is that true that sunsynk inverter(High frequency/Transformer-less) are not good for high load?? and should go with LF inverters(transformer)

and mofsets will burnouts..

i would like to run about 4000w for 10 hours a day..

can sunsynk handle that ? 

 

Edited February 11, 20223 yr by Mahdi

1 hour ago, Mahdi said:

is that true that sunsynk inverter(High frequency/Transformer-less) are not good for high load?? and should go with LF inverters(transformer)

and mofsets will burnouts..

i would like to run about 4000w for 10 hours a day..

can sunsynk handle that ? 

 

Who suggested that it wouldn't or couldn't?

"Low Frequency" is an older design that first uses MOSFETs to create an AC signal then drives it through a transformer to up the voltage.

"High Frequency" first boosts the voltage the voltage to > ~400v RMS AC (very high frequency AC), rectifies it (usually with MOSFETs for higher efficiency than a diodes) to 400v(ish) DC, then smoothes it (using capacitors) then uses MOFSETs again to create an AC signal (the same as the low frequency design but at much, much higher voltage so much lower current, so fewer MOSFETs required so higher efficiency)

If ever "low frequency" had an advantage that has long since disappeared.  Its most significant advantage is simplicity of design (so easier at the design phase).  The higher frequency designs have higher efficiency, lower inductance (and can thus respond faster to transients), lower weight (50/60hz transformer need to be huge).

Another big factor against low frequency is your design phase is cheaper but your manufacturing costs will be higher because it is dominated by the HUGE cost of that transformer.  And because the transformer is so important, it plays a huge role in your efficiency.  For example, if you look at a Voltronic 5kVA inverter's transformer, it is actually pretty small for the 5kVA rating.  I know this because I bought a toroidal transformer for lab purposes and the cost of a 5kVA toroidal transformer is so high that I considered actually buying a Victron inverter JUST for the transformer.  So they are sacrificing some efficiency to save cost and weight during manufacturing.

The high frequency inverter has a transformer in its boost converter stage but it is relatively tiny in comparison, so they don't need to cheap out on the transformer.

In theory and practice high frequency designs have lower impedance and lower losses.  In the early days before electronics were so good as now, maybe.  But these days there is no way a company can compete with the most expensive high frequency designs.

Unless your design criteria is simplicity, then, yeah sure.

Edited February 11, 20223 yr by Gnome

6 minutes ago, Gnome said:

"High Frequency" first boosts the voltage the voltage to > ~400v RMS AC (very high frequency AC), rectifies it (usually with MOSFETs for higher efficiency than a diodes) to 400v(ish) DC, then smoothes it (using capacitors) then uses MOFSETs again to create an AC signal (the same as the low frequency design but at much, much higher voltage so much lower current, so fewer MOSFETs required so higher efficiency)

"MOFSETs" in that bridge are usually IGBTs.

 

7 minutes ago, Gnome said:

I know this because I bought a toroidal transformer for lab purposes and the cost of a 5kVA toroidal transformer is so high that I considered actually buying a Victron inverter JUST for the transformer. 

The transformer you get in a 5kVA Victron will not match an actual 5kVA rated toroid - you can get a lot out of a transformer as long as you keep it cool - which is why Victrons have fans and derate very quickly.

3 hours ago, P1000 said:

"MOFSETs" in that bridge are usually IGBTs

Still a MOSFET 😛 I sort of see the type of FET as being an implementation detail.  They'll use whatever has the lowest on resistance and lowest switching losses with adequate current and voltage ratings.

3 hours ago, P1000 said:

The transformer you get in a 5kVA Victron will not match an actual 5kVA rated toroid - you can get a lot out of a transformer as long as you keep it cool - which is why Victrons have fans and derate very quickly.

I've actually wound a few transformers now and I'd say I'm somewhat versed in transformer theory (though I always feel like I could learn more about magnetics).  But the theory isn't so much important as the fact that they are wasting quite a lot of heat on these low frequency designs in the transformers. (I typed up a lot about how they are almost certainly driving the core to saturation due to its size, but I don't think it matters too much, the outcome matters)

Even if they were to really up the transformer size, I'm still not convinced by the design because your inductance must be mega.  Any transient you need to respond to in your circuit is bogged down by this enormous inductance.  I'm somewhat curious how a LF design like Victron responds to sudden demand changes (ie. troubling electronics like TRIACs) compared to a high quality HF design (maybe even compared to their own HF design).  Going through the various parts of the HF design, I can also think of reasons why it is likely easier to optimise there and get higher efficiency whereas LF design is somewhat fixed on what it can do.

Edited February 11, 20223 yr by Gnome

no one yet answer my question

7 minutes ago, Mahdi said:

no one yet answer my question

Simple answer is yes, the 5.5 kW or 8.8 kW Sunsynk inverter will easily handle your load (4 kW) without problems for long periods of time.

Just expect the fans to run while the load is been fed from solar or battery. 👍

Edited February 11, 20223 yr by TimCam

If you have a relatively large PV array and you are feeding into the grid the inverter will be maxed out at whatever capacity its rated for most of the day,4kw for 10 hours should be no problem at all. I have seen my inverter pinned at 5kw more than 6 hours a day.

2 hours ago, Nexuss said:

If you have a relatively large PV array and you are feeding into the grid the inverter will be maxed out at whatever capacity its rated for most of the day,4kw for 10 hours should be no problem at all. I have seen my inverter pinned at 5kw more than 6 hours a day.

Now what if I have a small array, 3000watts with a 5Kw Sunsynk inverter and want to use a 3000w Induction Cooker... would that tax my inverter too much? I'm really not sure if the Induction cooktops also have a high startup draw... would be interested to find out if anyone here knows.

10 hours ago, Gnome said:

13 hours ago, P1000 said:

"MOFSETs" in that bridge are usually IGBTs

Still a MOSFET 😛

Err, no. The "BT" of IGBT stands for Bipolar Transistor. That's why they have an emitter and a collector, just like a 2N2222A, but of course much bigger. The base is replaced by a gate, the IG (Insulated Gate) of IGBT. But that doesn't mean it's a Field Effect Transistor. FETs and MOSFETs conduct like a resistor; power MOSFETs measure in the milli-ohms. But a bipolar transistor (conventional or insulated gate) has a saturated junction voltage drop. So they behave somewhat differently. The 0.2-0.4 V Vcesat doesn't matter too much at higher voltages being switched. So for that and other reasons, MOSFETs tend to dominate at low voltages (e.g. the battery side), and IGBTs at higher voltages (e.g. the main DC-AC converter full bridge, if there is one).

7 hours ago, Moffat said:

Now what if I have a small array, 3000watts with a 5Kw Sunsynk inverter and want to use a 3000w Induction Cooker... would that tax my inverter too much? I'm really not sure if the Induction cooktops also have a high startup draw... would be interested to find out if anyone here knows.

I am thinking it shouldn't be a problem at all, I highly doubt it will pull more than 35 amps at startup. If you put it on the non essential side of the inverter there definitely wont be a problem. 

12 hours ago, Gnome said:

Even if they were to really up the transformer size, I'm still not convinced by the design because your inductance must be mega. 

Inductance is not the problem here. In fact - in the case of Victron they have a partially wound window to increase leakage inductance because it is so low for Toroids.

6 hours ago, Nexuss said:

I am thinking it shouldn't be a problem at all, I highly doubt it will pull more than 35 amps at startup. If you put it on the non essential side of the inverter there definitely wont be a problem. 

Ideally would want to put it on the Induction Cooktop (3000w) on the Essential side of the load, am trying to get away from the high costs of gas, locally in my jurisdiction. The hope is that with a big enough battery bank should be able to cook within an hour if grid is down and not pull power from grid when it's up either... gradually going off-grid, item by item.

9 hours ago, Coulomb said:

Err, no. The "BT" of IGBT stands for Bipolar Transistor. That's why they have an emitter and a collector, just like a 2N2222A, but of course much bigger. The base is replaced by a gate, the IG (Insulated Gate) of IGBT. But that doesn't mean it's a Field Effect Transistor. FETs and MOSFETs conduct like a resistor; power MOSFETs measure in the milli-ohms. But a bipolar transistor (conventional or insulated gate) has a saturated junction voltage drop. So they behave somewhat differently. The 0.2-0.4 V Vcesat doesn't matter too much at higher voltages being switched. So for that and other reasons, MOSFETs tend to dominate at low voltages (e.g. the battery side), and IGBTs at higher voltages (e.g. the main DC-AC converter full bridge, if there is one).

I did Google this before responding.  I've not used IGBTs before but, for example 2N2222A I find painful to drive with micros because it is current driven not voltage driven. So I've opted for voltage driven when required (obv. 2N2222A is great for really low current.  Darlington pair is not really my favorite, so I avoid that tbh).

I assumed that for IGBTs you'd ideally still not want a current amplifier because it is undesirable to some extent to drive, no?  Or are you saying it is essentially a current amplifier? (OFC there is no ideal world, logic level FETs are super expensive so I avoid those too unless I REALLY need it, linear region is a bitch on the cheaper FETs)

EDIT: OFC for such a big circuit as a inverter putting drivers in for current amplifiers or even drivers that operate at higher voltage for the voltage driven gates almost certainly becomes a must.  But I'm genuinely interested in the IGBT from the perspective of, how they are driven.

Edited February 12, 20223 yr by Gnome

7 hours ago, P1000 said:

partially wound window to increase leakage inductance because it is so low for Toroids.

Why would you want to increase your inductance?  It is effectively a huge resistor in series with your circuit.  Ideally you'd want nearly "no inductance" (obv. not no, but much lower), like the grid, basically supply "infinite" current at the drop of a hat.

7 hours ago, P1000 said:

partially wound window

Partially wound window, do tell, I'm not following (or perhaps not familiar with) that at all.

Edited February 12, 20223 yr by Gnome

1 hour ago, Moffat said:

Ideally would want to put it on the Induction Cooktop (3000w) on the Essential side of the load, am trying to get away from the high costs of gas, locally in my jurisdiction. The hope is that with a big enough battery bank should be able to cook within an hour if grid is down and not pull power from grid when it's up either... gradually going off-grid, item by item.

Sorry for derailing.  Yeah 3kW cooktop will work just fine.  Any resistive load like a oven, cooktop, etc. is nearly the ideal kind of load for an inverter because it is constant and changes slowly.  The WORST kind of load is TRIAC, so for example my heat gun uses a TRIAC to constantly turn on and off a heater and that just absolutely kills my inverters ability to control the voltage (so your lights flicker slightly which is super annoying).  Another case where TRIAC is used is in my coffee maker (ie. Nespresso uses them too).  It doesn't damage the inverter but it is really annoying.  Inductive loads like air conditioner, pumps, etc. pose a bit of a startup challenge for inverters but for Sunsynk, etc. it should really not be a problem at all.  If you turn AC on and off immediately it can pose a significant challenge because the rotor is locked by compression (old school kind).  But generally it really shouldn't be a problem.

Edited February 12, 20223 yr by Gnome

On 2022/02/11 at 3:20 PM, Mahdi said:

is that true that sunsynk inverter(High frequency/Transformer-less) are not good for high load?? and should go with LF inverters(transformer)

and mofsets will burnouts..

i would like to run about 4000w for 10 hours a day..

can sunsynk handle that ? 

 

I doubt if you will ever achieve 10 hrs/day at 4000W on a 5kW inverter for the simple reason that the PV delivers power in a bell shaped curve, the max power is a function of the PV ratings and the inverter. So you may only see max production from 10am to 3pm, the rest of the day the power produced will ramp up to the max and then down again..(bell shape).

I don't see why a Sunsync, or any other inverter for that matter will not provide its designed  rated output continuously as long as its kept cool. Even then, most inverters will derate (produce less) as the internal temp rises.

 

4 minutes ago, FixAMess said:

I doubt if you will ever achieve 10 hrs/day at 4000W on a 5kW inverter for the simple reason that the PV delivers power in a bell shaped curve, the max power is a function of the PV ratings and the inverter. So you may only see max production from 10am to 3pm, the rest of the day the power produced will ramp up to the max and then down again..(bell shape).

I don't see why a Sunsync, or any other inverter for that matter will not provide its designed  rated output continuously as long as its kept cool. Even then, most inverters will derate (produce less) as the internal temp rises.

 

i have a 30k lifepo4 batteries.. and i can use grid to charge them

 

Ok, the answer is the same, I'm sure any inverter can run at its rated capacity for any length of time within its rated life expectancy.

BUT your 30kW battery bank will not last 10hrs at a load of 4kW,  at best, you can probably draw down to 90% DoD, depending on your batteries, which is 27kW which is 2700W for 10 hrs, at 4000W draw, you will get 6-7 hrs....

1 hour ago, FixAMess said:

Ok, the answer is the same, I'm sure any inverter can run at its rated capacity for any length of time within its rated life expectancy.

BUT your 30kW battery bank will not last 10hrs at a load of 4kW,  at best, you can probably draw down to 90% DoD, depending on your batteries, which is 27kW which is 2700W for 10 hrs, at 4000W draw, you will get 6-7 hrs....

Let me weigh-in with my 2-cents here: I also would love to know at what ambient temperature(s) you expect to be running this load. I would hazard to say no matter how good or efficient the inverter is, if temperatures going to the load are coming either from Pv or from Battery should cause heat, non is good for the inverter and thus derating may occur or shorter hours due to inefficiency.

11 hours ago, Gnome said:

I agree that the IGBT has some characteristics of a MOSFET. But they are sufficiently different to be in a class all their own.

Quote

I assumed that for IGBTs you'd ideally still not want a current amplifier because it is undesirable to some extent to drive, no? 

When you switch them continuously, you need a gate driver (just like a MOSFET), because of the large capacitance of the gate. You need to put in or pull out amps of current for a short time (nanoseconds). The bigger ones want fairly significant gate voltage swings (like +15 V and -5 V), so you can't just drive those directly from a microcontroller.

Quote

Or are you saying it is essentially a current amplifier? 

Well, it's an insulated gate, so apart from charging and discharging the capacitance, there is no current, so current amplification is essentially infinite.

Quote

But I'm genuinely interested in the IGBT from the perspective of, how they are driven.

Exactly the same as MOSFETs. In fact, the buck transistor in Axperts is optionally a MOSFET or an IGBT, with I believe no changes to the driver circuit at all.

On 2022/02/12 at 10:34 PM, Gnome said:

Why would you want to increase your inductance? 

[ Edit: 13 hours after writing the below, I'm not so sure I have this correct. Transformers are tricky things. ]

When you drive these transformers from a full bridge or the like, you need a certain minimum inductance so that the current doesn't ramp up too fast. Too fast and you can't control the average and maximum currents safely. It's the same with electric motor controllers. I remember an EV conversion driving an ironless motor that needed large, heavy and expensive inductances in series with each phase between the controller and the motor.

I've seen it described once a bit like this. Imagine you're a gremlin that works inside a motor controller, in charge of turning on and off one MOSFET. At regular times, the boss yells "CLOCK!" and we all switch on our MOSFETs. There is another guy yelling out the current: "100A, ... 200 A, ... 300 A" and there's our limit. We turn off the MOSFETs. We get a bit of a rest now until the next clock cycle. That's with adequate inductance.

Now the same situation with too little current: "CLOCK!" We turn on our MOSFETs. But the guy yelling the current is yelling very fast: "500A! 1000A! BANG!" before the gremlins have a chance to turn the MOSFETs off.

No offense is implied by the childish example, it's just one that stuck in my mind. Possibly from someone famous on the Electric Vehicle Discussion List. Maybe Lee Hart.

[ Edit: I'm pretty sure that the explanation above is valid for motor controllers, but I'm now wondering whether it applies to transformers as well. The inductance on the secondary would be reflected back to the primary. And you always have an inductance for smoothing out the PWM pulses. ]

Edited February 13, 20223 yr by Coulomb

You need inductance, like @Coulomb said above, to filter the high frequency switching of from the MOSFETs. Yes, in "low frequency" designs, the bridge typically switches at 24kHz, so you want to filter that. The bridge, along with the partially wound toroid and an output capacitor acts like a buck converter in that configuration (with a couple of exceptions). In some designs, like Studer, you have a fully wound toroid, but with big inductors between it and the bridge. The other important role of inductance is - in order to act as a bi-directional inverter, you have to be able to boost the output of the transformer in order to charge the batteries.

Partially wound window refers to the fact that they only wind the low voltage winding to half the toroid.

23 hours ago, P1000 said:

You need inductance, like @Coulomb said above, to filter the high frequency switching of from the MOSFETs.

Yes, but that inductance (I may be wrong here) is usually separate and on the load side of the transformer. [ Edit: I'm not familiar with low frequency transformer designs. ] I'm now thinking that this inductance will transform to the MOSFET side. In other words, I'm now more siding with @Gnome and wondering why you would want to add leakage inductance to a transformer. Perhaps when this is done, there is no separate inductance for the LC filter. @P1000, you mention big inductors on the bridge side. Do they also have an inductor on the load side?

Edited February 14, 20223 yr by Coulomb

14 hours ago, Coulomb said:

Yes, but that inductance (I may be wrong here) is usually separate and on the load side of the transformer. I'm now thinking that this inductance will transform to the MOSFET side. In other words, I'm now more siding with @Gnome and wondering why you would want to add leakage inductance to a transformer. Perhaps when this is done, there is no separate inductance for the LC filter. @P1000, you mention big inductors on the bridge side. Do they also have an inductor on the load side?

On the load side only common mode chokes. The inductor is always on the LV side, filter cap on HV (for low frequency). If you don't do it that way, and try to couple HF through a LF transformer with low leakage inductance (as a normal toroid), you will have very high losses. Also bear in mind that the inductance we are talking about here has very little effect on 50Hz loads. And higher frequency components are handled by the capacitor. In the case of Victron, they increase the leakage inductance of the transformer (at the cost of copper loss) and don't have additional inductors.