losses in the added components , a negative gate current seems to have little effect... unlike sucking out hangers-on mobility carriers via a negative transistor base (NPN).
Basically it has overlap problems.... these  cost a fortune so reluctant to give in to using Mosfets  just yet.
Call me stubborn.

But then, arguably, why even bother with AC and forget any conversion at all. ?
Ninety percent of appliances, electronics, CFLs use a SMPS that happily runs off HV DC mains anyway. Washing machines, dryers, and other appliances no longer use AC motors, they now use steppers as you say, micro controlled and pulsed with lower voltages generated internally via their own internal  SMPS.
Another big efficiency saving right there......
It is almost possible to connect 10 or 20 x 12 volt  batteries straight into the house wiring via the existing switchboard with no other changes. Admittedly, the AC rated switches and circuit breakers may have a harder time, but Im pretty sure most everything will not notice any difference.
Vacuum cleaners, electric drills and skillsaws, grinders, cake mixers  and other lowtech stuff  predominantly use "universal" type brushed motors, and are also happy with DC.
The cost of the odd appliance that may object will adequately compensate for not having to buy an expensive inverter.

[starfire]

A cunning thought here to extract free power from the grid.....
A high current bridge rectifier placed before the electricity meter would  run the house on DC without spinning the meter.....  the old analog meters used AC eddy currents to spin the disk, smart meters use AC current transformers to clock the A/D converter.. DC cannot effect either.
No meter tampering is needed, so nothing illegal there...... and the required wiring  to cut into is easily acsessable.... if done at the pole, you would be home free...

[oldgoat]

If done at the pole here in aust you would probably end up with a fine big enough to negate any savings. They are very sensitive about that here because you may be using it to grow small crops in the back shed.

[starfire]

OG, you are right of course, Im just being difficult. Im fairly sure you would be noticed climbing a power pole in the dead of night, but then, if you wore a high vis vest, and had a white van with a bright yellow ladder, a few orange road cones and did it in the middle of the day...... I very much doubt anyone would suspect a thing.
There was the Masport factory here many years ago that had a major pilfering problem. The staff were searched after leaving each shift by security. One worker would always leave with a wheelbarrow full of sawdust for his garden, the guards would sift through it every time, and never find anything.
Years later they discovered he was stealing wheelbarrows.

[AdeV]

One worker would always leave with a wheelbarrow full of sawdust for his garden, the guards would sift through it every time, and never find anything.
Years later they discovered he was stealing wheelbarrows.

That made me laugh out loud! Thanks!

Agree about using a white van, cones, hi-viz everything. Also, if you keep a Health & Safety questionnaire on a clipboard nearby just in case anyone does come nosing, they'll soon scarper. For Health & Safety, feel free to swap your local organisation, e.g. OSHA, Elfen Safety, etc.

[BruceM]

Starfire, yes,  the essential nature of AC power in the home has dwindled.  Most of the large (computer) server farms have gone to direct 350V DC distribution instead of the 230VAC, since it is directly compatible with the computer power supplies and eliminates inverters entirely.

For home use, 120VDC is quite handy and has 4x the inherent safety of 120VAC. A shock feels more like a static discharge; surprising but not painful. I use various sized crock pots (no mods), a 300W rice cooker (modified), a 600 W small toaster oven, an old clothes iron (no mods), 300 W immersion heater for tea (no mods), coffee grinder (no mods), food processor (no mods), blender (no mods but I use the plug instead of switch),  Hair dryers and heat guns work fine, though I use the plug instead of the switch, with switch taped on to save me from my regular Sr. moments.  With the advent of the IXYS 200V+ P-channel Mosfets for high side switching, when modification is required, it is as simple as can be.  Fancy gate driving is not required for the simplest appliances that have no electronics, since some HV mosfets can now handle a very slow switch and even linear operation when needed.  A 10K ohm resistor to high side from gate, and 100K ohm to neutral (using the existing switching via bimetal thermostat switch or other timers or switches) does the trick nicely for my toaster oven/broiler.  Low side Mosfets are now also available that have a well defined safe operating area for linear/semi-linear/very slow switching. Fairchild has well priced ones.  Better to spend a couple bucks more on the switching transistor and save yourself a bunch of complicated circuitry.

Switching off DC power that doesn't pulse like brushed universal motors is problematic.  DC will nicely span via arc quite a large gap when a fair amount of current and voltage is involved...in testing with a 500 watt test load I fried many heavy duty AC switches on the very first "off". (I did find one DPDT wall switch that holds up well.)  Cooper made 150VDC rated wall switches but they are now out of production.  Fortunately, Leviton and others still make rotary lamp switches that were originally designed for DC and still work nicely, as do normal lamp switches and even the rotary cord switches; they have not bothered to change the design for AC. For higher currents the Midnight Solar DC breakers can be used and are rated for switching use.  Or you can use any puny toggle switch and one of the big SOA Mosfets.

You can do serious work with 120VDC with quite reasonable currents.  So low current that battery efficiency is much improved and wire losses nil.  Edison was no fool.  110/120V was and is a good compromise for safety and utility in the home, even over 100 years later.

 


 

[starfire]

Are you saying you are incorporating Mosfet switching into  these appliances to switch them on and off?
My answer to using AC switch  contacts with DC is simply a reversed diode attached directly across the contacts.  Inductive loads are the problem here, and the non quenched arc can/will weld contacts together very quickly. Once done, they are never a problem. Circuit breakers are even easier as the connections are more accessible.
Not having any grid connection means that NZ legislative practice can largely be ignored and common sense prevails. This saves a huge amount of time and money too.  God help anyone using these back on any AC circuit, would cause much head scratching.....
The main problem using 120/240 VDC direct from battery banks is fire safety, like short circuit current will easily exceed almost any  wire gauge rating, so good reliable fusing is VERY important.
A handy and cheap way to current limit is to use strips of stainless steel in series with the battery source.,  stainless is a poor but stable conductor. I use this idea when welding with batteries also.
Electrolysis too can be a problem in damp climates, I have that problem here at times.
I didnt know DC is becoming more popular, thats very interesting.
AC was and is the go for distribution, easy passive voltage conversion etc, but certainly of limited advantage when source and user are only meters apart.
All your appliances mentioned will happily run on DC.... this does little with your RFI//EMI concerns though, given the SMPS inherent in most electronic stuff.
Most of my entertainment electronics stuff runs off DC  12 volts, TV, radio,  stereo etc having removed the internal SMPSs and substituting a LVDC daughter board.
Lighting is 240 VDC CFLs, to reduce RFI, running off a DIY inverter, and the main 240 VAC 600w sine inverter runs the electronic workshop, as my sig gens, scopes etc use linear power supplies to prevent any RFI from that source.
Its messy using several voltage busses, but works well in practice..

It will be very confusing and possibly expensive mistakes for whoever inherits this place when I die......

I always hoped the Vacuum cleaner door to door  salesman would turn up, spread the sawdust and ash all over my carpets and rugs, and then attempt  to suck it all back up with a machine that wont work on 12 volts......
All I get are Jehovah bloody witnesses, so focused on the end times they care little about the here and now, and are bored shitless with all things Lister, solar and electric.... thats how I get rid of them.

[starfire]

And, health and safety..... we call it OSH.
Curiously, if  say, I paint your house for money, here in NZ, I need to fill out many forms, undertake a risk management assesment, use scaffolding too, are not permitted to paint off a ladder. Checks for lead paint and asbestos, electrical overhead line issues and other hazards.
If I paint my own house, none of this is required.
This seems to mean that falling off a ladder is unlikely if its here, but not there.
Or, my safety/life is worth less if Im home.
Or, gravity is less in my location, more at  yours.
Perhaps my ground is softer.
There is always a good side to this beaurocratic nonsense.
By having a sign designating your home as an industrial site or work area, immediately negates any implied right of access by anyone. They are required to report to the "site manager" before entering the property, and submit to any protective gear and safety restrictions deemed necessary, and have all known, and unknown dangers pointed out. This process can take considerable time. ....
For those like me that have say, unpermitted buildings, illegal mains wiring, illegal quantities of stored toxic chemicals and fuel, etc many sins can be "corrected" or hidden before entry actually takes place.
Usually it turns out that the initial importance and critical nature of the inspection turns into accepting a verbal assurance that everything is fine.... its only a job to them after all, with bigger fishes to fry....they are also vastly under resourced.

[BruceM]

Starfire, I'll  try your diode method on an AC switch tomorrow and see what happens with a 500 watt incandescent heat lamp load. I do use that commonly for 12V solenoids and relays, but didn't think to try it for 120VDC lamps. I'll let you know how it works out.  I

DC rated switches and relays often use a permanent magnet to stretch the arc to extinguish it, along with a snap action and larger open throw of the contacts.  Electric trolleys in Denmark use compressed air to blow out the arc! 

Midnight Solar has DC rated breakers for 150 and 300V.  I use those, plus a DC rated fuse right at the batteries, for as you suggested, an extremely high current short can possibly weld breakers.  The fuses are sold for high voltage PV systems and are fully rated for DC operation.  DC Breakers and fuses are readily available and fairly cheap here; single 150VDC breakers about $10 US, so I have no need to mess about.

Yes, you're right in that DC instead of AC does nothing special for the EMI from switching power supplies (or any other sources), with the big exception that there is zero power loss penalty for capacitance in passive filters.  That is a huge help.

Like you, I find 12VDC very convenient for low power DC use.  I distribute 12VDC for my home and other shop controls such as Lister controls, solar hot water and heating system and circ pumps, etc.  I use glass and plastic fiber for all control signals to/from the home.  I don't have ANY switching supplies (or any electronics gadgets) in the home at all.  My rear projection workstation is powered via switching supplies in a steel pedestal outdoors. The supply is heavily filtered, but there is also a surplus -110dB from 10K-4GHz military grade facility filter for house 120VDC and another for the 12VDC.  The LED light source "pocket projector" is in a shielded enclosure; a welded 21" cube of steel, honeycomb vents, ITO glass. The projection "booth"/shed is outside the house shield envelope.  Keyboard and trackball are powered from the house 12V, but they are both custom designed;  actively strobed keyboards and unshielded processors are a show stopper for me. Audio is limited to a single plastic fiber controlled speaker, 12V powered, as my audio processing is so screwed up I can't enjoy music any more.  All home wiring is in EMT steel conduit with compression fittings, also.  My home is designed as a shield room and provides a measured -110dB reduction (-55 dBm power) at 2.4Ghz.  It's much more shielding than I needed to sleep OK here now (peak levels below -70dBm seems to do the trick and it's about -62dBm outside now) but it was an experiment to see what could be done, practically, and insurance against new cell towers and MM wave cellular back links.  The 50 thread per inch stainless steel screens outside the windows is the only thing that shows as slightly unusual looking.

I did develop successfully a very simple ultra low EMI 60-100Hz switching supply (a push pull center tapped toroidal primary with stock small toroidal transformers intended for AC use) for 120V DC to DC conversion.  By using super slow op amps (30K bandwidth) and switching on/off time in the 50 usec range, only the diode noise in the DC rectification is an EMI issue, and it can be snubbed and filtered into submission.  There is some stray ELF magnetic field from the toroids for transformer and critical DC filtering inductance, but conducted emissions are virtually non-existent.   A disappointment is the lack of QC for the Talema brand toroids; the two sets of 115V primary windings can be off by a few turns so testing and tweaking of a few ohms resistance in the higher turns winding is needed so that I can use the stock transformers.  I don't like winding toroids when there are hundreds of turns!  This isolated supply design is intended for use with a low dropout linear regulator to remove ripple. I haven't messed with some regulation of output voltage via PWM with this simple op amp/comparator oscillator based design yet, step down is purely transformer turns ratio.  Adding a microcontroller adds so much complexity in shielding and filtering that I avoided it.  In my inverter design I did add a shielded and filtered AVR (Arduino) controller.  It's the simple way to do fancy logic and timing, RMS voltage calculation and allow adjustable dead time and delays as needed with minimal circuitry.

I'd like to experiment with a low frequency, ultra low EMI buck converter with a gapped tape wound toroid for the main conversion DC choke, running at a few hundred Hz but my project wings have been clipped, health-wise, lately.  My project backlog runneth over.



 

[starfire]

Bloody hell Bruce, can I suggest you live in a steel shipping container with chain mail vest and matching pants? Talk about the ultimate Faraday cage...... Why your concerns on RFI/EMI?
 If you are having welded contacts with resistive loads, I would think its current overload, rather than arcing?
The diode actually shorts out the reverse EMF from an inductance, preventing the arc from starting, I cant see a diode helping with resistive loads, maybe a capacitor here will work better? The old car distributor ignition type caps were handy for this, but getting hard to find now, their construction made them super reliable. A mains rated z/y cap will  do ok too. Just to soak up the opening spark that initiates the arc as the contacts open.  I have also seen a Triac in series, the gate connected to M1, the Triac turns off as the switch opens preventing arcing.  Any arcing is of an AC nature that will eventually turn the triac off, like in milliseconds.... I think its rude and inelegant, and have never tried that one.
Whatever, but ll be keen to know your results
The rest relating to your electronics stuff I cannot comprehend, its far tooooo modern for me.... I still struggle with my Sony MP3 player.