 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. |
Forum Index : Electronics : Various aspects of home brew inverters
  Page 18 of 48   |
|||||
| Author | Message | ||||
Tinker Guru  Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Regarding wiseguy's request for input where this all should be heading, here is what I am planning: For the moment, sit on the fence until our resident programmers have perfected the 2 x sine wave method and the associated controls. I do my own PCB design so a different mosfet power board and nano Atmega 328P based control board is no problem. I had planned to go the opto isolated power/control way but looking further into that think I'll abandon that idea - too many different voltage sources were required to run the chips. Just 12V for the drive and 5V for the logic is enough. The plan is to get the 5V from an isolated DC/DC brick anyway and the 12V from another (non isolated) DC/DC converter. Using the IR2184 half bridge driver chips is the plan, they arrived today - still waiting for the Atmega 328P. High DC voltage cutout is not required, I think that belongs to battery charging management. And the inverter will run fine with a raised (within limits) DC voltage. Under DC voltage cut out is required to stop excessive currents. This needs to be adjustable for the different battery types. Not sure if this should be self re setting, if so, there should be a long delay to stop hunting. Over temperature; if the NTC value keeps dropping markedly with the fan running the heat sinks cannot cope with the load and the inverter should shut down for a while. Just like the EG8010 does, and then try re starting for a set number of times before going into lock out which requires hard re starting. I'm not sure about the push button ON/OFF approach. To me a switch that visually shows where it is sounds more logical. This also could be one that's lockable off if that feature is required. Consideration should be made what happens after the inverter is *first* connected to the battery bank. The big caps will charge slowly up (if a suitable series resistor is used) and therefore the 5 & 12 V rails will increase slowly from zero. Can the nano cope with that? The nano has 30 pins and the EG8010 has 32 so any plug in adaptor board is tricky. Much better is a new control board IMO. Presently my inverters (I have 2 working and a test version) have totem pole drivers. They work fine without the cross conduction prevents. Since the 2 x sine wave drive looks much more efficient, a different power board (using IR2184's) is what I'm aiming for. I do not need another inverter so I will design that board to replace the ones in my existing inverters. Klaus |
||||
renewableMark Guru Joined: 09/12/2017 Location: AustraliaPosts: 1678 |
Good bloody question, really though how many of us are there? Who has a running Clockman or Mad inverter that they would want to fit a new control board to? Are there any that would want to be building one for the first time and choose to use this newer control board concept and the Mad power board? Boards are pretty cheap so I probably would buy one or two and stick them on the shelf for insurance if it plugged into the existing Mad power board. BTW my Clockman control board will power the Mad power board. And the Mad control board will power the Clockman power board (didn't try it with nano side populated though) You have my number if you want to borrow anything for testing. Cheers Caveman Mark Off grid eastern Melb |
||||
Ralph2k6 Senior Member  Joined: 24/09/2017 Location: AustraliaPosts: 129 |
If it is straight forward enough, why not look at integrating into the existing Mad control board as the design has already been proven in the field. (mentioning this one as I have one waiting to be used, and it already employs the option of an Arduino for Fan/volts/thermal etc control, not all eggs in the one basket). @Tinker; As per the low voltage startup, why not a separate supply just to the Arduino itself, protected with its own little fuse. (very small size 48v -> 5v supply board or whatever) The code can allow a time delay which could allow for a bleed resistor/heavy contactor pull-in to charge up the big caps. (Button to prep, light says ok, then a contactor pulls in or 2nd button is pressed or operator then closes 2nd breaker switch, whatever). This could lead to a '2 stage' startup, with 'stage 1' being only for empty caps to gently charge up or something.. Plus and minus to this automation is the options can be so very numerous, KISS can seem so boring sometimes.. @Poida @Gaspo @Wiseguy et all. These developments have been a brilliant read, keep up the good work! (I lurk almost every day reading up during smoko breaks  )Ralph |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Ralph, I already use a small 48 to 5V supply - I'm not a fan at all of oztule's original zener voltage regulator. Forget about contactors - the inverter operating current can be over 100Amps! What I have now works very well with the EG8010. As soon as the big caps start charging up the under voltage sense trips the SCR, locking the EG8010 off. It must be manually re set to soft start the EG8010. I do that after a little green LED tells me the caps are charged to battery voltage. As this initial start up happens very rarely in normal use (but often with my experimenting)there is no point to over complicate it. Just something that turns the nano on trouble free - I do not know how it behaves with less than 5V applied. Thinking about the latter, perhaps it could be powered from 12v at the Vin pin, then its internal regulator would power it when that puts out 5V. Klaus |
||||
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
Ralph, I considered that making something compatible with existing inverters would be a good way forward. Although I might be contributing to the ideas pool, I'm not trying to drive this in a particular direction and wanted others to have a say. I think that soon an arduino nano board that is pin compatible with the EG8010 adapter board will be created or if not that an existing nano board with an interface board to make it work. I have found that new designs can be less effective when trying to make them backwards compatible as they are interfacing to existing stuff sometimes warts and all, because starting afresh you possibly would do it a bit differently. And KISS is boring but it also avoids flack on the forum, a lot of posts ago I suggested a high current solenoid approach for trickle charge of capacitors and then closing the solenoid for full current but it was not well received. There is no reason though that the current regimes cant continue to be used by existing setups and anyone wanting to try something different should be encouraged to "roll their own" and have a go & report back their success or failure. If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| tinyt Guru Joined: 12/11/2017 Location: United StatesPosts: 438 |
Just some info I read: The Atmega328 spec shows that it has an operating voltage of 1.8V thru 5.5V. At 16 Mhz clock, there is a graph showing supply current starting at around 4V. To play safe, I asssume that the arduino can reliably operate specially its timers from 4V upwards. The chip is held reset (Power On Reset) until the VCC exceeds about 1.4V. There are fuse bits for setting VCC Brown Out Detector level settings: 1.8V, 2.7V, and 4.3V. The arduino nano I have has its BOD level set to 2.7V. So, I think productive code will not execute until VCC is higher than 2.7V. The RESET start up code can check what kind of reset it is by examining the MCUSR hardware register (BORF bit). So an elaborate code can check and decide what to do for different kinds of reset conditions. I don't know if the arduino start up code is doing this checking. I don't know also if we need to bother coding for it, and if it can be done in an arduino IDE environment. Maybe at power on, after some software delay, indirectly check for high enough VCC value via battery voltage monitoring. Then if it is good proceed to the timer setup. |
||||
| Zibe Newbie  Joined: 15/01/2018 Location: AustraliaPosts: 10 |
I haven’t looked at this forum for quite a few months and its good to see all the progress. There are a lot of similarities to an inverter I built earlier in the year and has been running flawlessly. It runs on an atmega 328p, has a 2 line LCD and logs all its data over WI-FI so it can be accessed anywhere and the entire operating history is logged. Some suggestions for the design are; keep it as simple as possible, the code and the hardware. Avoid the ability to make changes to the inverter while it is running e.g. tuning etc. Rather consider making all the changes, trip points or scaling before the inverter starts. E.g. enter a start-up routine and stay in there until you have finished making changes, then proceed to finishing a start-up routine. Serial coms is good for this. As for WIFI it’s very easy to do, do all the work on the Wi-Fi chip e.g. an esp8266 and just send a data string from the micro to the ESP. I can provide the design and code for the ESP chip and integration if interested. |
||||
poida Guru Joined: 02/02/2017 Location: AustraliaPosts: 1418 |
That's great to see. Can you share the code you are using here? wronger than a phone book full of wrong phone numbers |
||||
| Ralph2k6 Senior Member Joined: 24/09/2017 Location: AustraliaPosts: 129 |
That's great to see. Can you share the code you are using here? @Zibe I second that if you could post up idea/process of your inverter if you are happy to. More the merrier for formulating a agreed upon design, yes? Ralph |
||||
| gaspo Regular Member  Joined: 25/06/2018 Location: AustraliaPosts: 65 |
A little progress with STM32 inverter running on the simplest F030K4 board. Three waveforms, power adjustment and two complementary outputs. https://youtu.be/vpGJbCOpRzY |
||||
| tinyt Guru Joined: 12/11/2017 Location: United StatesPosts: 438 |
I don't know if this draft schematic is what we want. But here it is for review/comments/etc. It is still incomplete.2018-11-15_025354_TBS-PGWdraft1.zip |
||||
| poida Guru Joined: 02/02/2017 Location: AustraliaPosts: 1418 |
This looks great. It's nice to see the 3 modulation schemes illustrated so clearly. wronger than a phone book full of wrong phone numbers |
||||
| poida Guru Joined: 02/02/2017 Location: AustraliaPosts: 1418 |
This is good to see a circuit of this thing. The 12V Vfb diode bridge has an extra join node (shorting input) I want to add a 1K pull down resistor on pin 8 (D5). This is to force the IR2184 shutdown active LOW during reboot when output pins are high impedance with undetermined outputs. I would like to have the i2c bus available for an LCD. This uses pin 23, 24, 5V and ground. Maybe on a 0.1" header? All analog inputs could have a 5V zener, limiting input voltage. The temperature sensor could be based on a 10K NTC thermistor, fed into ADC6 maybe. How about a TMP36. No good for really hot heat sinks with the max temp of 150C? TMP36 is 10mV/degC which means the 10 bit ADC will resolve approx. 0.5 degC These are noisy as, so an RC filter will be needed. wronger than a phone book full of wrong phone numbers |
||||
| tinyt Guru Joined: 12/11/2017 Location: United StatesPosts: 438 |
OK, will work on this later. Is the 15V supply for the IR2184 OK? The limits are 10V - 20V. I was thinking that 12V is too close to the lower limit. But most fans are 12V. |
||||
| tinyt Guru Joined: 12/11/2017 Location: United StatesPosts: 438 |
On the 5V zener, my experience is they have to much spread on zener voltage and are also temperature sensitive. We might also have to consider leakage current. Most microcontroller chips have diodes at their pins that clamp the input to VCC and GND. I confirm that the nano analog pins have them using DMM in diode mode. Maybe just a low value series resistor to limit clamping current? The TMP36 has multiple pins except the TO-92 package which can be mounted on the heatsink. I don't know if we need high resolution temperature sensing if we just want to control a fan. |
||||
| poida Guru Joined: 02/02/2017 Location: AustraliaPosts: 1418 |
15V is good for the IR2184. I drive them at 12V due to availability of 12V from the bench supply. I'm happy to go with anything else you suggest re. temp sense. Which ever is easiest for you? We need to prevent more than Vcc + 0.5V entering ADC pins. I have put more than that in by mistake without destruction, must have been lucky.. Clearly to my mind, the Vfb ADC input could go higher than 5V with careless use of the trimpot. Series resistor probably a good choice. wronger than a phone book full of wrong phone numbers |
||||
| tinyt Guru Joined: 12/11/2017 Location: United StatesPosts: 438 |
After reading some more on the TMP36, I found out that it might need an opamp buffer to be able to send its output without loss to the ADC of the ATmega328. It can be affected by RFI and also needs bypass capacitor(s) right at its leads. Since the heatsink we want to mount it possibly is a big 20kHz radiator, I think it is better to use the simple 2-terminal NTC thermistor. Here is the updated schematic for review/comments/etc. 2018-11-15_140441_TBS-PGWdraft2.zip I moved the 4.7 uF capacitor from the trimpot wiper to the diode bridge because in the inverter I am testing I found it difficult to select the TBD resistors to get the right voltage range at the stops of the wiper. Also, the TBD resistors can limit the max voltage at the wiper stop. Note that this portion of the circuit is just a cut and paste of oztule's design except for the connection now of the 4.7 uF capacitor. |
||||
| poida Guru Joined: 02/02/2017 Location: AustraliaPosts: 1418 |
Tinyt: Your circuit is looking great. Here is some progress in the code. It is not ready for testing by others. The calibration values have no effect, etc. But I want to show where I'm heading. I have E2PROM read & store working, as well as the i2c LCD. Expect significant changes as I work through this. Interestingly, I have dispensed with the low pass filter applied to the AC output. I wonder if we need a potentiometer to set the AC output voltage. Why not just use a number, stored internally to set the output voltage? We can change that easily via the serial console. Speaking of which:  Here is video of the LCD display, showing dummy data for DC volts, DC current, DC watts and temperature. It shows the calculated output AC voltage and above that the AC setpoint. The Fluke DMM is there to keep it honest. Notice the nice waveform. here and the code, for a good laugh is 2018-11-15_205654_uno_inverter_wg_sine_i2c_1.ino.zip wronger than a phone book full of wrong phone numbers |
||||
| gaspo Regular Member Joined: 25/06/2018 Location: AustraliaPosts: 65 |
That is a good idea, set point is really just a number. Saves one pot and 1 pin for other purpose. Poida, what is the difference between "Low AC ouput" and "AC output low limit", and also "Low DC volts" and "LV cutoff voltage"? |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Could you elaborate what you physically mean by "serial console"? Having built several of these EG8010 based inverters and two running my house I'm always interested if they can be improved. And your and wiseguy's 2 sine wave method does show its possible. Presently my inverters have an on/off switch and a reset button on top for controls. Did you mean the nano version should have an external USB port so it could be pre set even if its works are housed fully enclosed inside the inverter? You see, I'm looking at this from the practical user side, not the breadboard version. Klaus |
||||
| Page 18 of 48 |
|||||
 Print this page |
