So as I mentioned before, after many many hours of constant full-power work, the plugs at the bottom started to foul up and make that arcing noise and the lights in the house would flicker. It might also have something to do with the stiff Surfix wire that was wired to them that might have caused it to pull slightly to one side. So while a disassembled Multiplus is not something you see every day, I took a few pictures.

Control board removed:

Power supply board pulled out for better access. Note my makeshift terminal blocks attached to the input/output, a temporary fix I used.

New plugs installed.

Reassembled, factory reset and tested.

@PJJ, I got your note of interest. Let me just think it over a bit more :-)

11 minutes ago, plonkster said:

I got your note of interest. Let me just think it over a bit more :-)

No problem

Take your time, as long as I am first in the Que

If you still want a spare Multi, we can always do a swap + cash from my side?

5 hours ago, plonkster said:

Let me just think it over a bit more :-)

 

I have sheet music for that! It's in A-flat major. I thought it might be funny to learn to play it, to mess with my kids, but didn't get around to it :-)

It's been a trusted partner for over half a decade :-)

@plonkster, im having trouble with a repair on an same unit, long out of warranty 2010 i think. anyways maybe some insight what i could check next. i swapped out all the relays on the main ac switching because i was convinced a contact was burned off or fused. symptom is that runs fine from battery, you give it grid and it synchronizes, checked all that, the wave-forms align nicely. when synced it latches hums and gives in with an error. the output voltage drops like its fighting the grid but i don't see how.

 

 

I'm afraid I am not THAT skilled in working on these things. If you sent it to one of the repair centers, they'd probably just swap out the entire affected module. Maybe something is wrong with the charging side? After it connects to the grid it usually starts charging.

Edit: In other words, disable the charger and see if it connects to the grid correctly without an error (I think it will go into passthru mode instead).

Edited September 2, 20186 yr by plonkster

On 2018/09/01 at 3:50 PM, plonkster said:

Let me just think it over a bit more :-)

Still thinking?

 

I have accidentally left out the MOV on the output side. It probably doesn't matter, but I'd prefer to put it back in. I also haven't tested the AC side, just checked that it turns on. At the very least I need to test THAT.

19 hours ago, plonkster said:

I have accidentally left out the MOV on the output side. It probably doesn't matter, but I'd prefer to put it back in. I also haven't tested the AC side, just checked that it turns on. At the very least I need to test THAT.

No worries, as you have seen, I have patience

Interesting, looks like that inverter is a low frequency design. Did not expect that.

Also explains a lot regarding the price

1 hour ago, Gnome said:

Interesting, looks like that inverter is a low frequency design. Did not expect that.

What makes you think that? The big toroid?

This is a 1.6kva unit. Have you seen a 1.6kva low-frequency transformer?

:-P

11 minutes ago, plonkster said:

What makes you think that? The big toroid?

This is a 1.6kva unit. Have you seen a 1.6kva low-frequency transformer?

:-P

Yes and yes.

Size is far too big for high frequency type, windings are obviously not litz wire.

Additionally the size they chose is suitable for a fan cooled transformer. You can make the size of the transformer smaller at a cost of higher magnetizing current and higher no load draw (relatively speaking for a toroidal transformer which typically have really low no-load draw values).

Another give away is the circuit board design, I haven't seen the detail of the board yet but there is no way they have a boost stage, AC->DC stage and then inverter stage in that tiny high power PCB. It is just just too small.

Then again I don't have much size reference but it certainly looks like a low frequency design

Edited November 28, 20186 yr by Gnome

Found some pics of the higher power units where you can clearly see they parallel up low frequency transformers:

 

Maybe on some of the newer models they are using a high frequency design, but typically that would not be a toroidal transformer for high frequency.

15 minutes ago, Gnome said:

Another give away is the circuit board design, I haven't seen the detail of the board yet but there is no way they have a boost stage, AC->DC stage and then inverter stage in that tiny high power PCB. It is just just too small.

Truth be told, I too have thought that it seems a bit too simple. Maybe it is low frequency... but I always thought it wasn't.

Edit: OK, this is embarrassing. I should know this stuff. Yes, it is a low frequency design.

Edited November 28, 20186 yr by plonkster

So after that revelation, I starter to wonder if this is necessarily a bad thing. Apparently it's a bit of a pros and cons thing.

High frequency costs less, weighs less, has a lower quiescent consumption and is 1%-2% more efficient.

Low frequency costs more, weighs more, is a bit less efficient, but has lower impedance on the output side and can start heavy loads with less fuss (apparently, as a rule of thumb, twice what the HF inverter does). And it is easier to tie grid-tied inverters to it :-)

This is actually interesting if we do another one of those infernal Victron/Axpert comparisons. Axpert is HF... yet the quiescent draw is not better, and it is not really much more efficient.

On the other hand, that's why you use toroids, to claw back some of what you lose.

1 hour ago, plonkster said:

So after that revelation, I starter to wonder if this is necessarily a bad thing. Apparently it's a bit of a pros and cons thing.

High frequency costs less, weighs less, has a lower quiescent consumption and is 1%-2% more efficient.

Low frequency costs more, weighs more, is a bit less efficient, but has lower impedance on the output side and can start heavy loads with less fuss (apparently, as a rule of thumb, twice what the HF inverter does). And it is easier to tie grid-tied inverters to it :-)

I didn't really say it was good or bad. I was just surprised because low frequency has a lower development cost but substantially higher manufacture cost.

Both designs have pros and cons as you say

The biggest pros are probably the THD and isolation provided by the transformer in my mind for LF. The control circuitry is also exceedingly simple in comparison the HF which is very complex. (ie> reliability)

I'm not really sold on the starting heavy loads because LF has a higher overall impedance, so you almost certainly get a voltage sag when you start a heavy load. Whereas a HF design can theoretically have incredible low impedance to the point that in the Google Inverter challenge they built inverters that didn't even need capacitors and had minimal inductance.

I think that is maybe something that was true in the past but is just marketing wank now.

1 hour ago, plonkster said:

This is actually interesting if we do another one of those infernal Victron/Axpert comparisons. Axpert is HF... yet the quiescent draw is not better, and it is not really much more efficient.

Yeah I suspect Axpert has a number of inefficiencies.

Ranging from cheaper older style MOSFETs, using as small as possible transformer and likely having power hungry control circuitry to name a few things I've noted.

For sure HF can be more efficient, but efficiency needs to actually be a design goal if you follow

1 hour ago, plonkster said:

On the other hand, that's why you use toroids, to claw back some of what you lose.

Could be but if I may say so. The transformer is on the smaller side. If you compare it to a 100% duty cycle 1:1 isolation transformer the core is undersized. And having wound my own transformers I can tell you this is where most companies cut costs sadly. There is a formula I typically use provided in a nice video about winding transformers

They want a smaller core size and lower copper count because it really increases cost quickly across multiple units. The inverter would likely have significantly better efficiency if they increased the core size and/or increase the winding count.

To magnetize the core a certain amount of magnetomotive force is needed. The fewer turns and smaller your core the more current is consumed, called the magnetization current.

To put it another way, your transformer pushes back when you put power into it without putting a load on it resisting change. But if you have fewer windings it requires more current to create that magnetic field to push back. That is why transformers have so many windings (eg. a 240v -> 12v transformer has thousands of windings not just 240:12 or 20:1).

That core is about the same size as my toroidal 1kVA isolation transformer core which is slightly undersized even for 1kVA (I think my calculation amounted  to around 800VA based on the circumference)

It isn't unusual tho, cheap UPS have even smaller transformers (APC is bad and brands like Mercer just shocking). I've salvaged a lot of transformers :p

The companies can save a HUGE amount of money by burning some power as heat in the transformer so they really push it to the point where it is good enough.

Edited November 28, 20186 yr by Gnome

Hmmm... .KitKat

For anyone interested, victron employ what they call a hybrid design.

The accuracy of a HF with the surge handling of a LF.

SinusMax - Superior engineering Developed for professional duty, the Phoenix range of inverters is suitable for the widest range of applications. The design criteria have been to produce a true sine wave inverter with optimised efficiency but without compromise in performance. Employing hybrid HF technology, the result is a top quality product with compact dimensions, light in weight and capable of supplying power, problem free, to any load.

 

Also detailed here by Victron

https://youtu.be/UPfUn5ki7OM

11 hours ago, Toopy said:

For anyone interested, victron employ what they call a hybrid design.

The accuracy of a HF with the surge handling of a LF.

SinusMax - Superior engineering Developed for professional duty, the Phoenix range of inverters is suitable for the widest range of applications. The design criteria have been to produce a true sine wave inverter with optimised efficiency but without compromise in performance. Employing hybrid HF technology, the result is a top quality product with compact dimensions, light in weight and capable of supplying power, problem free, to any load.

 

Also detailed here by Victron

https://youtu.be/UPfUn5ki7OM

Not sure about all this marketing w@nk but that topology is exactly the same as APC used roughly 20 years ago.

They are using a lot of discrete components.  Not only that, they are using discrete components from the 1960-1970s (package type) rather than more modern surface mount components.

It is unlikely their design has changed in a very long time.

Edited January 13, 20214 yr by Gnome

1 hour ago, Gnome said:

Not sure about all this marketing w@nk but that topology is exactly the same as APC used roughly 20 years ago.

They are using a lot of discrete components.  Not only that, they are using discrete components from the 1960-1970s (package type) rather than more modern surface mount components.

It is unlikely their design has changed in a very long time.

Yep, why fix something thats been proven to last.

Victron have been using that design for 20yrs as well, the newer multiplus II and smaller multi's are using more modern manufacturing methods which result in lower standby power, however I would guess the circuit designs are very similar.

Edited January 13, 20214 yr by Toopy

2 hours ago, Toopy said:

Yep, why fix something thats been proven to last.

Victron have been using that design for 20yrs as well, the newer multiplus II and smaller multi's are using more modern manufacturing methods which result in lower standby power, however I would guess the circuit designs are very similar.

Not to dis Victron. They are magnificent... pricey, but rock solid. However, that damned big transformer has a down side. They require more batteries than comparably specced systems in order to start up.

So they start out twice the price of other systems and then the money starts to really get out of hand!

Just my experience. Your mileage may vary.

Edited January 13, 20214 yr by PaulinNorthcliff

23 hours ago, Toopy said:

Yep, why fix something thats been proven to last.

Well it just means that they are making pure profit.  They aren't investing in R&D.

From a design perspective, something newer with surface mount components would be much cheaper to build and require fewer components.  It would also allow them to do things like respond to transients faster.  With a huge transformer like that, it must have a lot of surging/sagging when a huge load is turned on, the inductance is insane.

GaN based inverters are going to catch on in the low cost sector sooner rather than later at which point they'll be hard pressed to compete anymore.  I expect at that point they'll redesign.

Edited January 14, 20214 yr by Gnome

On 2021/01/14 at 8:32 PM, Gnome said:

Well it just means that they are making pure profit.  They aren't investing in R&D.

From a design perspective, something newer with surface mount components would be much cheaper to build and require fewer components.  It would also allow them to do things like respond to transients faster.  With a huge transformer like that, it must have a lot of surging/sagging when a huge load is turned on, the inductance is insane.

GaN based inverters are going to catch on in the low cost sector sooner rather than later at which point they'll be hard pressed to compete anymore.  I expect at that point they'll redesign.

They may have been making pure profit on their inverters however they certainly have been putting it back in and spending on R&D. Just look at the amount of new devices they have released in the last few years.

Im sure victron will keep up and release it’s own gear utilising GaN E-HEMT‘s

Your still going to need a transformer to bring the voltage up in a low voltage system and I would still prefer the large toroidal due to its surge handling capacity.

 

Edited January 26, 20214 yr by Toopy

16 hours ago, Toopy said:

Your still going to need a transformer to bring the voltage up in a low voltage system and I would still prefer the large toroidal due to its surge handling capacity.

Why would a large transformer help surge capability?

In my mind it would reduce surge capability, or at least the large inductance would dramatically decrease the ability to respond quickly to a surge.  The lower your inductance the faster you can supply current.  You'd want as little inductance as possible with a large capacitance to respond quickly to a surge and keep the voltage stable.

High inductance would add significant resistance in responding to a transient so you would see a large voltage drop and ringing.  Maybe that is where this idea that a large transformer is better for surge comes from.  Because the faster you respond to a surge the more likely you are to trip a breaker as the voltage drop means the current drops too preventing a breaker from tripping.

But it is only better if you like dimming lights and brownouts when switching on high current devices...

Edited January 27, 20214 yr by Gnome