PIP-4048MS and PIP-5048MS inverters

[lopezjm2001]

[quote="weber"] Lopezjm2001, you have an unusually high bypass current of 2.25 A because of your 3 parallel strings.

I would expect this bleeding current to decrease as the battery packs balances.

But the BMS master current would usually be less than 0.1 A.

According to my centralised mini-BMS manual it's operating current is less than 10mA. So 3 x 10mA total 30mA.
Dimitri stopped making the centralised mini-BMS years ago so only the more common distributed mini-BMS is available that uses a separate head board.

So that still leaves about 0.65 amps in the budget for contactors. These typically (EV200 or GX14) pull a large current for a short time when turning on, but use less than 0.2 A each at 12 V to hold. If there were two of them and your DC-DC was only 70% efficient, that would still only be 0.15 A at 50 V.

I measured all subsystem charger control current to be about 1.2 amps from 48V battery packs using a DC clamp meter. So I expect that it just takes a long time for cells to balance.

So it should still come in under 3 A and therefore the existing LiFePO₄ patch should work now that you've bumped parameter 2 up to 120 A.

Sill checking. Clouds keep interfering.

UPDATE: A HVC occurs (HVC=3.65volts) when clouds pass by dropping out a contactor which disconnects the solar array. It appears to cause the PIP4048MS to go automatically into float mode, I need to verify this a second time. Can you verify this in the Lifepo4 patch?

What kind of contactors do you have?

Using 3 (not 2 as previously reported) contactors Nanfeng ZJW200A. They each pull about 1 amp at 12Vdc continuously.

What voltages do you have for parameters 26 and 27 (absorb and float)?

Absorb - 55.0 volts
Float - 53.7 volts

At what cell voltage does your BMS start to bypass?

Being an analogue BMS they start burning at 3.40 - 3.45 volts.

[weber]

That all sounds good. Except you have unusually high contactor coil drive (as well as high BMS bypass) because of using contactors with no economiser. So I'd expect the final current to be around 3.3 A, which hopefully the PIP will read as 3 A and not 4 A.

DC clamp meters are very inaccurate for such low currents. Mine is only good to ±1 A. Coulomb has one that's good to ±0.5 A I think.

Yes, I should have asked whether they were balanced. As you suggest, since they are not balanced, it may just be doing exactly what we'd want it to do, and may go to float when they are finally balanced.

The PIP we're running off-grid, with 180 Ah cells, only has 0.8 A bypass and less than 0.6 A contactor drive (3 contactors with low hold currents). Total, less than 1.4 A.

[paulvk]

A relay needs far more current to first operate than it needs to hold so a trick to reduce this hold current is to put a resistor in series with the coil shorted by a set of normally closed contacts.
My telco cabinet uses this to reduce the current drawn by a large 500amp 48v relay that disconnects the batteries when the voltage goes below a set value as this relay has only one large set of contacts it is driven by a small PCB relay that inserts the resistor.

[coulomb]

... when clouds pass by dropping out a contactor which disconnects the solar array. It appears to cause the PIP4048MS to go automatically into float mode, I need to verify this a second time. Can you verify this in the Lifepo4 patch?

It's not something I explicitly changed for the patch, no. Maybe the battery is near or above the CV voltage, so it just needs to see the charge current go below the threshold. Maybe that never happens because of your high non-PIP battery loads. So it takes a cloud to reduce the available charge current below the threshold for the 50 seconds to go to float mode.

That doesn't sound ideal to me, more like a defect in the charge strategy that was really designed for mains charging, but maybe that's the temporary "get out of jail" card that you are looking for.

[ Edit: 50 seconds or two minutes -> 50 seconds ]

[lopezjm2001]

a trick to reduce this hold current is to put a resistor in series with the coil shorted by a set of normally closed contacts.

Good idea. Problem is that my contactors only have one normally open contact. I could use a small relay in conjunction with the contactor. But just how much current will I decrease the holding current of 1 amp by?
What value resistor? Watts?

It is explained here:- Economy Resistor in DC Coil

Just need to find the coil's holding current.

Update:

Found the holding current to drop out at about 165mA, so settled for 200mA, So economy resistor = (12v/0.2amps) - 12ohm = 48ohm,
Power = 0.2amps x 0.2amps x 48ohms = 1.92 watt, so use 5 watt resistor.

Using a 50ohm ceramic resistor 10 watts and an Omron time delay (0-10sec) relay H3Y-2 12volts. Timer power consumption: 75mA.

[paulvk]

You need to experiment to find the holding voltage
A variable power supply is good for this.
The relay in my cabinet has 1/2 inch silver contacts its fairly large its like this one http://www.electriccarpartscompany.com/ ... _p_26.html
And yes it is driven by a small relay which inserts the resistor but you could drive the small relay from the larger ones output so when the larger one operates it turns on the small one that takes the short off the resistor.
You will be surprised how little current it will take to hold up the relay.

[weber]

A few days ago the Australian Clean Energy Council released the following guidelines for Grid-Connected Energy Systems with Battery Storage. Most of it is also relevant for off-grid systems.

http://www.solaraccreditation.com.au/da ... nstall.pdf

It addresses Lithium Ion cells to some degree, which is good because existing standards don't even mention them, but I was disappointed with this:
"Using the manufacturer’s data, the installer shall determine the correct response to a fire either initiated by the battery or an external source which may pose a hazard to the battery. Either as part of the installation process or by formal agreement with the owner, a suitable fire response must be provided adjacent to the battery with simple operating directions in the event of a fire."

It's good that fire response instructions must be posted. But it's bad that it's up to the installer to figure out what that response should be, "using manufacturer's data". So these guidelines give no guidance at all in regard to fire response.

Can anyone provide an example of a recommended fire response from a battery manufacturer (or battery-based equipment manufacturer) for Lithium Iron Phosphate?

Surely the best response is the same for the same battery chemistry irrespective of manufacturer. And surely someone has determined what that best response is by now.

[lopezjm2001]

Can anyone provide an example of a recommended fire response from a battery manufacturer (or battery-based equipment manufacturer) for Lithium Iron Phosphate?

The only example I have seen is in the ENGINER PHEV 48Vdc Lifepo4 Battery pack for PHEV cars. The whole kit including Lifepo4 batteries are installed inside a stainless steel case with an automatic fire extinguisher inside. The kit can also be used as a UPS for the home. I am using these as an off-grid solar power system with a PIP4048MS.
ENGINER went bankrupt about 3 years ago.

Putting the onus on the installer is the CECs response to everything. They make you earn those STCs.
What about a response to a battery explosion?? My instructions would be to exit as fast as possible. Call fire brigade. (Get advice from a solicitor about directions you post).
Maybe a copy of the battery pack's MSDS should be posted outside so the fire brigade know what chemicals they have to deal with. Maybe install a fire alarm in the battery room which alerts the occupants to exit the premises and which automatically calls the fire brigade. If you can get a signed statement from the owner that he will install a fire alarm/"000"caller in the battery room then your arse is covered. Probably what unethical solar retailers (you know them) will likely do.

[weber]

Thanks lopez. The competitor to the CEC, the ESC, have just put out a battery "guide" as well. It is even more useless in regard to enclosure design and fire response.
http://issuu.com/solarcouncil/docs/batt ... 3/35308740

The word "Lithium" only appears once, in the preface, to tell us that "Lithium-ion based batteries are quickly gaining market share in this sector".

Page 28 is particularly breathtaking in its uselessness.

[kolargol]

Hi to all forum members from France (Sorry for my english).

Thanks for all your contribution about wks (and similar inverter) inverter

I am electrician and one of my hobbies is software development and technologies

I develop principally in java,and vb.net, android, php...

My question is about wks COM port.

I need to develop and app to retrieve data and send value (power, etc) to a server

Actually , i do that with a personal system (sensors, arduino mega, wifi chip, etc...), but i want do that directly retrieving this information from the inverter with a computer (old one potable PC with windows 7)

Before trying to connect with the inverter, i test personal app in "Csharp" and "java" with serial port (usb connector of arduino) with an arduino.
All work fine with com port, no problem.

When i try to connect inverter with personnal app, (Csharp app and java) app never find COM3 and give me an error.
It s the same with this command line app https://github.com/scottwday/AxpertTest.

My wks is 3kva and the data output is directly USB port, not a serial like exemple in this forum.

This morning, i try to connect with my personal code to a little inverter (http://www.ldlc.com/fiche/PB00115604.html)   and the apps never find com3 to.

I doesn't understand what is different in communication with and arduino and an inverter.

Someone could help me?

Thanks
[Edit: Removed excess blank lines from end]

[coulomb]

... All work fine with com port, no problem.

When i try to connect inverter with personnal app, (Csharp app and java) app never find COM3 and give me an error.
It s the same with this command line app https://github.com/scottwday/AxpertTest.

I've had the same problem. I believe that some on the South African Power Forum site have solved this. I can't find the definitive answer (I just use the COM port and not USB), but this might help:

http://powerforum.co.za/topic/290-axper ... re/?page=2

I assume you know to kill the WatchPower software (find it in the system tray and use File Exit), as it will auto-connect to the serial port and make it busy. This will make it not appear in the serial port list.

The answer may be in some of the third party replacements for WatchPower; there are at least two links in the index (first post of this thread).

Edit: I also assume you realise that the comm port and the USB port use the same serial port of the DSP (the inverter's main processor). So if the comm port is busy with some other software, you won't be able to use or even see the USB port. They are two ways of accessing the same thing.

[kolargol]

For information, My new investigation about com port and usb:


go to installation directory of app

ex:  D:\app\WatchPower\lib

take  file:  "bak_WatchPower.jar"

go to website   :   http://www.javadecompilers.com/

decompile file "bak_WatchPower.jar"  and save result to your computer

Unzip result  and you have all java directory and source  and the start of app and search and connect to port com is in:

\bak_WatchPower_source_from_cfr\cn\com\voltronic\solar\console\windows\MainConsole  (for window)

and in this file  the start of procedure is line 117

        SearchDevice searchdevice = new SearchDevice();
        searchdevice.start();

after that search class:

\bak_WatchPower_source_from_cfr\cn\com\voltronic\solar\work\SearchDevice

and after 

\bak_WatchPower_source_from_cfr\usb\USBCommEx

etc

[Northland]

reads in next days weather forecast < when running from batts at night, voltage level based on next days weather i.e. if you know its gonna be sunny and get a good charge, why not use more battery or less if its gonna rain etc

This may be the thing I've been looking for. I've been running my SCR water heater system for 6 months now, and only had 1 day where the water wasn't hot enough (had a bath instead of a shower, probably would have been ok otherwise). But that's summer and autumn, I should design for winter conditions. So I'm thinking of buying a temperature controller module off eBay, and jacking up the thermostat on a sunny day if I know tomorrow is rainy.

A little inconvenient and automated would be better. So your software would be the go, plus a few mods. Will upgrade my SCR while I'm at it as I think 2.4kw will be ok.

Questions
- I have arduino sitting around, how different is raspberry pi?
- would it be possible to have the pi adjust the controller?

On another note, we were discussing fuses a while back. I found these 60A fuses on ebay for $2.

[kolargol]

Original Arduino is not a computer like Raspberry, it's a microcrontroller with usb connetor to upload your program and a serial com
There are lots of models of arduino
The basic version, have digit and analog imput , digital output , and low memory
More expansive arduino like PCDUINO is similar to a raspberry

Arduino are more oriented to automation developpement and value mesurement that raspberry

[Northland]

Here is a cheap 12v contactor . No extra set of contacts though

48v contactor with N.C contacts

[Northland]

I have arduino mega. One issue stopping me from using thus far is the input voltage limited to about 18v. I need to measure 48-56v. Anyone know a solution or if pi can do it?

[coulomb]

One issue stopping me from using thus far is the input voltage limited to about 18v.

You just need a voltage divider. Say divide by 4. So then 48-56 V becomes 12-14 V.

It's just a little trickier because of the existing divider on the Arduino; the processor itself can't read 18 V. But as soon as you know the input resistance, you can calculate what "upper" resistor you will need. See any basic book on electronics. [ Edit: this advice was based on a wrong assumption; see below. ]

[Northland]

Or simpler, just measure across one battery?

[paulvk]

Questions
- I have arduino sitting around, how different is raspberry pi?
- would it be possible to have the pi adjust the controller?

Which Arduino?
I have written a lot of code for use in the AVRs (chip in Arduino)
I write in Bascom its much more stable than the Arduino it has even passed space flight testing!
Working on web server system to control my systems at the present so I am solving the interfacing problems now if your interested.

[neilg]

Northland wrote:

One issue stopping me from using thus far is the input voltage limited to about 18v.

The 18 volts is the maximum power supply voltage on the power connector.
ALL input and output pins on the Arduino Mega are a maximum of 5 volts.
The CPU internal analogue reference voltage is a max of 4.096volts so each analogue pin can only measure up to this value or 5 volts if you use the CPU voltage as the reference.
For simplicity, a voltage divider of 150k and 10k will allow you to measure up to 65 volts.
That is: +65v - 150k - Analogue pin - 10k - ground.

Hope this helps.
Neil

[coulomb]

That is: +65v - 150k - Analogue pin - 10k - ground.

I thought 18 V maximum was a strange value for an analog input. 4.096 V makes much more sense. Thanks, neilg.

Note that the 15:1 ratio resistance values cause a 16:1 voltage reduction; the 10 k is 1/16th of the total resistance (including its own: 10 + 150 = 160, equals 16 x 10). So if a full scale reading from an analog port is 1023 (assuming 10 bit resolution), this will represent 4.096 V times 16, or 65.536 V. So you can multiply the reading by (65.536/4.096)/1023 to get a floating point value in volts.

[Northland]

I have written a lot of code for use in the AVRs (chip in Arduino)
I write in Bascom its much more stable than the Arduino it has even passed space flight testing!
Working on web server system to control my systems at the present so I am solving the interfacing problems now if your interested.

Yes I'm interested. I assume bascom writes to the mega the same?

Would I be better off buying a pi than trying to solve all these arduino problems?

Maybe I should state my goals:

1. Have the "computer" detect battery voltage and adjust an SCR to increase the load (duty cycle) of hot water element
2. Have the computer read the weather report and adjust the water heater thermostat
3. Log input and output power for 30 days via a webpage or android app. Maybe water temp also

[Northland]

This measuring voltage method maybe isn't best if using Arduino. Instead, I could use a shunt and measure the 0-75mv to see what current is going into the battery. Then make the Arduino decrease the resistance across the SCR (how?....are we just increasing current? Can Arduino control current?), which will increase element current, which will decrease current going to the battery.

For simplicity, let's say we want zero current going to the battery and all available capacity to go to the element (until water hot enough, then recharge battery). There will need to be a sample rate & delay to avoid flip flop. Every 5 sec?

Is 75mv (max...probably spend most time around 5-10mv) too small for the Arduino to work with?

Yes I know I could use comms to query the pip for charge current. I'm a better sparky than a coder...

Can the Arduino detect water temp? I assume it's just reading a resistance across the probe. So is the thermostat controller redundant?

[paulvk]

?

Yes I know I could use comms to query the pip for charge current. I'm a better sparky than a coder...

Can the Arduino detect water temp? I assume it's just reading a resistance across the probe. So is the thermostat controller redundant?

Bascom writes code for all AVRs including the XMega its only a matter of changing a one line in the code.

I already have the code written to get the data from the PIP it displays it to a LCD and even puts it on a web page stored on an mini SD card.

Temp measurement can be done with a DS18B20 sensor I have also the code for this I intend to make a small hole in the outside case of my hot water tanks and insert the probe so it touches the tank then seal around the cable

Logging data is also simple just write the responses from the PIP to the SD card which could be called up via the web interface.

Note I have also the code written to turn things on/off via the web interface and to read things that are on/off

The AVR (Arduino) is better suited to the task than the Pi as it has all the on board hardware for this, runs code only for the tasks and not a full operating system full of stuff we do not need that may cause stability problems.

You will also need a W5100 shield which gives the network access and mini SD card socket but these are cheap on ebay as are the DS18B20 cables ready made

But what Arduino do you have?

[Northland]

I did state above but I think you missed it. It's a mega, though I also have some uno's. I also have a 16 relay board if required. Yes I think you are right - arduino is better suited for the job.

The mega has 8 comms pins so assume I don't need a network shield? The SD card socket and sensor +3m are on their way.

I spent a lot of time over the weekend learning about the pi. It doesn't seem simple at all. You have to install one software, to install another so that you can install the operating system. Then I have to install and learn another language for code, and there were many options so just deciding on which wasted a few hours. And I figured I would get the ball rolling by ordering 2 on ebay plus 2x micro SD cards.

But all is not lost. My wife's office computer is an all in one, not a huge draw except that it often turns itself on or stays on after she hits shutdown. On day's like today with rain and soc at 81% - having a computer that uses 5w instead of 150w, and only when in use - could make a big difference. And the other pi?...a media centre.



MPP discussion on the pi forum