Confused About AC vs. DC Coupling

[digitalabacus]

Hello all,

First time posting here, been lurking and learning.

I am a bit confused about when AC vs. DC coupling is better for a given setup. From what I have read so far, this coupling refers to the mode of energy ingress to the primary inverter from a battery backup or similar, either AC or DC. Makes sense so far. (side note: does anybody know what is being coupled in this terminology? Battery to primary inverter, battery to grid, grid to household via battery, etc?).

What I can't wrap my head around is why you would prefer one to another. AC holds the advantage for power transmission efficiency, sure, but trying to sync up transfer between AC sources sounds like a real bear when considering things like glitching and phase differences between the two AC waveforms. Given that just about every picture I have seen (a comparatively small sample size, sure) each has all of the various system components mounted right next to each other, it seems that transmission losses are minimal anyway. Why wouldn't you just DC couple and save the cost of the extra inverter?

This confusion is further compounded by items like the Sol-Ark 15k brochure (link) which shows both a DC battery input and a 24kW AC coupled input. My gut is telling me that this product is all about options so you can choose either or perhaps wire up the AC output of a generator or something.

So, in short, why would a given installation choose DC coupling vs. AC coupling?

 

[wattmatters]

So, in short, why would a given installation choose DC coupling vs. AC coupling?

There are probably a multitude of reasons no doubt.

In Australia (where vast majority of solar PV systems are grid tied), then the choice of an AC or DC coupled system can often be due to local regulatory restrictions on the size of inverter which is permitted to be connected to the grid.

e.g. in some areas (QLD for instance) there is a 10 kVA limit on inverter capacity which may be connected per phase, and the total inverter capacity includes any battery inverter.

So for instance if you had an existing 8 kW solar PV inverter then you would not be permitted to instal a Tesla Powerwall 2 as it has its own 5 kW inverter. You'd be limited to an AC coupled battery system with a maximum inverter capacity of 2 kW.

But with a hybrid DC coupled system you can have a 10 kVA inverter which manages both the solar PV and the battery in one, enabling the battery to supply high power demands and recharge more quickly as well.

Such regulatory limitations are not the same across Australia, in many areas battery inverter capacity is not included in the grid connection limits for small scale generation systems.

AC coupled batteries are convenient in a grid tied scenario as you can install a solar PV system now (which here are very financially beneficial) and add a battery later when/if it makes financial sense (they don't). Some have tried to "future proof" by installing a hybrid inverter in the expectation they will add a battery later. Unfortunately the experience is that after a couple of years there are no more compatible batteries available for that older hybrid inverter model and so the extra up front expense of the hybrid was not worth it.

 

[digitalabacus]

I see. So, if I am understanding correctly, it seems like the regulatory concern is mostly about ability to backfeed the grid? Does this imply that, in an AC coupled system, the AC coming off of the panel inverter(s) can "see" the grid AC and vice-versa, hence the regulatory regulatory involvement? And, if so, how is AC waveform distortion prevented? If I understand correctly, much effort is spent on "syncing" up the national electrical grid so all of the generation is in-phase (which is why it takes time to "restart" the electrical grid after a regional outage).

For my specific case, I am working backwards. I work from home as an engineer and, unfortunately, power outages are becoming more commonplace. Obviously, this will not work for business so my plan is to install an inverter and battery backup now and then add panels a bit later. Should satisfy the immediate need of not losing power while leaving options for installing panels a few months later.

 

[SparkyJJO]

I'm doing AC coupling because I already have grid tie microinverters. If I was DIYing my whole system, I would not do AC coupling.

 

[wattmatters]

So, if I am understanding correctly, it seems like the regulatory concern is mostly about ability to backfeed the grid?

Both DC and AC coupled system can feed the grid, if permitted.

To me the only difference between them is that AC coupled just means the battery has its own inverter. But happy to learn what others think.

Any inverter which is grid-tied has to go through a process of syncing to the grid's frequency. They are grid following devices. It's a normal part of the start up cycle. They will go through a process to monitor the grid and adjust their internal frequency until it matches and then it will connect.

When you are off grid (be it permanently or temporarily), then one of your inverters becomes the grid forming supply source, and anything else using or connecting to the AC supply has to sync to it.

For example, in grid tied set up with a solar PV system and a Tesla Powerwall 2 (which is an AC coupled battery), when the grid goes off the Powerwall's Gateway and inverter immediately takes over supplying power to the home and it becomes the grid forming supply source. The solar PV inverter then follows the grid forming inverter. The Powerwall then uses frequency control to manage the solar PV inverter's output.

When grid power returns, the Powerwall will take a bit of time to adjust its frequency so that it matches the grid before switching back to grid connection.

In a DC coupled system, the battery shares the same inverter as the PV array and grid connection, if any.

 

[digitalabacus]

I'm doing AC coupling because I already have grid tie microinverters. If I was DIYing my whole system, I would not do AC coupling.

Gotchya, so this is a particularly beneficial option if you already have existing infrastructure / a microgrid setup / a Tesla powerwall / etc. which already produces AC.

Any inverter which is grid-tied has to go through a process of syncing to the grid's frequency. They are grid following devices. It's a normal part of the start up cycle. They will go through a process to monitor the grid and adjust their internal frequency until it matches and then it will connect.

When you are off grid (be it permanently or temporarily), then one of your inverters becomes the grid forming supply source, and anything else using or connecting to the AC supply has to sync to it.

I see, I hadn't even considered that these would be "smart" on the AC following side. In hindsight, that makes obvious sense.

Thank you both for the feedback; this makes much more sense now. It seems like, for my particular use-case, the fact that I already have battery capacity in mind and am planning to use a monolithic hybrid inverter solution "from scratch" (e.g. minimal preexisting infrastructure) indicates that I should just use DC coupling (pending review of applicable capacity regulations, as described by @wattmatters).

 

[wattmatters]

Gotchya, so this is a particularly beneficial option if you already have existing infrastructure / a microgrid setup / a Tesla powerwall / etc. which already produces AC.

Yes. This is very often the case in Australia, where 25-30% of homes have a grid tied solar PV system but only a tiny fraction have a battery.

Of those much smaller number of homes with a battery, the most popular battery choice is a Tesla Powerwall 2. Here it is available as an independent retrofitted device. Expensive though, typically ~A$15,000 installed for 13.5 kWh of capacity, but that's about a typical cost for such battery installations here.

 

[toms]

Reasons why i use AC coupling (20kwh off grid)

- inexpensive second hand PV and GTI availability
- compatibility with batteries
- simplicity of system setup

I also have a 10kwh off grid system that is solely DC coupled, for different reasons. It really is dependent on your individual case.

 

[Samnarly]

Hey, I'm wanting to power up my house with an off grid system. Victron 10kva inverter and some eve 280ah batteries.
I have 10kw in panels connected to a 8kva growwatt inverter feeding 5kva back to grid. Can I AC couple my on grid inverter to my victron, to charge batteries etc? If I can how do I do it?

 

[Judoal]

I have a Sol-Ark 12K system. Because of a Voc mismatch between one array and my other 3, I am using an Outback FM60-150 CC to DC coupled the array to the the inverter. Setting the Time-of-Day to daytime hours, the discharge voltage is set to just below float level. This way only power from the array is used and when there is limited/no sun, my batteries won't be drawn down. The battery bank is limited to one of two strings because of a bad battery, so capacity is limited.

 

[RCinFLA]

DC coupling puts a lot of ripple current on batteries when back feeding single phase AC grid. The batteries are effectively large filter capacitors for the DC charge controller.

Conversion losses are usually greater for DC coupling for grid exporting.

 

[robbob2112]

So, I am reading a lot about this topic and there is one thing that still confuses me. The Schneider conext XM pro goes out of it's way to point out that it is not an 'on grid' inverter. From the discussion here I would have thought is was a DC coupled inverter. Instead it talks about increasing the frequency of the ac output to 62hz to trip the 'on grid' inverter offline. From other things in the manual it makes it sound like the ac out is hooked to this other inverter. Then the discussion goes on to talk about this other inverter like it is sacrificed in the event the grid power fluctuates repeatedly in a very short time.

But it seems like everyone is just using AIO inverters and either tying into a breaker in the main panel or in between the grid and the AIO.

Can someone please explain the distinction?

@Hedges
@sunshine_eggo

 

[Hedges]

Apparently HF HV AIO have more trouble with AC coupling, can't change between charging battery and inverting from battery as quickly as LF inverter. Connecting GT PV to "generator" port which has relay lets it be disconnected.

AC coupling between Sunny Boy and Sunny Island seems to work fine. I use AGM. Some people use lithium, which needs enough charge headroom to accept a couple seconds of load dump.

When GT PV does frequency-watts, the battery inverter ramps up frequency over a couple seconds, and GT PV inverter ramps down power. I think that is from 60.5 Hz to 61 Hz under Rule-21, or 61 Hz to 62 Hz for Sunny Boy with offgrid settings.

Schneider specifically, I don't know about. Other people here have some model of Schneider AC coupled, but some issues getting it to charge battery the way they want.

With Sunny Island, the issue is more how to export from DC coupled PV to grid the way you want.

 

[RCinFLA]

If you are primarily off-grid operation, use DC coupling to avoid the difficulties with dealing with sudden over-production for AC coupling.

If primarily on-grid operation use AC coupling. Gives better efficiency and less battery stress than DC coupling.

HF AIO inverters are effectively AC coupling with direct wired hardware control to control over-production situations. I am just not a fan of HF inverters because of their 'delicacy'.

My idea all-in-one inverter would be a low freq inverter with hardware control of a parallel GT inverter to solve the sudden overproduction issue.

Present LF AIO products on the market are DC coupled with direct hardware control on charge controller that feeds battery terminals of inverter directly. You have higher charge controller losses bucking down to battery voltage, plus higher conversion losses in LF inverter boosting back up to AC output.

Total path efficiency for AC coupling is routinely 95-98% range.
Total path efficiency for DC coupling is 87-92% range.

Best PV conversion to AC output efficiency is for HF inverter with PV array voltage very close to inverter's HV DC bus voltage (HV DC bus voltage usually sets the maximum PV Voc input voltage due to DC to DC boost only SCC). Because the boost charge controller has to do little voltage boost conversion when PV array voltage is high, the SCC efficiency is about 99%. Then all there is to do is IGBT H-bridge PWM sinewave chopping and filtering of HV DC which has about 97-99% HV DC to AC output efficiency, resulting in overall efficiency of 96-98%. Efficiency drops a bit for lower PV array voltage requiring greater SCC DC to DC converter voltage boosting to HV DC bus voltage.

This is how most GT inverters and HF AIO inverters work. There are a few GT inverters that require very high PV voltage, always above inverter's HV DC bus, that have buck DC to DC converters for SCC to HV DC bus. They have the best conversion efficiency.

 

[CRenner]

So, I am reading a lot about this topic and there is one thing that still confuses me. The Schneider conext XM pro goes out of it's way to point out that it is not an 'on grid' inverter. From the discussion here I would have thought is was a DC coupled inverter. Instead it talks about increasing the frequency of the ac output to 62hz to trip the 'on grid' inverter offline. From other things in the manual it makes it sound like the ac out is hooked to this other inverter. Then the discussion goes on to talk about this other inverter like it is sacrificed in the event the grid power fluctuates repeatedly in a very short time.

But it seems like everyone is just using AIO inverters and either tying into a breaker in the main panel or in between the grid and the AIO.

Can someone please explain the distinction?

@Hedges
@sunshine_eggo

Before I found this site, I had an AC Coupled system installed. It has micro inverters that output 100%.
When off-grid, you add an AIO that can control the "other inverter" (micro inverters) by changing the frequency, up or down.
When power is not needed shift frequency up, when needed down to 60.

 

[400bird]

So, I am reading a lot about this topic and there is one thing that still confuses me. The Schneider conext XM pro goes out of it's way to point out that it is not an 'on grid' inverter.

Really? The XW has lots of setting related to grid input and it is compliant to many of the grid regulations.

From the discussion here I would have thought is was a DC coupled inverter.

My interpretation is that it's in hopes of explaining AC vs DC coupling and the benefits/reasons to DC couple.

Instead it talks about increasing the frequency of the ac output to 62hz to trip the 'on grid' inverter offline. From other things in the manual it makes it sound like the ac out is hooked to this other inverter. Then the discussion goes on to talk about this other inverter like it is sacrificed in the event the grid power fluctuates repeatedly in a very short time.

But it seems like everyone is just using AIO inverters and either tying into a breaker in the main panel or in between the grid and the AIO.

People are lazy and cheap, AIOs fill both those needs. They are easier to install and mostly engineered overseas and imported, hence lower cost.
If not cheap, they are often huge (15k) at reasonable cost, but still engineered and built overseas.

Can someone please explain the distinction?

@Hedges
@sunshine_eggo

 

[robbob2112]

Watching this video at 19 minutes they are talking about the BCS which is connected to the load and is basically a 200amp auto transfer switch "to keep from overloading the 60 amp relays" ... and their diagram is of directly connected to the grid.

When they are pointing out it not being an "on grid" they are no saying it can't be grid tied, but they seem to be saying they aren't meant to be directly connected to the grid before the main panel. All the diagrams show it being fed into breakers unless the BCS is used. But even then it isn't like the Sol-Ark and EG4 which connect to the grid feed on one connection and the load feed on another... so it requires the BCS to combine into a 200amp relay.

Prior to this I was thinking you could just split the grid feed into 4 (for a 4 XW install) using a small box and 60 amp breakers and wires to each XW. Makes me wonder if that would work and each XW would switch their own 60amp relays when grid power is lost.

So are they trying to sell the BCS or is it required for grid tie with more than 60 amp service?

 

[Tulex]

One thing to consider, at least with the Sol-Ark 15k, is that AC coupling overage will go to grid, you can't not sell it.

 

[MajicDiver]

I am off grid with 2 inverters, one “ac coupled” to the other. These are Growatt MIN-11400TL-XH-US inverters that are capable of only 10kW ac output when only on battery input. On very infrequent occasions our load will go above 10kW after sunset - when we have people over and are running ovens, microwaves, heat or ac and other things at the same time. I do not want to have to tell people to be careful what they turn on at those times. Thus I buIlt our system to be capable of sustaining up to 20kW at nighttime but only rarely exceed 11-12kW and have never gotten to 13.
In this system, and I assume all NERC 1741SA compliant systems, when ac coupled the grid forming inverter raises/lowers its output frequency if its battery is full to control the output of the ac coupled inverter. For my inverters, once battery is full if load on the grid forming inverter is above about 800 watts it keeps frequency at 60hz. As it’s load drops it slowly increases frequency and the load on the ac coupled inverter lowers accordingly. This continues such that at about 400 watts on the grid forming inverter it’s output frequency is 63.1 hz - the point at which the output of the ac coupled inverter trips if it is set according to the 1741 requirements in this area, southeast US. If the load goes up, frequency drops and the ac coupled inverter will re-sync and help share the load. There are a number of settings that can be used to control how quickly these things occur if you want but it takes a lot of understanding if you want to deviate from 1741 and make it work without hiccups.

All of this works well even with a 4.5-5 kW load cycling on/off with batteries full.

Another advantage of this arrangement is I have an installed spare inverter in a sense. If my grid forming inverter were to fail I can open two breakers and shut one and be back up without having to rewire anything and it’s simple enough my wife can do it without understanding what an inverter is.

 

[robbob2112]

More to read - good thing I am still a year out from designing things....

My sysem will be grid tied and zero/minimal export - the rural electric coop will only allow excess generation of %20 and that is banked as KW and then used later to offset. Once a year true up at wholesale rates that can't exceed $500.

Like you I plan to have a system that can do everything at once even if it is at night. I figure the grid tie is more cost effective than a much larger battery bank.

I also plan to have a notebook with a first page for switching over and the second page for startup and shutdown. Following pages with detailed diagrams and drawings. All laminated. And it may extend a few wires or take a few contactors but I would like to do it all from a single labeled control panel. Clean, simple, efficient is the goal.