Parallel Tristar MPPT 60 Problem

degraaffray
degraaffray Registered Users Posts: 3
Hi Guys,

New poster here although I have been around looking and gathering some fantastic information for some time. Great forum and a wealth of knowledge, so seemed to be the natural place to post my question.

I have an off grid system consisting of a Studer Innotec inverter, 1150AH FLA battery bank @ 24v and 2 Tristar MPPT 60 CC's. Each CC is fed from a separately located and switched/fused array, one of which is very slightly smaller than the other. The smaller array goes in to the shade about 1 hour before the larger.

My problem appears to be more of a niggle or a worry than an actual problem, but whilst Morningstar support has bee absolutely great, they don't seem to be able to answer my query.

The MPPT CC's run on the MODBUS system, and whilst I have the correct hub, and get the aggregated data on my display, the CC's do not actually communicate with each other, which I knew, so each one has no idea what the other is actually doing. This means that each will behave in its own way based on the custom configuration settings.

What is happening on my system is that all is working perfectly, with batteries fully charging every day and hitting float etc. However, my absorb time is set at 120 minutes (2 hours), and depending what I have drawn the night before, if the batteries have dipped below 24.8v and extra 30 minutes is added to this. Now the main controller in the morning hits absorb generally around 11:00am and is very quickly followed by the second controller, maybe just a couple of minutes behind as that is how much longer it takes the sun to hit those panels. Now when the first controller times through to float (120 minutes), the second controller is still in absorb for another few minutes. Therefore when the first drops down to 27.2v for float, the first sees this as a drop and pushes straight back to bulk, lifting the volts to 28.6 (temperature compensated from 28.8 ).

This shouldn't be a problem according to Morningstar, but to me it now means my batteries are getting far more bulk charge than I want. Although the first controller does drop off the current to 0, it still means the second controller is holding up the volts as it is yet to complete its absorb phase, and in fact is now back in bulk.

I hope this all makes sense as it is hard to explain, but looking at both units simultaneously on MSView, it is really obvious what is happening, and I worry that it isn't going to be doing my batteries much good staying at the absorb and bulk charge rates longer than I (or Hawker) want.

Any help or advice would be welcomed.

PS I have considered timing the sun profile on each array and adjusting the timers so that they match the actual time the sun is on the panels. Maybe this would help?


Many thanks.

Ray

Comments

  • Vic
    Vic Solar Expert Posts: 3,208 ✭✭✭✭
    Hi Ray,

    Welcome to the Forum.

    This is a common situation when using multiple Charge Controllers (CCs), when the CCs are unable to coordinate charge states, or at least be able to coordinate the transition to Float. Your description was quite understandable, BTW.

    Sometimes it is advantageous when one CC has a somewhat larger PV array than others. Your shading issues adds an additional variable that might change with seasonal variations, but, you might try making the CC that has the largest PV array the laggard. You could try this, by setting the Absorb voltage slightly higher than the setting on the other CC. This can make the CC with the smaller array go to Absorb first, and could allow the lagging CC to complete Absorb (although a bit delayed) due to it having greater PV power.

    Sometimes there is not any guarantee that this will always work, or even ever work ... but this has helped on one system, here, before going to two MidNite Classic CCs, which can coordinate charge states.

    You might need to use MS View software to be able to customize the Absorb voltage, as believe that Morningstar is still using DIP Switches. FWIW, Vic
    Off Grid - Two systems -- 4 SW+ 5548 Inverters, Surrette 4KS25 1280 AH X2@48V, 11.1 KW STC PV, 4X MidNite Classic 150 w/ WBjrs, Beta KID on S-530s, MX-60s, MN Bkrs/Boxes.  25 KVA Polyphase Kubota diesel,  Honda Eu6500isa,  Eu3000is-es, Eu2000,  Eu1000 gensets.  Thanks Wind-Sun for this great Forum.
  • degraaffray
    degraaffray Registered Users Posts: 3
    Hi Vic,

    Thanks so much for your response it is really useful. I hadn't considered what you suggested so I will give that a go.

    I always use MSView anyway so making the changes is not a problem. I actually went for the Tristar over everything else as they seem to be the only ones that provide passive cooling, and as my system is in my holiday home in a very dusty national park in Southern Spain, dust really is an issue (a neighbour has ready had 2 fans fail on his outback Flexmax 60).

    Anyway thanks for the info and I'll apply that and see how I get on.

    Many thanks

    Ray.
  • Vic
    Vic Solar Expert Posts: 3,208 ✭✭✭✭
    Hi Ray,

    Thanks for the info on your situation.

    You might need to play with the Absorb voltage settings a bit to try to tune things, and perhaps, when you get it to work well, changes in loads and/or season might mean a little bit of retuning might be needed -- all part of off-grid husbandry!.

    Have Fun, and please let us know how you are doing. Vic
    Off Grid - Two systems -- 4 SW+ 5548 Inverters, Surrette 4KS25 1280 AH X2@48V, 11.1 KW STC PV, 4X MidNite Classic 150 w/ WBjrs, Beta KID on S-530s, MX-60s, MN Bkrs/Boxes.  25 KVA Polyphase Kubota diesel,  Honda Eu6500isa,  Eu3000is-es, Eu2000,  Eu1000 gensets.  Thanks Wind-Sun for this great Forum.
  • degraaffray
    degraaffray Registered Users Posts: 3
    Ok by way of update, it transpires the issue is related to the battery sense cables being unequal lengths (by quite a bit).

    The units apparently "watch and mimic" each other's status pretty closely by way of voltage measured on the battery sense cable, with each of the units needing to have identical custom settings programmed (voltages and timings). I have 2 separate battery sense cables as I felt this was going to be more accurate than a parallel link, but as one was only put on temporarily as I was coming home it wasn't trimmed to length, but instead left partially coiled. As such when I log on remotely to view both units side-by-side I can see the voltages measured can differ by up to 0.25v. This means the "lagging" unit is seeing absorb later, and subsequently timing to float later as this is simply a pre-defined time delay from absorb. Hopefully the trimming of the cable to the same length will resolve this issue (fingers crossed).

    With the above noted, I still can't find out for the life of me where the TS MPPT 60 gets its date/time stamp from? There is nowhere in the settings that I can see for adjusting or setting the time, so assume it gets the information from the http port, but what if this isn't connected? When I download the log files each log has a date/time stamp, which is in UTC time standard, so it must pick this information up from somewhere, otherwise any internal clock would be prone to drift and not relevant in different time zones?

    Thanks for your help and patience as I really like to know how things work to understand them better.

    Ray
  • MStar1
    MStar1 Registered Users Posts: 24 ✭✭
    edited September 2016 #6
    Hi Ray,

    The voltage sense circuit has very low current so it is designed to detect the voltage with no voltage drop so wire length will not matter.
    Though the accuracy of the controllers is good, <.15V for each controller in a 48V system, it will not always exactly match. If you can custom program the controllers you can make a small voltage adjustment to one or both of the set-points. You might want to measure the voltage with a meter to compare to the controller readings and make adjustments accordingly. If the TS is indicating lower than what is measured set the Absorption lower. If indicates higher set it higher
    Example: measured Vb = 52.15V; TS indicates 52.28V, set the Vabsorption to .13 higher than you want. 

    As Vic suggested, creating a slightly lower voltage setpoint and shorter time for the smaller controller will allow the larger controller to stay in Absorption charge which would be less likely to drop into Bulk since it has a larger array. 

    Transitioning out of Absorption when the charge current is high may be a problem with your system meaning that the Absorption time might be too short. By the time you get through Absorption with the first controller the current requirement to stay in Absorption should be low enough not to let the second controller drop out of Absorption at least on a sunny day. When you factor in Load current, if the net current is somewhat high it could mean the battery is not fully charged. That is assuming that it has not become sulfated to the point where it will not reduce current during Absorption along with not holding a charge.

    2 hours or 2.5 hours seems a rather short Absorption time. Our presets are 3 hours or 5-6 hours with extension time for FLA batteries.
    According to one of the biggest battery manufacturers in N. Amer, FLA batteries that stay watered basically can't get overcharged at reasonable Absorption voltage levels though they may need to be watered more often. I am also currently working with another major US battery manufacturer at their request to provide custom settings that are longer than our presets for both Sealed and Flooded batteries.

    For parallel controller systems, what if the effect if one of the controllers enters sooner? This means that it will tend to extend the Absorption time if the battery had reached a very low SOC the previous night. If the batteries are not discharged very much both controllers will enter Absorption early in the day since it will take very little current to maintain the regulation voltage. It will tend to extend charging when the battery needs it to be extended. Is this maybe better than forcing one of the controllers out of Absorption when the battery really needs more charging?

    -MStar Applications EE

  • mike95490
    mike95490 Solar Expert Posts: 9,583 ✭✭✭✭✭
    Have you used the web monitor for the controllers ?   A simple Hub and 3 cables with your laptop will show what each is doing and for how long,   
    And I think you should extend the Absorb time, 1 hour seems short,

    There is NO time/date stamp, the internal clock restarts a DAY when the morning sun wakes up the panels.   Try switching off the panels in the daytime for 30 minutes, and then back on, it will log a new day then.    There IS a random 30 minute update cycle of data storage in RAM before it writes to ROM (long term logging memory) so there is a slight chance of loosing up to 30 minutes of logging if power is cut.
    Powerfab top of pole PV mount | Listeroid 6/1 w/st5 gen head | XW6048 inverter/chgr | Iota 48V/15A charger | Morningstar 60A MPPT | 48V, 800A NiFe Battery (in series)| 15, Evergreen 205w "12V" PV array on pole | Midnight ePanel | Grundfos 10 SO5-9 with 3 wire Franklin Electric motor (1/2hp 240V 1ph ) on a timer for 3 hr noontime run - Runs off PV ||
    || Midnight Classic 200 | 10, Evergreen 200w in a 160VOC array ||
    || VEC1093 12V Charger | Maha C401 aa/aaa Charger | SureSine | Sunsaver MPPT 15A

    solar: http://tinyurl.com/LMR-Solar
    gen: http://tinyurl.com/LMR-Lister ,

  • MStar1
    MStar1 Registered Users Posts: 24 ✭✭
    edited September 2016 #8
    There is NO time/date stamp, the internal clock restarts a DAY when the morning sun wakes up the panels.  
    There is an internal Hourmeter which counts the total hours of operation. When the Logged data is download all the internally logged data gets a UTC time based on the Hourmeter stamps in the logged data subtracting from the time on the PC. One of the problems with this is that if the controller is not in operation it will not count that time.

    If you want to create local data and time columns in Excel we made a white paper showing how to do this. 
    http://www.morningstarcorp.com/adjusting-morningstar-data-log-file-time-zone/

    -MStar Applications EE

  • Iceni John
    Iceni John Solar Expert Posts: 103 ✭✭

    How's this for an oblique way to solve your dilemma:

    Split your battery bank into two halves, have each CC charge its own half entirely separately of the other, and combine the two battery banks with a Schottky diode on each to prevent one bank back-feeding into the other.   This is what I've done on my bus  -  each row of four panels feeds its own CC and battery bank, as much for redundancy as anything.   If one bank charges slower than the other, or if one array produces less power than the other (quite likely on a bus roof!), it doesn't matter.   Essentially I have two completely separate systems running in parallel, and each contributes what it can to the combined DC load center.   Another incidental side benefit is that I will change out half my batteries but twice as frequently, so I'll never have all of them approaching the end of their life at the same time.

    It sounds very socialist, but it's working for me.   When you convert a bus, you learn to think outside the box, WAY outside!

    John 

    40' Crown bus with 2kW of tiltable panels on the roof:

    Eight Sharp 255W, two Morningstar TS-MPPT-60, Magnum MS2000, Champion C46540 generator converted to propane, eight golfcart batteries, and maybe a small Exeltech inverter for the fridger.

    Southern California

  • mike95490
    mike95490 Solar Expert Posts: 9,583 ✭✭✭✭✭
    edited September 2016 #10
    MStar1 said:
    There is an internal Hourmeter which counts the total hours of operation. When the Logged data is download all the internally logged data gets a UTC time based on the Hourmeter stamps in the logged data subtracting from the time on the PC. One of the problems with this is that if the controller is not in operation it will not count that time.

    If you want to create local data and time columns in Excel we made a white paper showing how to do this. 
    http://www.morningstarcorp.com/adjusting-morningstar-data-log-file-time-zone/
    Awesome, I'll have to look at that.  Thanks
    Powerfab top of pole PV mount | Listeroid 6/1 w/st5 gen head | XW6048 inverter/chgr | Iota 48V/15A charger | Morningstar 60A MPPT | 48V, 800A NiFe Battery (in series)| 15, Evergreen 205w "12V" PV array on pole | Midnight ePanel | Grundfos 10 SO5-9 with 3 wire Franklin Electric motor (1/2hp 240V 1ph ) on a timer for 3 hr noontime run - Runs off PV ||
    || Midnight Classic 200 | 10, Evergreen 200w in a 160VOC array ||
    || VEC1093 12V Charger | Maha C401 aa/aaa Charger | SureSine | Sunsaver MPPT 15A

    solar: http://tinyurl.com/LMR-Solar
    gen: http://tinyurl.com/LMR-Lister ,

  • mike95490
    mike95490 Solar Expert Posts: 9,583 ✭✭✭✭✭
    Iceni John said:

    .....Split your battery bank into two halves, have each CC charge its own half entirely separately of the other, and combine the two battery banks with a Schottky diode on each to prevent one bank back-feeding into the other.   This is what I've done on my bus  -  each row of four panels feeds its own CC and battery bank, as much for redundancy as anything.   If one bank charges slower than the other, or if one array produces less power than the other (quite likely on a bus roof!), it doesn't matter.   Essentially I have two completely separate systems running in parallel, and each contributes what it can to the combined DC load center. .....

    What wattage and how do you heat sink that diode ?   Even a Schottky with 30A going through it,  (60A load total) is going to dissipate 20 watts and require heatsinking,  I suppose it does give some redundancy and isolation between banks,
    Powerfab top of pole PV mount | Listeroid 6/1 w/st5 gen head | XW6048 inverter/chgr | Iota 48V/15A charger | Morningstar 60A MPPT | 48V, 800A NiFe Battery (in series)| 15, Evergreen 205w "12V" PV array on pole | Midnight ePanel | Grundfos 10 SO5-9 with 3 wire Franklin Electric motor (1/2hp 240V 1ph ) on a timer for 3 hr noontime run - Runs off PV ||
    || Midnight Classic 200 | 10, Evergreen 200w in a 160VOC array ||
    || VEC1093 12V Charger | Maha C401 aa/aaa Charger | SureSine | Sunsaver MPPT 15A

    solar: http://tinyurl.com/LMR-Solar
    gen: http://tinyurl.com/LMR-Lister ,

  • Iceni John
    Iceni John Solar Expert Posts: 103 ✭✭
    edited September 2016 #12
    mike95490 said:
    Iceni John said:

    .....Split your battery bank into two halves, have each CC charge its own half entirely separately of the other, and combine the two battery banks with a Schottky diode on each to prevent one bank back-feeding into the other.   This is what I've done on my bus  -  each row of four panels feeds its own CC and battery bank, as much for redundancy as anything.   If one bank charges slower than the other, or if one array produces less power than the other (quite likely on a bus roof!), it doesn't matter.   Essentially I have two completely separate systems running in parallel, and each contributes what it can to the combined DC load center. .....

    What wattage and how do you heat sink that diode ?   Even a Schottky with 30A going through it,  (60A load total) is going to dissipate 20 watts and require heatsinking,  I suppose it does give some redundancy and isolation between banks,


    I have two Cole-Hersee 48161 diodes that are rated for 250 amps each.   Both sides of both diodes are tied together with 1" x 1/4" copper busbars, with another busbar to take power to a Blue Sea 9002 switch that can bypass the diodes if I need to charge the batteries from my Magnum inverter/charger.   I think they're intended for use in emergency vehicles  -  they meet some arcane federal regulation for use in ambulances.   Even when drawing 2000W total from the batteries their heatsinks get only slightly warm, so I guess they're a lot more efficient than the usual silicon diodes that have much larger heatsinks in comparison.   I've never measured voltage loss through them under different loads, but if I lose a few tenths of a volt it doesn't matter because they're not in a charging system, and a slightly lower house DC voltage may help prolong the life of 12V lights.   So far, so good!


    John 

    40' Crown bus with 2kW of tiltable panels on the roof:

    Eight Sharp 255W, two Morningstar TS-MPPT-60, Magnum MS2000, Champion C46540 generator converted to propane, eight golfcart batteries, and maybe a small Exeltech inverter for the fridger.

    Southern California