Connected two 48V chargers in parallel, internal spark

[Jalle19]

I would like to connect three Sunpower SP-750-48 (rebranded Mean Well) chargers in parallel (ultimately the idea is to build a 2S3P three-phase charger for an EV, but that's beside the point right now).

When experimenting, I connected two chargers in parallel on the DC side, then connected one of the chargers to AC. Both chargers turned on, but after around two seconds, a big spark happened inside the charger that wasn't connected to AC. After disconnecting everything and inspecting the sparked charger, no visual damage was found, and in fact the charger still worked normally. I didn't dare to try again.

What could be the reason behind this sparking? Would it have been better to connect both chargers to AC first, and only then completing the DC circuit to the battery?

FWIW the BMS was configured to not allow charging when I turned on the chargers. My only theory so far is that the first charger (the one connected to AC) started feeding DC into the second charger, overwhelming it somehow. The chargers have diodes in them though, so AFAIK it shouldn't be a problem to parallel connect them.

Any advice appreciated.

 

[Browse]

Likely a capacitor discharge of something like that. You should be able to parallel outputs of chargers, but the DC side should be fully discharged before connecting anything.

 

[hwy17]

More details about the charger? Model number?

 

[Jalle19]

Likely a capacitor discharge of something like that. You should be able to parallel outputs of chargers, but the DC side should be fully discharged before connecting anything.

It's this one: https://www.meanwelldirect.co.uk/wp-content/uploads/sites/55/2018/04/r401r401_3.pdf I updated the first post to include the full make and model

but the DC side should be fully discharged before connecting anything.

What do you mean by "fully discharged"?

 

[Browse]

It's this one: https://www.meanwelldirect.co.uk/wp-content/uploads/sites/55/2018/04/r401r401_3.pdf I updated the first post to include the full make and model



What do you mean by "fully discharged"?

I know that my Growatt inverter retains voltage on the busbars of my battery rack (after the batteries are turned off) until the capacitors bleed off. It can take an hour or more to see the voltage go back down near zero. I'm assuming most chargers keep some voltage on the DC side for a bit but I could be wrong. You can check this with a voltmeter.

 

[DJSmiley]

In the datasheets I can't find anything on supporting parallel (and thus allowing backfeeding which basicly happens when a single psu is powered on)

Also, as far as I can find, they are power supplies, not chargers. And not all power supplies can handle backfeeding.

 

[Texas-Mark]

In the datasheets I can't find anything on supporting parallel

Why would you? Having two in parallel is really no different than having one connected to a battery. However, whether or not a charger likes to be connected to a battery (or another charger) when not plugged into AC is another story.

I have some PowerMax charger/converters that do strange things when connected to a battery but not AC (like the fan runs when it does not need to even though it's not even connected to AC). But the units still work fine.

Now, I can see some strange things happening if two chargers are connected in parallel, but not connected to a battery. The voltages they would sense from each other might cause them to do strange things. But as long as there is a battery load, I don't see an issue (unless the battery BMS opens and the previous scenario happens).

 

[Browse]

That model doesn't look like it has active current sharing. Some of their models do and will list it on the data sheet:

https://www.meanwell.com/Upload/PDF/RSP-1000/RSP-1000-SPEC.PDF

 

[Jalle19]

Now, I can see some strange things happening if two chargers are connected in parallel, but not connected to a battery. The voltages they would sense from each other might cause them to do strange things. But as long as there is a battery load, I don't see an issue (unless the battery BMS opens and the previous scenario happens).

Okay, so next time I try something I should make sure the chargers have somewhere to put the power they put out (i.e. ensure the BMS is accepting charge)?

I have some PowerMax charger/converters that do strange things when connected to a battery but not AC (like the fan runs when it does not need to even though it's not even connected to AC). But the units still work fine.

I have done this (connecting a battery to the DC terminals on the charger, without the charger being plugged into AC), it produces a fairly big spark at the terminal due to inrush current, then the fan turns on and the LED on the charger turns green. Doesn't harm the device as far as I can tell, which is why I was surprised a second charger (acting as a battery in a sense) could damage a second charger.

@Browse according to https://www.meanwelldirect.co.uk/glossary/what-is-current-sharing/ , active current sharing is a fancy feature that I don't particularly need (the "brute-force" way works just fine).

 

[Jalle19]

Anyway, is the general consensus that everything should work as long as both chargers are connected to AC before connecting the DC side to a battery?

 

[hwy17]

The block diagram for these is a bit different than the Mean Well supplies I have used like these so I don't know.

I think it's slightly concerning that they don't call it a constant current supply with a constant current output graph specification.

 

[garetwo]

Anyway, is the general consensus that everything should work as long as both chargers are connected to AC before connecting the DC side to a battery?

Was this ever answered? Can we have multiple varying make/model/wattage chargers in parallel?

 

[Jalle19]

I've been telling myself that the charger that blew was defective anyway and would have done the same thing even if I had connected it exclusively to the battery. I'll be installing two telecom rectifiers in parallel sometime this winter so I can make an update to this post then.

 

[zanydroid]

Was this ever answered? Can we have multiple varying make/model/wattage chargers in parallel?

It’s probably 80% presumed safe in the solar world (where prior to AIOs systems were aggressively integrated from separate components via DC coupling). Until someone gets burned. Like in situations where people generalize to things other than solar inverter/charger and SCCs

Telco and data center rectifiers are often designed as components for active active or hot swap redundancy so they’re likely more likely to work than other random stuff

 

[garetwo]

It’s probably 80% presumed safe in the solar world (where prior to AIOs systems were aggressively integrated from separate components via DC coupling). Until someone gets burned. Like in situations where people generalize to things other than solar inverter/charger and SCCs

Sadly, as an electrician, we tend to generalize, its why I ask stuff in here. You guys not only know these particular scenarios, but can often provide an answer in electrical layperson words. The 80% is not good looking from where I sit. That may well translate to 100% for those who know what they are doing, averaged with X% for those who do not. I hope I am only near the 80 because I assume the worst and would not do it.

Telco and data center rectifiers are often designed as components for active active or hot swap redundancy so they’re likely more likely to work than other random stuff

Totally above the folds in my brain. Would a component data sheet say "designed for hot swap redundancy?" Is "active" redundancy meaning they are all on line at the same time? Could active redundancy be the equivalent of a manufacturer stating it may be paralleled with other chargers? Can a power supply be a charger (or vice a versa) ? I can probably find the answer to that last question elsewhere.

 

[garetwo]

I've been telling myself that the charger that blew was defective anyway and would have done the same thing even if I had connected it exclusively to the battery. I'll be installing two telecom rectifiers in parallel sometime this winter so I can make an update to this post then.

Is there a point where one must parallel because there is no singular solution? I clearly see with inverters that their charging function is paralleled when other of the same inverters are connected and communicating with one another. I just assumed they then are perfectly (greater than 80%) fine getting battery banks to properly charged levels without causing issues for the batteries or themselves.

 

[zanydroid]

Did you read this pinned post yet?

Thread 'Charging from multiple sources'

Mar 6, 2021

One can have as many independent charging sources running concurrently as they like provided they are programmed properly for the voltage of one's battery bank.

Examples (some or all at once, even multiples of a single type, different brands, etc.):

1. Solar PV via MPPT
2. Solar PV via PWM
3. AC-DC input via MPPT (using a AC to DC power supply feeding an MPPT solar controller with voltage/current consistent with a solar panel)
4. Wind with suitable controller
5. Grid/Generator AC via inverter/charger
6. Grid/Generator AC via RV AC-DC converter
7. Hydro with suitable...

• snoobler
• Replies: 62
• Forum: General Battery Discussion

• 

That may well translate to 100% for those who know what they are doing, averaged with X% for those who do not. I hope I am only near the 80 because I assume the worst and would not do it.

I mentioned it primarily because people assume everything can be DC coupled against the battery, but I'm a little reluctant to go crazy with mixing and matching. For instance, discrete SCCs were paralleled for a decades with each other and with inverter/charger. DC coupling/paralleling is also easier than AC coupling/paralleling.

So people sort of assume AIOs can be used this way (which have inverter/charger and SCC in them, albeit in a different power topology unless it's literally just a bunch of pre-existing boxes bolted together, which Victron does but nobody else does). The Chargeverter is marketed assuming this is the case.

But, why would an AIO not designed for parallel stacking, necessarily be safe against another charger being in the mix.

(Fortunately most AIOs and hybrids support parallel stacking on a shared battery bank)

Totally above the folds in my brain. Would a component data sheet say "designed for hot swap redundancy?"

I know if you look at Meanwell datasheets/application notes you can figure it out directly from what they write.

Is "active" redundancy meaning they are all on line at the same time? Could active redundancy be the equivalent of a manufacturer stating it may be paralleled with other chargers?

To me it is a very good sign. Note, I'm piecing together this conclusion from knowing stuff about Datacenter equipment from going on tours in them as a software engineer (who likes looking at these, but it's different from an engineer designing or living in the datacenter).
Active redundancy to me means they have the right protection to not get killed by another power supply. So either passive or active rectifiers, and whatever else is called for. I think they need to be spooled up already to minimize disruption to the server, rather than reactively starting up. If a PSU is marketed as hot swap, I should be able to pull out one of the two without a problem.

Now, active redundancy does not require them to share current evenly when both are up. In this case, it is more of a nice-to-have, since it puts thermal wear on the power supplies more evenly, than if one was always carrying 100%; the backup power supply can also do some more sanity checks (IE, is power being drawn through me? if not something weird is happening).

Can a power supply be a charger (or vice a versa) ? I can probably find the answer to that last question elsewhere.

A power supply can be a charger, but it kind of gets into issues with English.

Chargers have extra logic like the charging curve config. A pure power supply would only have max current and max voltage (which could be fine). A charger may have battery detection logic, in which case it cannot be used as a power supply without a battery.

Plenty of people use benchtop power supplies to charge & balance batteries when assembling them.

In some contexts (USB chargers) there are those that argue that they're only rated for 2 hours of full load operation, since by then the portable device / battery is fully charged. While power supplies, people probably use a 80% load rule if expected to run for a long time.

 

[garetwo]

but it's different from an engineer designing or living in the datacenter).

Ha! I know many who "live" at what they do...See, I do read everything! Thank you for the in-depth comments

 

[garetwo]

I think they need to be spooled up already to minimize disruption to the server, rather than reactively starting up

How would you "spool it up" and then connect to the load?

 

[garetwo]

Chargers have extra logic like the charging curve config. A pure power supply would only have max current and max voltage (which could be fine). A charger may have battery detection logic, in which case it cannot be used as a power supply without a battery.

Might it construe another charger as a battery?

So a good charger would not necessarily have built in programming for various battery types, via a selector switch, but provide the means for the end-user to configure voltage settings (sort of a question)?. Either way, if one charger's built in settings do not exactly match the settings of any it is paralleled with, no big deal? If it is sort of a deal, I can see an argument for same make, model, etc. because even if they may be programmed exactly the same, there is a degree of deviation each may have.

I think I am just rambling now and putting immediate thoughts here. I saw someplace where they said a charger's fan ran while not in charging mode but was connected to the bus, it seems it minimally implies wasting energy.

 

[GXMnow]

Was this ever answered? Can we have multiple varying make/model/wattage chargers in parallel?

Chargers are a bit different from power supplies.

As mentioned in this thread, there are "Data Center" power supplies that are meant for redundant connection. These typically have an isolation diode at the output so only the supply producing the most voltage is actually powering the load. Even a 0.1 volt difference can cause one supply to be taking all the load while the other(s) are just sitting idle. But if the loaded supply fails, the next highest voltage one then takes the load. This works just fine to keep a server powered up.

Charging is quite different. We typically want a current source with a voltage limit. As long as the battery is below th voltage limit, then the currents of all chargers will add up and help charge the battery. And since the battery WILL have voltage on it's terminals, the charger has to be able to handle voltage on it's outputs. So back feed is not an issue.

A power supply that is nit designed to have back fed voltage could certainly be destroyed with back fed power. Think about a pass transistor regulator circuit. An NPN transistor would have a higher positive voltage on it's collector, with the emitter going to the output. If the supply is turned off, you now have the higher voltage on the emitter with the collector at a low voltage. This could fry it. There are many different output circuits for power supplies, and many of them will not like being back fed.

 

[zanydroid]

How would you "spool it up" and then connect to the load?

I don’t know, hypothetically one could consider putting it on cold standby instead of with the SMPS already started. It would probably interrupt power for 100s of milliseconds.

 

[zanydroid]

Might it construe another charger as a battery?

It might but a charger behaves pretty differently. It will not accept power on its output without smoking

So a good charger would not necessarily have built in programming for various battery types, via a selector switch, but provide the means for the end-user to configure voltage settings (sort of a question)?.

Search for the Chargenectifier thread here, it will cover the programming of a data center rectifier.

The charger inside the Chargeverter is this type of rectifier, except the programming is simplified into front panel settings

Most AIO inverter/charger/SCC have customizable target voltage and current

Either way, if one charger's built in settings do not exactly match the settings of any it is paralleled with, no big deal?

If they have input protection, I think the main risk is that some of the chargers don’t deliver power.

Matching chargers within a brand/family helps with avoiding having a hodgepodge of heterogeneous crap to have to understand

 

[GXMnow]

Matching chargers within a brand/family helps with avoiding having a hodgepodge of heterogeneous crap to have to understand

Most "Chargers" will work together just fine on Lead Acid and Lithium batteries. They both do Constant Current Bulk charge. Parallel a 20 amp charger and a 30 amp charger and you are charging at 50 amps. The only time this changes is when the charge voltage comes close to full. I currently have my Schneider XW-Pro AND a Victron MPPT feeding charge current to my battery bank. They are very different devices. The Schneider charges at the current set in the "Maximum Charge Current" register. In my case, that is adjusted remotely and can vary from 7 amps to about 60 amps. And if another load comes on and is pulling too much current, the XW can lower the current so it does not exceed the current draw of the AC1 grid feed breaker setting. The Victron MPPT can feed up to 35 amps, but it is based on how much sunlight is hitting the DC solar array. It can range from nothing to 35 amps and change very quickly when the amount of sun changes.

Both devices fully expect battery voltage to be on their output leads at all times, even if they are not charging. And both devices can handle up to 60 volts without a problem, even if the charge voltage is set lower, it won't hurt them.

If the Victron DC charge controller sees the battery becoming full, it has no choice but to throw away the extra energy. So I intentionally have the charge voltages set different for the two devices. The Schneider XW-Pro is pulling the charge power from the AC system of my home. In most cases, that is actually extra power coming from my Enphase microinverters. I have a controller that monitors power exporting to the grid and it then commands the XW-Pro to go into charge mode and take most of that energy and store it in the battery. But when it sees the battery becoming full, it just uses less of the AC power, and more of it will export to the grid and I will get credit for it. So I have it set for a Bulk charge voltage of 56.7 volts. On good sunny days, it will hit that before 2 pm and start exporting to the grid. But the batteries can still hold more and the Victron DC charge controller is set for a Bulk charge voltage of 57.6 volts. So instead of throwing away more power, it will keep bulk charging for about 2 more hours and pushing all that energy into the battery. The WX-Pro reports the voltage still climbing even though it is not charging.

With two different chargers, they should also act like this. Even if their Bulk voltage is set the same, it is very possible that one of them will switch from Bulk to Absorb before the other one. That is not a problem and it will work just fine. It may take another few millivolts, or in my case nearly a full volt, but the other charger will eventually go to Absorb as well. And they will then also switch to Float, if set up for that, and again, one may switch before the other. It's not a problem, it will work just fine.

But again, this is CURRENT source chargers, not power supplies. On a quality power supply, even a few millivolts difference could cause one supply to provide all of the current. Adding a small amount of series resistance can help the current sharing, but it will still be very picky. If the supplies have a voltage trimmer, you can try to dial them in to match, but they could drift enough to make the current sharing swing a lot between the two or more supplies.

When we use a Bench Power Supply to charge batteries, they should have a CC CV setting where you dial in the voltage and current limits. They they act more like a charger. First set the CV voltage to the maximum Bulk voltage you want. Then dial in the CC current into a heavy load. The voltage will drop below the CV voltage and the supply should then be just running at the set CC current. The current will drop off as the voltage comes close to the CV voltage. If you are using 2 supplies like this, yes, one supply is likely to drop current before the other. Even millivolts of difference can cause a huge current different.