Monday, July 26, 2010

Article posting

If you have some ideas or suggestions that you think would help others in their Grid Tie endeavors, please contact me and I will add you to this blogs 'Author' list. I'm fairly sure that will allow you to publish your own posts here.

Cheers,
Sparky

Wednesday, July 14, 2010

The importance of GTI shutdown & suggestions / modifications

The PSWGT series of GTI’s all consume varying amounts of AC power from your grid when not generating. 

During standby (at night or very overcast conditions), the 300w GTI’s consume around 5-6 watts, the 600w unit consumes around 12 watts and the 1200w GTI, around 20 watts. I have not tested the larger 2000, 2500 or 3000w GTI’s but it stands to reason that the power draw of these units increases proportionately. This small power draw may not sound like a big deal, but it is a serious problem and can add up to a significant proportion of your daily generation capacity, particularly during the winter months as I will now try to demonstrate..

The following calculation is based on the 1200w GTI. You can transpose the figures for your own GTI if you have a smaller or larger unit. If you have multiple GTI’s in use, you will have to multiply the final figure by the number of GTI’s in use..The final figure may surprise you!.

GTI power draw in standby mode (at night and very overcast light conditions) = 20W
Winter months nighttime hours (no sun)         =   17-18hrs per day.
Winter power consumption  =  17hrs x 20w  340Wh per night!

Now ask yourself how long it would take for your system to produce that 340Wh’s of energy used every night by the GTI before you actually saw a net benefit?

But, it can get worse! The above example assumes good light levels throughout the day. With a small solar panel array and in very overcast conditions, the GTI may well consume more power than it generates during the day too! The GTI will not start generating power until the DC input is within the MPPT operating voltage band. This is generally 25-55v for the higher voltage GTI and probably between 12-20v for the lower voltage models. If you have a digital AC power meter inline with your GTI, (Killa-Watt style) you should not assume that those small watt figures being displayed are what you are generating. AC power meters cannot differentiate between power in and power out, so any reading less than ~20 watts is probably power draw rather than power generation.

So, now you can see the importance of reducing or eliminating the standby power consumption of these devices when they are not generating.

Option 1: (may be better for multiple GTI’s)

Use a commonly available AC timer switch on the AC side of the GTI/s. Set the switch to come on a little before the sun will be shining on your solar panel array and switch off a little after the panels begin to be shaded in the afternoon. This will eliminate all night GTI power consumption.
The disadvantage of this system is that the timer requires resetting at different times of year as the days become longer or shorter. The advantage is that it does not involve hardware modification, is cheap and one timer switch can be used for multiple GTI’s.

Option 2 (May be better for single/larger GTI’s):
Caution! This option requires internal GTI modification and should not be attempted by anyone who is not comfortable working on high voltage equipment! Make sure you disconnect the inverter from AC before attempting this one!!
Depending on the DC operating voltage of your GTI, you will require a n AC voltage rated power relay or a Solid State Relay (SSR)  for this mod.

If using a conventional relay, select one with a low coil operating current as this will consume less power from your array (higher coil impedance/resistance = less power consumption). Fit the SPST relay (N.O circuit) into one of the AC input lines inside the GTI. Run light duty DC wire from the DC/solar input to the relay coil (make sure you fit an appropriate diode across the relay coil to prevent back EMF from being generated when the relay closes). Despite the DC voltage reaching ~40V at times on my 28-55v GTI, a 24v relay seemed to handle the higher coil voltage without any problems. If you are concerned about this a resistor could be placed in series with one of the relay coil control lines.

As pointed out by 'XRing', using a Solid State Relay (SSR) may well have several advantages over a conventional relay in this situation as Solid State Relays have the following important characteristics:

Eliminates contact arcing (may protect the GTI FET's during AC switch-on).
Improved reliability (no contacts or coils to burn out)
No moving parts
Reduced current draw from the DC/Solar supply (SSR's only require a few mA on their control line)
More flexible adjustment of turn on voltage on the control lines (a simple resistor voltage divider circuit across the Solar input will permit the user to select the turn on or wake-up voltage for the SSR control lines).

SSR's have an inherent 'On Resitance' which is higher than mechanical relay contacts. This resistance produces some  degree of AC power loss, which results in the SSR generating heat at higher current levels. (hense the need for SSR's to be mounted to a heatsink). I have not actually tested the degree of power loss in this situation, so it may not be a significant factor, nevertheless it should be mentioned here. 

Conventional Relay Theory: With no DC voltage applied to the GTI input, the relay contacts are open and no AC voltage is applied to the inverter (no AC power is consumed). As the sun begins to shine on the solar array, the DC input voltage gradually rises until the relay energises and the AC contacts close (about 19v on my 24v relay). At this point, AC is applied to the inverter and normal GTI operation begins. As the relay only requires 20% of it’s rated coil voltage to maintain it’s closed state once energised, the relay contacts will stay closed if the DC voltage drops due to overcast conditions. At the end of the day, as the sun sets, the DC input voltage will slowly drop. When the DC input voltage drops to 5-10v, the relay contacts will open and AC is removed from the GTI, effectively putting it into sleep mode until the following day.

SSR theory: With a simple resistor voltage divider circuit (variable resistor on one side of the divider) across the DC/Solar input, the user can adjust the 'Wake-Up' voltage to any point between 3.5v - 30V. On the 24-55v GTI, I would suggest somewhere around 20V is a good point to turn the GTI on and off. Any higher than this and the GTI may cycle off everytime a cloud goes past the sun.

Sunday, June 13, 2010

Solar panel angle and effiency calculator

Handy calculator for working out how much money you will save on your power bill, depending on array size and location. It is also useful comparing different solar panel tilt angles for a given location and shows the variation in kWh's produced.
 
http://rredc.nrel.gov/solar/calculators/PVWATTS/version1/

Interesting looking new 1200w Grid Tie Inverter from PowerJack

Very Interesting new 1200w Grid Tie inverter from PowerJack. Incorporating an LCD display, larger case and larger DC terminal posts. Supposedly has 'bigger' mosfets and transformer etc.


Intersting that it retains the same PSWGT-1200 part number as the older version.

Excellent price of US$250. Will be interesting to see if this is an introductory price or if it stays at this after July.

Available in 110v or 220/230/240v versions.

Hopefully, they will produce a 50-100v DC input version too.

Ebay seller saying that these units will ship 1st July.

Replacement Filtered Mains IEC Socket

Another Modification for reducing EMI / Radio Interference, is to replace the original mains IEC socket with a filtered socket. The socket below was purchased from Ebay for 0.99 cents (plus $1.99 for postage).
After the original socket was removed, the hole had to be filed out slightly as the new socket is about 1mm wider and higher than the original. The new socket is also screwed to the case rather than clipped in. Fitting this filter will reduce the amount of switching noise that enters and is radiated by the mains wiring .


As for the wiring, simply cut the plastic plug and earth terminal off. Connect the active, neutral and earth wires up to the other mains socket as per the original socket.

Tuesday, June 8, 2010

Power Jack Grid Tie Inverter RF noise suppression mod

PowerJack EMI Filter 

The Power Jack 600w Grid Tie inverter produces a high level of broadband RF noise when operating.  I would be very surprised if the smaller 300w or larger 1200w+ models didn't suffer the same shortcomings. The noise is caused by the high frequency switching circuits used to convert the DC supply to high voltage AC. In more expensive inverters, the input and output circuits are filtered to prevent this noise from being radiated by cabling. Unfortunately, the PowerJack inverters provide no such filtering and so much of this noise escapes the inverter cabinet and is then radiated by the solar panel and mains wiring. As the solar panel wiring in particular can act as an efficient antenna, the noise can cause significant interference to AM and shortwave band radio reception for quite some distance around your house. The attached circuit provides individual low pass filters for both +ve and –ve solar array lines. This circuit will prevent most of tthe broadband noise from entering the solar array wiring. In addition, 100-150v 10kA varistors have been added on the filters input terminals to offer a degree of protection from high voltage spikes that may occur from nearby lightning strikes. It is important to note that the varistors will not protect the inverter from a direct lightning strike and they will offer no protection unless the filter case and inverter are earthed via a dedicated earth stake (the mains earth wire in the IEC plug is not directly connected to the inverter chassis or earth post - ref text in original article). The varistors are optional and do not affect the filter characteristics. Not shown in the circuit diagram, is a clip-on/hinged , 30mm square, ferrite core (see below)  that is attached to the Positive and Negative leads that run between the filter and inverter input. This core reduces the level of radiated interference further still.
Inside layout of Solar/DC side noise filter. Cable on right goes to inverter input (note clip-on ferrite core). Solar panels connect to the red & black terminals on the left. The green Earth wire connects to the inverters earth stud. Capacitors and varistors (if fitted) are soldered to brass screws near the input & output terminals.

Critical aspects of this circuit:
Ensure that the filter grounds (capacitors and varistors) are well bonded to its enclosure and the enclosure is then connected to the earth post on the inverter. Without this connection, the filter will do nothing! Keep the DC lines between the filter and the inverters input as short as possible! Preferably, no more than 150mm. Longer interconnect leads will result in higher levels of RF noise being radiated.

Inductors: 25-30mm OD powdered-iron core with 7-8 turns of insulated 3-4mm wire. 3-4mm wire is required to reduce DC resistance losses or voltage drop at high current levels.

Capacitors: 0.1uF @ 100v. 
Varistors: 100-150V @ 10kA varistors (optional). 

Enclosure: This circuit is enclosed in a suitable diecast aluminium enclosure (120x95x50mm works well). The filters are ‘grounded’ to the enclosure and enclosure is earthed to the inverters ‘Earth’ post and to an external earth stake (earth stake required only if varistors are installed).
After construction, ensure all terminal posts are tight and check for minimal voltage drop between input and output terminals under load. With 36v @ 3 Amps of current flow, my unit induced a voltage drop of only 30mV on each power rail (60mV total).

Monday, June 7, 2010

Power Jack Grid Tie inverter initial observations
Construction quality:
These inverters are built to a (very low) price. When you remove the lid and take a closer inspection of the circuit, you can see the evidence of this. Critical components such as power diodes and Mosfets are held against the case/heatsink with a single strip of metal that do not compensate for varying component size tolerances. The result of this is that some diodes/Mosfets are clamped tightly to the case whereas others sit loosely, not benefitting from effective heat transference. Several owners of the smaller, 300w inverters have reported Mosfets and diodes failing in these units. It is likely that these components have failed due to overheating caused by poor heatsinking and or load sharing. There is a mod for the Mosfet/diode mounting issue that I noticed on a popular video sharing site that looks like it might be effective, involving small adhesive rubber pads being placed between the Mosfets/diodes and the metal clamping strip. When I first activated my 600w inverter, I was getting an electric shock each time I touched the inverter case and earth. I place a multimeter between the inverter case and the household earth connection and got a reading of 160V AC! The problem turned out to be a small piece of aluminium swarf that was shorting one of the power Mosfet tabs to the inverter cassis. The unit worked fine in this situation but obviously the QA process in the factory doesn’t check for chassis isolation. The lack of any EMC filtering on the input and output is also a serious issue with these devices – see next paragraph! Before you hook this thing up to AC and if you feel comfortable opening the case, I suggest that you use a multimeter to confirm that all diodes and Mosfet metal tabs are electrically isolated from the case.

Electromagnetic Interference (Radio Noise)!
These inverters are about as RF dirty as you can get! If you are an Ham Radio operator or you live in the vicinity of one or even if you just enjoy listening to AM or Shortwave radio during the day, you will probably want to consider using some additional filtering on the DC and AC sides of the inverter. Hook these inverters up to the AC start generating and you will be generating wideband radio interference across a wide frequency spectrum for two blocks! Compounding the problem is that the solar panel wiring in particular acts like an efficient antenna, radiating the noise for all and sunder to hear. I am in the process of developing some ferrite/capacitor based low pass filters for the solar panel side, which should help. As for the AC side, I will try investing in a replacement EMI filtered, IEC style, AC socket to replace the cheap original socket. Hopefully this will tame the unit. Stay tuned for more info..

Rated Output:
It is probably safe(er) to assume that these Chinese grid tie inverters are  very overrated on their published output power figures. The 600w Grid Tie inverter I refer to here may be able to provide up to 600w of peak output (downhill, with a tailwind!), but  it cannot produce this kind of power for extended periods of time without overheating. Based on this assumption and the expeiences of others, it is probably best not to exceed 400w worth of solar panels on this device. With 400w worth of panels connected to the input, the AC output power should be around 320W at it’s peak during the summer months. In many countries, summertime ambient temperatures can reach 40+ degrees C and derating the unit in these conditions would be prudent course of action if you want to obtain a reasonable service life.  Tests  from other users indicate that efficiency drops considerably anyway once AC output exceeds 300w. If the smaller 300w inverters have similar characteristics, I would suggest not running any more than 200w worth of solar panels on those. 

Input Operating Voltage (20-55v DC?)
Well no, in actual fact the input/solar operating voltage band of these inverters is actually quite narrow and requires that the user carefully select the configuration of his solar array. On bench tests, these inverters do not start generating meaningful power until the input DC/solar array voltage exceeds 28 volts! In practice, this requires that parallel strings of two series connected 12v or single 24v panels be used. Either arrangement should provide an open circuit voltage of around 40-42v DC and a maximum power point of the solar array (under load) of around 36V. A string of three 12V panels in series, while being advantageous as far as DC wiring losses are concerned, would produce an open circuit voltage of  >60V and would therefore potentially exceed the rated input voltage of these inverters. The big electrolytic filter capacitors on the input circuit of the inverter are only rated at 63V, so you can see why the manufacturer has nominated 55V as the maximum input voltage.

Efficiency
Bench tested efficiency of these units appears to range from about 85-91%. Efficiency figures vary depending on input voltage and current and are to be expected with any grid tie inverter. It is likely that the specified maximum efficiency of 91% would only occur with solar array voltages ranging between 34-40V DC, which is where this inverters MPPT circuit seems to operate the best.

Power consumption
I know you are asking yourself ‘Power Consumption’, what is this guy talking about? Well, for some reason my 600w inverter does consume AC power when it is not generating i.e at night and under very poor light conditions. According to my inline AC power consumption meter, this inverter consumes around 9 watts of energy when not generating power. For the unit to actually start generating power, it must have at least 28v applied to it’s Solar input terminals. With only a small solar array connected and under very poor light conditions, it is possible for this unit to actually consume more power than it generates! These inverters will consume 108Wh’s of power every night (12hours). Assuming you have ~200watts worth of Solar panels connected up to the inverter, the first hour or so of low elevation morning sunlight will simply be replacing what the inverter itself has consumed the previous night. If you live in a region that does not see a lot of regular sunlight and you are only planning on putting up a small solar array, you might not end up with any net financial gain at all, particularly after the costs of the inverter, solar panels and associated hardware are taken into account.. Food for thought..?
Although another 600w inverter owner reckons that his unit does not consume any power during standby.. Go Figure..

Safety & Earthing
Interestingly, the earth pin on the inverters mains socket is not directly earthed to the inverters chassis or circuit board.
An earth wire does come off the socket and connects to the Printed Circuit Board, but from there, it simply connects to chassis via a series .001uF capacitor. It would seem as though the only function of this earth is to shunt (some of the) RF noise, generated in the inverter, to ground as it offers no personal protection against internal shorts etc.
The inverter chassis does have a dedicated earth terminal. In my installation, this is connected to the metal cabinet that encloses everything and then an earth cable is run to a nearby earth stake driven into the ground. I’m not sure if this improves the safety of the unit, but it may help down the track with interference filtering. There are issues here with balanced, un-balanced and grounding of neutral lines in AC circuits that I need to put some more thought into..

Earth Leakage questions:
This issue applies to any ‘plug-in’ type Grid Tie Inverter.
Many homes nowadays contain RCD’s (residual current devices) or ‘ground fault circuit interrupter’ sensors on their various mains circuits. These devices compare the current flowing in the Active and Neutral lines of the house wiring. If there is a disparity or imbalance of more than 5-30mA of current flowing in these lines, the RCD immediately shuts the power off. These devices likely prevent the electrocution of many people every year. However, if a circuit, normally protected by an RCD, has a grid tie inverter connected to it, will that RCD still activate if an earth leakage fault occurs? If a child for arguments sake, sticks a metal object into a mains socket, some of the current flowing through the body would be coming from the RCD protected mains circuit and some, I assume, would be flowing from the inverter. The RCD will only shut the power off once the residual current reaches 5-30mA, however, a similar amount of, if not more, current could be flowing from the inverter. Granted, the Grid Tie inverter will shut down (anti-islanding) once the RCD shuts down, but this may not occur, since not all the leakage current would be flowing through the RCD.