Charge Controller Question

So yesterday I had a disturbing moment.
I have 2 MakeSkyBlue 50A (MSB) charge controllers collecting from 2 arrays.
1 array faces easterly and the other westerly. Each array consists of 6 250W panels.
I have the MSKs set to 4.07V for absorption and 4.06V for float.

Since I am a stickler for having all my sensor readings recorded and available live and historical at any moment in time I noticed yesterday that one of my MSB units was not stopping at 4.07V but continuing to charge. My batrium was going ballistic trying to bypass at 4.08V so I immediately raised the bypass to 4.15V and continued to watch.
Here is the trace of this incident.
There is a glitch on cell 1 dropping to 3.95V at ~13:00 just after I set the bypass to 4.15V. I am attributing that to the Batrium accepting the change as I have not seen it again or have noticed any other cell issues.



Here is the Battery V in perspective.



The Temperature topped out at ~101F in the plant (Shed) and am wondering if that could have caused the MSB to go wonky.
Here is the trace of the 2 MSB amperage outputs.
As you can see Controller 2 (orange trace) went out of "control".
After about 90 min it dropped right down to nothing realizing it had been a bad controller and meandered till the V was low enough on the pack for it to charge again. Also you can see Controller 3 (purple trace) was out a little early and did not participate in this mutiny.
I do understand that the MSB controllers aren't the top of the line and I will ask for a replacement suggestion at the end.
I will keep an eye on it today as it is another hot day in Maine and see if there is a repeat of this shameful disobedience of its programing.




So this brings me to the question of recommendations.
I am not married to the MSB controllers and I bought these in the early days of my solar adventure so there was little research done except for the price.
Now I know better.
As my panels face as described initially in 2 different hemispheres although at noon both are well saturated by the sun I thought that 2 controllers would be beneficial as one would pick up after the other one was no longer at an optimum angle for good solar collection.
My question is should I just combine the 2 arrays giving me 3000W into 1 quality controller (I am thinking Victron but open to suggestions) or should I still split the arrays as they are and go with 2 controllers.

Any input is appreciated.

Wolf

From my experience they do not retain thier settings through a power cycle, so if your DC side has been disconnected at any point overnight they may have reset to defaults, which may be the "fry your pack" output level.

I have 3 units fed from 7 strings in different combinations for the same reason as you.

Also, try to avoid reliance on one point of control / failure.

completelycharged said:

From my experience they do not retain thier settings through a power cycle, so if your DC side has been disconnected at any point overnight they may have reset to defaults, which may be the "fry your pack" output level.

Click to expand...

Yes I have noticed that too. But in this case there was no interruption overnight otherwise I would have seen that with all my data collection.I know overkill.
So no changes were made yesterday to anything and today it is even warmer by 2 and both seem to be behaving as of now.

We will see if Controller 2 decides to throttle back at the appropriate time. At least I have 0.1V to play with if it goes out of control and I can drive home and do a disconnect, soon to be managed by2 ABB shunt trips on the arrays.

Controllers - I have 3 x Midnite Classic 150s, the oldest is 2 yrs and counting. Besides being rock-solid/feature rich in general (e.g. 'the real thing') I particularly like:
- Full internet (RJ45 cable) access
- Simple Modbus access to retrieve/set anything you want over the internet remotely- so simple I could do it .
- GCIF... but also Arc Fault (detects loose/arcing wiring in the array hookup)
- Aux1 (and Aux2) external relays to control things - e.g. I use this to turn on/off my inverters based on battery voltage.
- Works smoothly in unexpected ways - for example I temporarily hooked up 30 panels to 1 unit (8400 watt potential > 4200w max) last winter while waiting for the 3rd unit to arrive.... and found that when I got short bursts of sun and the panels exceeded the input - the unit continued to work smoothly and simply ignored the extra power... Amazing.

The only other controller I have experience with is a small 24v/500w so I clearly have not done any comparisons But I can get behind Midnite Classics as seem rock solid. I haven't noticed any setting issues or loss of settings or anything close. The only things I've had to do is
- Reduce output to 80a (instead of allowing 86a max) becasue my 80a circuit breakers tripped (and 4awg wire won't handle it)when the unit actually achieved 80a+ one day.
- Tweak Arc Fault sensitivityabout 10% to avoid false alarms.

OffGridInTheCity said:

Controllers - I have 3 x Midnite Classic 150s, the oldest is 2 yrs..........................................

Click to expand...

Yea I have looked at them and they do look nice. Would you get 2 for my small system? I don't mind spending money but ~1400 for 2 with only 3000W of panel. Or get 1 and keep the MSB as backup. LOL
Also if I do use only 1 how do you think it will work with my panels easterly and westerly orientation?

On another note today the MSBs worked just fine. Shrugg

Wolf

Wolf said:

I would have seen that with all my data collection.I know overkill.

Click to expand...

No overkill, not in the slightest, the more data the better as it is the outliers that sit on the edge of the data that cause all the problems.... and data that highlights the problems before the fire engine turns up...


With 48V and typical panels the economics on small systems (less than a few kW) using PWM controllers is economic, even with the loss of arund 10-20% on not tracking the MPPV. Typical 200-300W panels just need 2 in series. For me MSB was at the margin and the higher cost controllers to me kjust don't make economic sense anymore now that panels have dropped in price so much, Just buy more panels and use cheap as chips PWM controller or even a FET array switch controlled from the BMS.

---

OffGridInTheCity said:

becasue my 80a circuit breakers tripped (and 4awg wire won't handle it)

Click to expand...

4AWG is rated for 130A and I have installed km of it for house incomming mains (100A fused)

My solar is fed off two 8AWG and gets upto about 55A on each cable, granted it gets warm... not hot. and at the limit as to what I find comfortable to use it with. 8AWG is rated for over 70A but that's hot to the touch and would cause thermal expansion/contraction issues if cycled daily.

If your breakers are tripping due to heat at 80A then your terminations may be the issue because 4AWG cable will transfer qute a bit of heat away from the breaker.

Just a thought looking at the output from controller 3 vs the battery voltage do you have the float and bulk voltages on the MSB set to the same voltage ? The question is why did controller 3 not pickup again when the battery voltage declined ?

---

Looking at the cell voltage chart the timing is different as to the peak compared to the chart voltage with the controllers in. Do you have the same day data on each chart ? The data does not match ?

Wolf said:

OffGridInTheCity said:

Controllers - I have 3 x Midnite Classic 150s, the oldest is 2 yrs..........................................

Click to expand...

Yea I have looked at them and they do look nice. Would you get 2 for my small system? I don't mind spending money but ~1400 for 2 with only 3000W of panel. Or get 1 and keep the MSB as backup. LOL
Also if I do use only 1 how do you think it will work with my panels easterly and westerly orientation?

On another note today the MSBs worked just fine. Shrugg

Wolf

Click to expand...

I would start with one. The Classics handle a 4000w range - https://www.midnitesolar.com/sizingTool/. Assuming your arrays (East and West) are similar panels/wired for a similar voltage, then I don't see why one wouldn't work. My panels wired in 3s. Each Midnite does a3s5p set of panels. Some of the 3s(s) are shaded in first 1/2 of the daybut the voltage is fine and parallels in with the other (unshaded) 3s(s) and things seems to work very efficiently - e.g. a common voltage (+/- 10%)but differing amps is the best design for the controller's MPPT.

The Classics have more than one MPPT tracking algorithm with 'custom adjustments'to fool with if you're so inclined. I use "Legacy P&O".

Solar
This is the default mode for PV systems and has a very fast sweep (typically1/2 second or less) that will re sweep at user adjustable sweep intervals, unless the Classic finds that it needs to do a sweep on its own because of changing conditions. The timed sweep interval is user adjustable and is in units of minutes. SOLAR mode is typically best for PV systems, especially if there is partial shading at times during the day. SOLAR mode is best suited for shaded or un shaded PV arrays that are at least one nominal voltage above the battery voltage. For severe partial shading or PV arrays with nominal voltage equal to battery voltage, you may also want to try Legacy P&O (Perturb and Observe) MPPT mode.

Legacy P&O
Legacy P&O (Perturb and Observe) mode is a slow tracking mode similar to the Micro Hydro mode but with the difference that it is slightly faster and will shut off if the power source goes off.
It has two settings that are user adjustable. Sweep Interval is the time between mini sweeps, in minutes, and sweeps around the present (i.e. the last found) MPP Voltage. The range of this sweep is determined by the Sweep Depth user adjustment and is expressed as a percentage of Watts that the sweep started from. For example, if in Legacy P&O mode, the Classic was outputting 500 Watts and the Sweep Depth percentage was set for 10% (50 Watts), the sweep will bring the input voltage DOWN until the output power drops down to 450 Watts, then will sweep UP in voltage until the power drops again down to 450 Watts and then go back to the newly found MPP Voltage, waiting for the next sweep.

Dynamic
This is typically used for PV (solar) input sources and tries to follow, on a slow dynamic basis, the changing conditions of the input source. This mode has one user adjustment which is a forced sweep perturb trigger interval for times when the input condition changes do not trigger a dynamic sweep. The interval is in units of minutes

U Set VOC%
This is a fully manual mode based on a percentage of VOC. The Classic will sweep based on the user set time in minutes and then park at a user set percentage of the VOC the Classic found on that sweep. This mode is useful for testing or constant voltage sources.

---

completelycharged said:

OffGridInTheCity said:

becasue my 80a circuit breakers tripped (and 4awg wire won't handle it)

Click to expand...

4AWG is rated for 130A and I have installed km of it for house incomming mains (100A fused)

My solar is fed off two 8AWG and gets upto about 55A on each cable, granted it gets warm... not hot. and at the limit as to what I find comfortable to use it with. 8AWG is rated for over 70A but that's hot to the touch and would cause thermal expansion/contraction issues if cycled daily.

If your breakers are tripping due to heat at 80A then your terminations may be the issue because 4AWG cable will transfer qute a bit of heat away from the breaker.

Click to expand...

I have 3 x Midnites - each has a +/- 4AWG that's15ft from the controller to control box bus. 4 of the wires share a1" flexible conduit (the other 2 have their own 3/4 EMT). The 4 wires coming out of the 1" flexible conduit peakaround 70-80a /55C (130F) inside the control box atmid-day. Here's a pic showing the wires coming into the control box and the temperature probe showing 55C temps....




I'm going with <60C is 'OK' in my mind. After all, wire is rated for this andthe Midnites run 55C PCB and FET temps each summer day andthe 4 AWG wire connects into the PCB at the Midnite itself.

Its amazing how much the wire length affects things - e.g. 1 inch vs 1 foot vs 15 feet can really make a difference. In the control box above I had a 1 inch 4 AWG segment (for a short while) connecting the bus to the shunt - and it handled 200a OK. Didn't even get warm! But I couldn't take it so I replaced it with a 2/0 segment

completelycharged said:

Just a thought looking at the output from controller 3 vs the battery voltage do you have the float and bulk voltages on the MSB set to the same voltage ? The question is why did controller 3 not pickup again when the battery voltage declined ?

---

Looking at the cell voltage chart the timing is different as to the peak compared to the chart voltage with the controllers in. Do you have the same day data on each chart ? The data does not match ?

Click to expand...

---

CC,

Yea its the same data but you know you can look at my Grafana dashboards at any time.
http://wolftech.mynetgear.com:4562/d/NiaQdVemz/westbrook-plant?orgId=1&refresh=10s
user -guest pw - guestw
I even put together the pertinent graphs in a dashboard called Playground,
Just select the last 2 days on the right or whatever and then you can drill down to the incident and examine it if you want.

6/18/2020 between 12:00and 14:00, May be different in your time zone though.

Ah yes the MSB issue with voltages. If have both of them to set to14.2 Absorption and Float (I lied in my initial post) which is 56.8V =4.057 per cell.
The problem is that the units are 0.2 Volts different from each other and tend battle once the system goes to float so 1 will drop out and it is usually Controller#2 the east side and #3 the west will pick up and continue. Also the V readings on the MSB controllers isn't the most accurate either.
During the "incident" the battery voltage was high so no controllers where charging till the V on the pack went below float.

OffGridInTheCity said:

I would start with one. The Classics handle a 4000w range - https://www.midnitesolar.com/sizingTool/. Assuming your arrays (East and West) are similar panels/wired for a similar voltage, then I don't see why one wouldn't work. My panels wired in 3s. Each Midnite does a3s5p set of panels. Some of the 3s(s) are shaded in first 1/2 of the daybut the voltage is fine and parallels in with the other (unshaded) 3s(s) and things seems to work very efficiently - e.g. a common voltage (+/- 10%)but differing amps is the best design for the controller's MPPT.

Click to expand...

OffGrid
Yes I am running them in a 3s2p config. if I went with 1 it would then of course be 3s4p.
Something to mull over.
Thanks for the info.

Wolf

The timing of the peak in this chart for the battery is showing at 10am compared to the 2pm peak on the other battery chart.

The difference looks partly due to voltage drops correlated with what might be expected with the current flows, although it does not look right.

The fridge is great for showing a relative consistent volt drop delta , which interestingly dissapears between 10am and 12:30. This is another pointer to me that something might be amiss with the wiring with volt drops.

Check the wiring out for voltage discrepencies caused by loose connections. Too much data ? Naaaahhhh.





I had noticed the rather poor consistency in volt readings with the MSB controllers. Even if controller 2 had drifted higher the concistency of output relative to the voltage seems inconsistent.

CC

I now understand your and my confusion.
You are looking at amps not volts on that chart.
So Yea at 10:00 AM the batteries where being charged hence the red trace climbing.



All that chart shows isthe amps correlation between both Charge controllers the Grid ties and the battery.

Wolf

Ah, yes, my mistake.

I was looking at it from how I have seen voltage drops create some od values on my setup. The smooth current profile makes sense as well as the battery was not providing the change in current as it was all from the controllers.

Looking at the load spike just as the full output started I'm wondering if thier implementation of float vs absorb timing has an issue and the brief spike and voltage drop that it would create caused it to flip back into bulk mode but at the wrong voltage level expectation ?

It is very much indeed a curious incident that I'm wondering about a bit more now.

Been out to have a look at my units and..... one of them was indicating voltages all over the place for the battery output. 52, 54, 57, 51, 54... adjusted the voltage setting to 13.2V, no difference. switched off some of the solar input the output settled a bit, switched more of the solar off, random numbers stopped.. hmmmmm.... the unit temperature at the time was reading 47C with the fan on. The unit has 4kW of solar attached and was reading about 1800W at the time. Each string was capable of about 600W under current sun, so 2400W available and close (probably too close) to the unit limit. I forgot to disconnect one of the strings onto another controller (from the winter) so may have fried some of the internals, which is impacting the cut-off ?

The other 2 units are not impacted.

Makes me wonder if this is a similar situation as to what you have had. So now figuring out how to connect the solar as I may revert to re-wiring with a simpler PWM and buy another 4 panels instead.

---

"Output power will be normal when its less 55 C." - so the unit should not have been derating the output at only 47C...

---

I'm actually quite happy you reported your issue because I would have not checked my units for the output voltage !!

completelycharged said:

I'm actually quite happy you reported your issue because I would have not checked my units for the output voltage !!

Click to expand...

Anything for a friend

This anomaly has not happened again since the first incident. Don't know if it ever happened before as I was running on my pb bank with different absorption and float V but I never noticed it.
I am keeping a close eye on it now though.
In my case I only have a max of 1500W of panel for each controller. I figured since it was rated at 50A and made in china I was safe with that input.
I have seen the controllers put out ?25A each at optimum MPPT. Still only of its "rated" capacity so I shouldn't have fried anything. :huh:

Wolf