Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.
LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408
Posted: 08:15pm 28 Jan 2019
Copy link to clipboard
Print this post
Hello!
I stumbled across this forum and will spend this whole weekend devouring the very interesting posts but I have 3 questions that everyone else around me says is just crazy talk:
I want to build a "PV Direct" interface (inverter?).
The behavior of this interface is as follows:
i. It prioritizes power from the PV panels using the grid as fallback - so if the PV panels can provide only P kWh, but the load needs L kWh where L > P, then L - P is sourced from the grid. Is this interface (inverter?) possible? From my basic understanding, it seems like all GTIs are built to behave exactly like this?
ii. It does NOT feed back into the grid. It consumes power FROM the grid ONLY when needed and diverts any excess to dump load(s) NEVER to feed it back
iii. This interface supports load sharing: it's possible to connect these units in series or in parallel to power loads that exceed the individual capacity of each unit by itself
The most important part of this interface is that it does not require batteries to be hooked up to the system to work (if possible).
Most PV interfaces on the market I see are assumed to be hooked to batteries and are designed that way. This interface cannot make that assumption (unless having a battery is absolutely necessary to even have a shot at this interface being a reality)
Is this all crazy talk?
Are any of these thoughts possible or is this all an unrealistic dream stemming from my ignorance of power electronics?
ryanm Senior Member Joined: 25/09/2015 Location: AustraliaPosts: 202
Posted: 11:46pm 28 Jan 2019
Copy link to clipboard
Print this post
Sounds like a standard grid tie inverter with a zero export device fitted. You need them in Australia on certain connections where the grid can't handle the excess power. They might be called something different in the US.
LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408
Posted: 11:59pm 28 Jan 2019
Copy link to clipboard
Print this post
Thank you.
Maybe in the U.S. they call them GTI with limiter?
Zero export sounds more correct - does any GTI come with that feature or that's an additional feature?
I ask because an inverter has to dump power somehow unless it's designed to handle 0 output load
Ralph2k6 Senior Member Joined: 24/09/2017 Location: AustraliaPosts: 129
Posted: 12:00am 29 Jan 2019
Copy link to clipboard
Print this post
Or something along the lines of Warspeeds very clever arrangement. Although more complex than getting something off the shelf. Spend enough time looking around here, some very clever setups have been designed and (b)logged here.Ralph
zaphod Regular Member Joined: 03/06/2018 Location: United KingdomPosts: 93
Posted: 07:00am 29 Jan 2019
Copy link to clipboard
Print this post
No that is not correct, the inverter will only draw from source (PV) the load*its efficiency. A PV panel has a maximum output at any given illumination but if the load is lower than the maximum there is no problem.
I have a 1Kw system with a no export setup and it works perfectly.Cheers Roger 1Kwp DIY PV + Woodburner + Rainwater scavanger :)
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 02:44pm 29 Jan 2019
Copy link to clipboard
Print this post
Yes its possible and has been done, I am running a system like that myself right now.
The inverter needs to work over an unusually wide input voltage range, that is not a problem, but the inverter needs to be designed to do that right from the very start.
At dusk or during really bad weather the solar output obviously falls, and so does the solar panel voltage under any kind of significant load. So you need additional dc power to run the inverter from at night. That can come from either the usual bank of batteries or from a grid powered dc rectifier.
I have three dc power sources isolated by three diodes, and the inverter runs of whichever voltage is highest at the particular time. During a normal day, solar easily provides all the required power for the inverter, at a fairly high voltage plus a sufficient surplus to recharge the battery.
At night the grid powered rectifier (can) provide all the night time power if its switched on. The rectifier dc output voltage is arranged to be a just few volts higher than the maximum battery voltage when its fully charged. No power can ever escape back onto the grid via the rectifier !
If the rectifier is switched off, power at night is drawn from the battery.
Even before full sunrise, the solar system may be contributing just milliamps of current, which very quickly swells high enough to take the load right off the battery during the day. Its a gradual seamless power sharing changeover at dawn and dusk.
On a very cloudy day, solar will struggle, but any power surge demands such as refrigerator start up for example, will be readily provided by the battery.
My system has 6Kw of solar panels, a 5Kw inverter and 5Kwh of Lithium cells. My night time load is about 1.4Kwh in summer and 3.5Kwh in winter. Its not that large a system compared to some others here. But it could easily be scaled upwards.
A third future power source will be from a 5Kw home brew gasoline/natural gas generator I am currently putting together. This will have both a 5Kw high voltage dc output, and 5Kw of ac mains voltage available. This will give me a lot of extra options to play around with once its up and running.
My system uses no solar controller. Its one less thing to go wrong.
People have criticized my system because its less efficient than an super software controlled MPPT system. So what is better a 100% efficient system with a high dollar solar controller, or a crappy 80% efficient system that requires maybe only a very few extra low cost solar panels to make up the lost 20% ? Dollar for dollar probably about the same, but a much simpler system with fewer parts has to be more long term reliable.
Cheers, Tony.
tinyt Guru Joined: 12/11/2017 Location: United StatesPosts: 438
Posted: 08:06pm 29 Jan 2019
Copy link to clipboard
Print this post
This is similar to the last step of my year long project, but it is not direct from PV, still needs a battery-inverter combination. The PV, batteries, inverter, charger and load are all hooked up. Still learning how to code in arduino sketch to control it.
Unfortunately, I am too busy with other things right now.
Warpspeed's simple idea is probably my next project if I finish my first project.Edited by tinyt 2019-01-31
LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408
Posted: 09:59pm 29 Jan 2019
Copy link to clipboard
Print this post
Thank you tinyt! Did you document it anywhere?
I would really like to work with you on the arduino sketch. I am very comfortable with arduino. I used PIC18F before that.
I currently use ESP32 based arduino and experimenting with the blue pill.
Yes!
Why?
OK, I just wanted to clarify that I want to draw power at the same time from more than one source, not just one, even if has the highest at the particular time.
OK, but in my case, I want to prioritize power from the battery with any slack taken from the grid in a load sharing fashion.
OK, but in my case, I dont have a battery so the grid has to be my battery and I want to minimize the power drawn from the grid.
That is very interesting! Did you document it anywhere?
I don't think there are any consumer grade MPPT systems out there that are more than 80% efficient from PV to load output, so if your system is 80% efficient from PV to load output, I would REALLY like to read about it
Where do I read more! Give me more!
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 10:02pm 29 Jan 2019
Copy link to clipboard
Print this post
Its not that complicated, but the required operating voltage ranges need to be very carefully thought through at the initial design stage.
I decided to build a 100 volt nominal system rather than the more usual 48 volt system to keep all the dc currents down, and the overall efficiency high. Not having a solar controller also helps with overall power efficiency during the day.
I discovered that thirty 50Ah Winston Lithium cells, when physically stacked 6x5 fit very nicely into a standard filing cabinet drawer, with just a few mm to spare all the way around after some 6mm steel clamping plates have been fitted. The height is ideal too.
So my battery voltage operating range will be 30 x 3.1v minimum (93.0v) And 30 x 3.45v maximum (103.5v). Four 24v solar panels in series produce around 140v full open circuit voltage in a clear blue sky. Maximum power voltage for these is specified as being 120v.
The inverter should see a dc input voltage maximum of 140v at no load, down to 93v at the battery exhaustion end point. Actual inverter design limits were 90v to 180v input voltage range.
I am very fortunate to have all three phases available here, and my three phase dc rectifier produces a fairly consistent 106 volts dc with very low ripple.
Edited by Warpspeed 2019-01-31Cheers, Tony.
LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408
Posted: 07:52pm 30 Jan 2019
Copy link to clipboard
Print this post
Very interesting! Did you document it anywhere?
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 11:45pm 30 Jan 2019
Copy link to clipboard
Print this post
Its all here somewhere but in difficult of find fragments.
The problem is, that I have been slowly developing my unconventional ideas over the last few years, and keep thinking up better or simpler ways to do things that includes some some pretty original engineering.
Its now all complete as far as I am concerned, working fine and I see no reason for any further development of either the inverter or the battery management system. I have been off grid for several months now, with so far perfect reliability.
Writing it all up would be a huge task, but I am planning to slowly drip feed Clockman with all the gruesome details. He can then pick through my ravings, and possibly present some of it as a sort of supplementary chapter to his excellent book.
I had thought of starting a new thread on the Forum, but the details would very quickly be diluted with other peoples comments, and reading through dozens of pages would become a pretty hard slog for anyone that is interested. So far there has not been enough interest in it to motivate me into action.
The consensus seems to be that PWM works, uses fewer parts and is simpler and cheaper to implement. That is true if high power PWM can be build without the continuous blow ups that most first time inverter builders seem to invariably experience.
My system requires combining the outputs of either three or four very basic low frequency square wave inverters. More individual parts, more custom wound transformers, and more work, mainly in the transformers.
But is far less critical of layout and noise problems, and discounting major "duh" type mistakes, it is vastly more robust and blow up free. Build it once and it will almost certainly work very first go !
Big difference between switching at 23.5Khz and switching at only 50Hz. No critical primary chokes required either. Performance wise there is nothing to choose between high frequency PWM and low frequency multistep approach. Idling current and efficiency should be about the same if you put the same amount of effort into the transformers.
PWM is probably better at low power, but as the power rises, PWM is much more difficult to scale upwards. Many people here can get a four mosfet bridge to switch cleanly at PWM frequency quite easily. Trying to get multiple mosfets in each leg to switch simultaneously can result in a lot of tears unless its approached in just the right way.
Getting a dozen mosfets to switch together simultaneously (but much more slowly) at 50Hz is child's play. Much easier to drive and far fewer problems.Cheers, Tony.
nickskethisniks Guru Joined: 17/10/2017 Location: BelgiumPosts: 458
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 07:14pm 31 Jan 2019
Copy link to clipboard
Print this post
Very interesting Nick.
One point it does not make very clear, what happens if the grid goes down ? This is still a grid tie system, but backflow restricted.
Does it still work if you switch off the main breaker, so that any connection to the grid no longer exists ?
Cheers, Tony.
LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408
Posted: 09:23pm 31 Jan 2019
Copy link to clipboard
Print this post
Things I like about this:
1. Your appliances will take energy made from your own production 2. The rest energy is taken from public grid in case the output power of your own production will not be enough for all appliances
# 2 is EXACTLY what I want to achieve and I THINK Warpspeed Stage 1 design (which we talk about here: https://www.thebackshed.com/forum/forum_posts.asp?TID=11086&PN=1) does that too
Things I DON'T like about this:
3. The inverter switches off immediately in case of public grid failure.
That's a deal breaker RIGHT THERE.
The interface I am thinking of using NEEDS to keep working regardless of grid failure: it only uses the grid when my appliances draw way more power than my PV panels can generate.
LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408
Posted: 09:52pm 31 Jan 2019
Copy link to clipboard
Print this post
deleteEdited by LadyN 2019-02-02
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 10:28pm 31 Jan 2019
Copy link to clipboard
Print this post
I doubt if an off the shelf boost converter is going to be easy to make work.
These have a negative input impedance, and draw higher current as the input voltage falls trying to maintain a constant dc output voltage and constant output power.
Its possible to do it, but its going to require some major surgery and redesign of an off the shelf boost converter.
Its not going to be a case of buying an e-bay unit and just connecting it up to a solar panel and getting a higher regulated output voltage.Edited by Warpspeed 2019-02-02Cheers, Tony.
LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408
Posted: 11:48pm 31 Jan 2019
Copy link to clipboard
Print this post
PLEASE always make me an exception to your rule. I am used to hearing "Natasha! You are wrong!". That's absolutely ok :D
I gain nothing if you (or anyone else here) don't tell me what I am doing wrong.
Obviously I WILL be making a lot of mistakes as I don't know even a tiny fraction of what you all know but they only way I know how to learn and be better is by doing projects, making mistakes until the project does what I wanted to do in the first place
One thing I can tell you is I WILL keep on learning, until it hurts, and then some more until it all becomes clear and does not hurt anymore.
My parents could not afford to pair up a caregiver with me so I could never be outdoors and forced to be indoors I taught myself how to program. I could not even compile a program the first few weeks. I slept only 2 hours a day. Now I take part in coding competitions at school. statewide. and we win.
Understood and Up for the challenge.
I am working on proving it to you so you can see I am not wasting your time.
Will post updates here: https://www.thebackshed.com/forum/forum_posts.asp?TID=11075&PN=1
Consider my thoughts: PV Panels are current source devices. Try and draw more than Imp for the given ambient conditions and the PV Panel collapses.
So I ensure there is a microcontoller monitoring the power draw from the PV Panel and ensure the convertor never draws more than Imp so panels don't collapse.
The output regulation of such a boost converter is going to be poor, but this boost converter is not being designed to power a load directly.
This boost converter will be hooked up to 1 - 3 PV Panels to draw MPP from them.
That is it.
We will have a cluster of these boost converters feeding a downstream inverter that tries to achieve output regulation and dips into the grid when required to maintain quality output regulation.
Ok, so back to the equivalent resistance of the BOOST convertor being negative
http://users.ece.utexas.edu/~kwasinski/EE462LS14.html Look for "Boost converter class notes (ppt file)" under "Week 6 documents:"
If you click on it, you will get the file _7_EE462L_DC_DC_Boost_PPT.ppt
From page 16 onwards they talk about Impedance matching and very next slide about PV and BOOST
As you can see: the computed equivalent resistance of the BOOST convertor is positive unless Rload can go negative.
So my questions are:
1. Is their derivation of the equivalent resistance of the BOOST convertor wrong? 2. IF NOT, under what condition does Rload go negative?
I also saw an article that tries to explain why the equivalent resistance of the BOOST convertor is negative: https://www.element14.com/community/docs/DOC-65505/l/anp008-negative-input-resistance-of-switching-regulators
However, I did not understand their point although I understood the utexas document.
Thus, Can you explain or send me links to understand negative input impedance bits?
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 12:57am 01 Feb 2019
Copy link to clipboard
Print this post
It depends on how the feedback is arranged.
Usual application of a boost converter is to step up some dc source to a higher voltage which we wish to tightly regulate. So we do not care about the input voltage (as long as its sufficient) and we wish to tightly regulate the output voltage for our purpose.
A system like that would draw twice the current if the input voltage falls by half in order to keep the output voltage constant.
What an MPPT system attempts to do is modulate the loading on the input for maximum possible power output. But what if there is zero load on the output to absorb all that power ? The output voltage just rises uncontrollably.
The solution is to use conventional voltage feedback to control the output voltage in the usual way, but have a system that overrides that when the input voltage tries to fall below some set threshold voltage.
It should work similar to a current limit on a bench power supply, where the output voltage is regulated, but when it reaches current limit the output gets pulled down to maintain a constant current beyond that point.
In this case we want a regulated output voltage, but instead of a current limit, we have an input voltage limiter that allows the output voltage to fall but not the input voltage.
I have built a system like that, still have circuits and circuit boards for it around here somewhere for three 2Kw boost converters that had 220v nominal input and produced +/- 240v dc output, voltage regulated. There were eight 24v solar panels in series powering each boost converter. Open circuit input voltage could almost reach 300v but the input voltage was prevented from falling below 220v if I remember correctly.
At dawn the solar voltage would spring up to 220v pretty quickly, and the output current would slowly creep up from zero once it reached 220v dc input. With no load, or very light load the output voltages were tightly regulated to +240v and -240v dc.
The output stayed regulated unless the load became too heavy for the solar panels to support. The outputs then fell away below +/- 240v, but the input could never be pulled lower than 220v which was the maximum power point.
I tossed all that out because I wanted to add a battery, and as it was, I would have needed 480v worth of cells to produce +/- 240 a thought which terrified me even if I could have afforded to do it.
So I split my solar array in half to two strings of four in series, and settled for 120v maximum power point from the panels and a 100v battery.
Cheers, Tony.
nickskethisniks Guru Joined: 17/10/2017 Location: BelgiumPosts: 458
Posted: 02:53pm 01 Feb 2019
Copy link to clipboard
Print this post
It stop's working if the grid is down. I use it do dump energy and do capacity test of my battery with it... It's easy to set minimum voltage of a dicharge test, you can also set a begin voltage. You can also limit the power or current manualy.
LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408
Posted: 07:05pm 01 Feb 2019
Copy link to clipboard
Print this post
Yes, so that the boost converter has good output regulation, which one would typically like from any PSU (but not for this PV one. Not the first stage atleast).
Or it could halve the current output while keeping the output voltage constant.
Yes, the output regulation would be VERY poor but it is not stressing the input much by trading off regulation.
Do we agree on this? I don't want to be wrong on this and you know everything to do with this and I DON'T
Are you saying that under 0 loading, a boost convertor's output voltage starts to rise uncontrollably and that leads the input voltage to sag?
Ok, the TL494 has a ULVO ability where it will turn the convertor off if it makes the input voltage fall below the ULVO threshold.
That is because the TL494 does not have the smarts of a microcontroller.
A microcontroller could direct the TL494 to increase its duty cycle to ensure reduced loading of the input.
OK. So why don't we build a LEARN_MODULE "3INPUTS" with the following specs:
1. Supports 3 inputs, all isolated from each other (I am assuming they are isolated to reduce the complication of the inverter design. If this assumption does not help reduce complexity, let's not make this assumption): i. Rectified DC grid power: Say 12V DC with a tight tolerance (10%) ii. DC from battery: Say 12V DC with a lower voltage cutoff but otherwise a 10% tolerance iii. DC from solar/PV panel: Say 12V DC with terrible tolerance (No load of 20% over nominal, in this case, 15V, with a similar 20% below nominal, in this case, 10V)
2. Output of 24V DC (or 48, 2 or 96: whichever is easier) with "good" regulation. 3INPUTS achieves regulation by dipping into the above inputs according to the following priority: i. PV panel: we always draw MPPT from our PV panels ii. battery: this is a small battery meant to provide backup when PV panels are absolutely not available. As a result, we dip into this only when PV panel power is below a few watts. If PV panel power above a few watts (but not enough to power a load), we do not draw from this source. I am almost driven to treat PV and this battery input as one input that's routed depending on time of day or weather. However, if supporting this complicated logic severly complicates our design, let's relax this constraint iii. grid: any slack between load demands and PV panel is drawn from the grid. We always want to draw AS LITTLE AS POSSIBLE from the grid Bottomline: we don't want to draw a lot from the battery specially when PV panels provide juice. We always want to draw AS LITTLE AS POSSIBLE from the grid. Our main source of power will be power from the PV panel
How do we go about designing this? I am not afraid of introducing multiple microcontrollers into this design. They can talk to each other to share information
Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.
Print this post
