DIY pseudo-PWM SCC using relay?

[fafrd]

(Spoiler alert - there is no PWM)

I’m adding a small 750W DC-coupled ‘bonus’ array to my grid-tied array and want to use relays to switch off a grid-tied Microinverter in late morning when my large grid-tied array starts getting up close to peak output (due to export power limit).

From ~noon to 4pm. I want to use that bonus array to charge a 24V LiFePO4 battery and have been planning to get an MPPT that will be switched on and off through another (larger) relay connecting output to the 24V battery.

Going over cost-benefit and knowing that my string voltage is only 2-3 V above my battery voltage, I’ve decided it makes sense to save money by getting a PWM SCC rather than an MPPT.

But when I think about the fact that I’m switching the PWM SCC on and off using a relay and reviewing how a PWM SCC works, it occurs to me that I could just directly connect the 30V string to the battery.

String voltage will be at Vbat which is 2-3V below Vmp, but since current will increase from Imp to Isc, total power loss will only be ~5% (same as it would be with a PWM SCC).

I’ll need to monitor battery voltage to turn off the relay as the battery gets charged up enough to enter CV mode (I have another string charging the battery through an MPPT so I’ll know when to disconnect the ‘bonus’ relay.

I need yo think about fault tolerance to assure that the bonus relay is always disconnected once the battery is fully-charged, but other than that, is there any other reason direct charging of an 8S / 24V LiFePO4 battery from a 60-cell / 30V PV array will not work?

Has anyone heard of anyone else performed direct charging of batteries from PV arrays like this?

If I’m only direct-charging for a few hours near midday when the battery should be in CC mode, does anyone see a problem with this that I’m missing? (I realize there is no logging / reporting and that is not important to me).

 

[fafrd]

While I’m waiting for someone to hopefully chime in, zi though I’d add a question regarding PWM SCCs: I’d appreciate confirmation that I’ve correctly understood power consumption / dissipation of a PWM SCC:

CC (bulk / boost charge mode): a PWM SCC is just a switch directly connecting PV input to charge outout at battery voltage while in CC mode, so the only power dissipation / heat-generation within the SCC is switch resistance times charge current.

In the case of a 30A charger, as long as charge current is below 30A, maximum heat generation should be below switch resistance x 30A x 30A (I^2R).

The first case of higher power consumption to consider is when the PWM SCC enters CV mode, at which point it opens the switch for some % of cycle-time (during which string voltage charges up towards Voc) then closes the switch goes the remainder of the cycle time allowing Cinput x (Vstring -Vbat) of charge to transfer through the switch into the battery.

Heat generation within the switch resistance in this situation will be higher because charge currents will be higher for at least a portion of the time, but I’m not interested to dive into a detailed analysis of that since in my application, I will pretty much never be charging in CV mode.

The second case of higher power consumption is when string current exceeds maximum current rating of the SCC. An MPPT SCC increases string voltage to reduce input current and power in this situation, while I’m not sure what a PWM will do (if anything), but this situation is also not of particular interest to me (since in my case, Isc of my string is only ~80% of PWM max charge current rating).

So to summarize, I’m looking for confirmation that power-consumption / heat dissipation of a 30A PWM SCC will actually be very modest when in CC charging mode.

The motivation for the question is reviewing very inexpensive PWM SCCs such as this:

https://www.amazon.com/EpRec-Controller-Compatible-Lithium-ion-Phosphate/dp/B07VDWTWTW/ref=asc_df_B07VDWTWTW/?tag=hyprod-20&linkCode=df0&hvadid=366282008862&hvpos=&hvnetw=g&hvrand=5315844063302396777&hvpone=&hvptwo=&hvqmt=&hvdev=m&hvdvcmdl=&hvlocint=&hvlocphy=9032083&hvtargid=pla-828783364474&psc=1&tag=&ref=&adgrpid=84691875748&hvpone=&hvptwo=&hvadid=366282008862&hvpos=&hvnetw=g&hvrand=5315844063302396777&hvqmt=&hvdev=m&hvdvcmdl=&hvlocint=&hvlocphy=9032083&hvtargid=pla-828783364474

A tiny unit like that is not going to be able to dissipate much of any heat generation.

It has a LiFePO4 setting and a maximum input power of 720W @ 24V.

The LiFePO4 setting is for ‘25.6V’ (3.2V per cell) so unclear whether that is the CC voltage limit or it charges in CC mode to a higher voltage limit than that, but the point is that 720W / 25.6V = 28.126A (meaning it could charge at as little as 24V / 3.0V pet cell before exceeding maximum current level with maximum input power).

It states: ‘-Dual MOSFET Reverse current’ so heat dissipation within a power MOSFET almost certainly dominates heat generation.

A power MOSFET has an ON resistance of ~125mOhms:



So at a maximum current of 30A, heat generation should be no more than 30^2 x 0.125 = 112.5W.

Is that a lot or a little?

[EDIT: just found this: https://www.electronics-tutorials.ws/transistor/tran_7.html

‘Power MOSFETs generally have a RDS(on) value of less than 0.01Ω which allows them to run cooler, extending their operational life span.’

With a resistance of less than 10mOhm, power consumption was thin the power MOSFET n my 30A example would be led than 9W (which is certainly low enough to be passively dissipated with a small PWM SCC box like the one I linked to…]