Saturday at home, installing an inverter

[tec0]

I got hold of my old electronics books and sure enough i have something that might actually work and it will not cost me anything close to what we are being charged commercially. I just need to proof concept and if i don't set my workbench on fire or end up extra-crispy i will share the build. Oh this is going to be fun playing with 220 Volts and a Amp potential of almost 8 Amps and up...

yea... some proper PPE will be a must...

[AndyD]

This is so you get the so called time of operation on the brochure. Following this route means that after approximately 30 cycles, your battery is dead.

I have had a problem with one of my customers, precisely because of this. He reckons the China Mall Solar inverter and charger is better than mine, because it runs for 2 hours longer than mine. What he does not understand is that my unit cuts out at 11.5V thereby prolonging the battery lifespan, as opposed to the competitive unit cutting out at 10.5V. I am now waiting and hoping for him to come back after his battery is dead and reverse his statement.

The Axpert inverters don't really give time of run figures in their brochure, the only time they give these kind of figures is in the owner/installation manual which contains a quick guide to suggested battery bank requirements but this isn't sales type info.

i see the axpert default cut off dc voltage is 21 volt on the 24 volt 3kva. the 48 volt 5 kva default setting is 42 volt. i wonder why they set it so low.

The default setting is too low for my liking as well and I always set them higher. I did mention this issue somewhere else on the forum, maybe even on this thread (I haven't searched) but it's just one of the perameters you've got to set up when installing them. Also that cut-off voltage is being measured when the batteries are under load and I'm guessing it would be an uncompensated value so the figures in my last post won't necessarily apply.

[HR Solutions]

I have just purchased my very own inverter - works like a dream.
tec if have the time to "build" your own inverter, good for you - I look at it entirely different. Secondly if you build you own inverter and your house burns down or something happens, insurance will certainly not cover you !

[Justloadit]

I got hold of my old electronics books and sure enough i have something that might actually work and it will not cost me anything close to what we are being charged commercially. I just need to proof concept and if i don't set my workbench on fire or end up extra-crispy i will share the build. Oh this is going to be fun playing with 220 Volts and a Amp potential of almost 8 Amps and up...

yea... some proper PPE will be a must...

There are many circuits on the web, but all small stuff, and most of them are square wave, and probably involve a transformer which is driven at 12V and steps it up to 220V.
When you talk the large inverters, usually 2KW upwards, they tend to be pure sinewave, as there is no saving in the electronics by going modified sine wave. Whilst the H bridge to generate the sine wave from a DC link of 400V is relatively easy to get right, as opposed to converting/boosting the battery voltage from 12/24/48V to 400V link becomes another mission. Efficiency at these high power levels starts biting hard on your available battery energy. The biggest cost factor is the switching devices on the low end, along with the isolating transformers, along with respectable rectifying diodes which can handle the relatively high frequencies that one wants to use to reduce and improve the cost on transformers. This is where the heat of any inverter lies.

Anyway good luck and have fun. When you finished, you will see that the prices may not be so outrages as thought on first glance.

As a matter of interest, I am currently looking at FETs which can switch 46A at 1200V at frequencies above 500KHz, with switching on times of 25nS and switch off times of 42nS and an on resistance of 85milli ohms. The only issue here is that they cost over R1K each. For the techies this is really a super new device. Lets just do simple calculation, lets say that we are using it in a 3Kw inverter, which translates to approximately 13Amps flowing through the device when it is on, so loss or power = Amps^2 * resistance.
13*13*.085 = 14.3Watts. Now in heating terms, this will be a very small heat sink, because heat is to the power of 2 when one needs to dissipate this heat. Lets take a current FET which would work in the same environment as the FET discussed, which has an internal on resistance of 427milli ohms. doing the same calculation as above,
13*13*.427 = 72Watts. The heatsink required to keep the device cool will be very large with a fan. This is almost as much energy a modern LCD TV uses.

[tec0]

As a matter of interest, I am currently looking at FETs which can switch 46A at 1200V at frequencies above 500KHz, with switching on times of 25nS and switch off times of 42nS and an on resistance of 85milli ohms. The only issue here is that they cost over R1K each. For the techies this is really a super new device. Lets just do simple calculation, lets say that we are using it in a 3Kw inverter, which translates to approximately 13Amps flowing through the device when it is on, so loss or power = Amps^2 * resistance.
13*13*.085 = 14.3Watts. Now in heating terms, this will be a very small heat sink, because heat is to the power of 2 when one needs to dissipate this heat. Lets take a current FET which would work in the same environment as the FET discussed, which has an internal on resistance of 427milli ohms. doing the same calculation as above,
13*13*.427 = 72Watts. The heatsink required to keep the device cool will be very large with a fan. This is almost as much energy a modern LCD TV uses.

yea my thinking was basically using a none conductive oil cool... Might work right... "same type of oil used to submerge PCs in. Think it is mineral oil not sure i have notes on it on my server will have a look get back to you on that. I know it is said that it doesn't do a good job at cooling stuff but that was the idea...

FETs which can switch 46A at 1200V

do you have fact sheet on them if you don't mind please.

[IMHO]

Very interesting review on an ammeter/volt/watt meter. At R480 local pricing the Turnigy will have to do! Will it work on a lead acid battery, as it is really aimed at RC type batts, like Lipo etc.?

[IMHO]

With respect to measuring the battery capacity, it is possible, but you require some very expense electronic equipment. I have been looking for one for a while, and found one at RS. ACT Battery Tester

At R5300 incl. you're right, it is damn expensive, wish I could buy one! It gives you only one thing and that is capacity left in the batt, but that is what I so desperately desire!

[Justloadit]

There is a number of new devices coming out. Have a look at this one C2M0025120D

Be aware of the on and off switching requirements, +20V and -5V for proper operation.

[tec0]

The part that get me the most is that from tomorrow we may actually be without power for 8 hours a day...

[IMHO]

The part that get me the most is that from tomorrow we may actually be without power for 8 hours a day...

[ATTACH=CONFIG]5655[/ATTACH]

[tec0]

@ IMHO

On the 5th of June, some municipalities that couldn't pay Eskom on time will have be without power. The info i got from our local paper is 6:00 AM to 10:00 AM and then from 5:00 PM to 9:00 PM every day... Now tell me that doesn't suck. When you wake up no power.... When you go to bed no power... Our municipality is yet to make a statement we will see how that goes.

[IMHO]

Only if you live in one of these two places.

Free State residents living in Harrismith and Parys and surrounds are in for hours of blackouts daily, due to their municipalities' Eskom debt.

Ngwathe and Maluti-a-Phofung municipalities' electricity will be cut tomorrow at 10am, unless they reach an agreement with the power utility.
Eskom’s spokesperson, Khulu Phasiwe says the two municipalities collectively owe close to a billion rand.

Phasiwe explains that Maluti-a-Phofung owes over R600 million, and Ngwathe over R300 million.


Cheers!

[tec0]

Well according to official word it is they still need to pay in R94million by next week or will pay it i am not sure. Problem is without proper official statements it is impossible to figure out if the sword is hanging over our heads or what is really happening. It is sickening that Eskom to date made no statement towards our area and our municipalities said they are doing there best but that R 97 million needs to be paid over.

However, a decision by Eskom on when they will start to introduce the bulk interruptions remains a misery. We simply don't know and they simply cannot be bothered. So it is a "are the lights on today" scenario each time we wake up.

If households and business fail to pay for whatever reason the situation may well go towards the fan...

sorry couldn't get my animation to work

[smitty]

Which municipality do you reside in?

Do you realise that Eskom has been in talks with all the municipalities that owe it money?

Why isn't your municipality keeping you informed with regard to the outcomes of those talks?

Why do you think that Eskom should be communicating with you instead of the municipality who is their customer?

Have you asked your municipality if they're paid up with Eskom?

Have you asked your municipality if they're going to be cut off for non payment to Eskom and failure to agree to pay back what they owe?

Why do you think that Eskom (and by extension its paying customers) continue to supply your municipality electricity for free?

Bottom line is:
It's your municipality and they are supposed to render services to you for which you pay them. If you're paying for electricity and you're not getting it because they're not paying their supplier (Eskom) for it then you need to give your municipality hell.

[IMHO]

Sjoe Smitty, maar jy is kwaai so vroeg in die oggend? haha