12VDC to 24VAC inverter

This is the place for any magazine-related discussions that don't fit in any of the column discussion boards below.
Post Reply
rf burns
Posts: 6
Joined: Thu Nov 22, 2007 2:47 pm
Contact:

12VDC to 24VAC inverter

Post by rf burns » Sat Mar 07, 2009 7:31 am

Hi all,
I have a refrigerator in my old motorhome that uses a 12VDC to 24VAC inverter to run a swing motor compressor in it. The inverter is toast and a new one is no longer available. The old one just used a multi-vibrator circuit driving some switching transistors.

I want to replace the old circuitry with a PIC driving some MOSFETS into the step-up transformer from the old inverter. I'd like to get a 60 hz sine wave (or something close) out of the transformer, not the funky looking square wave (ish), using PWM or something.

I'll use the PIC to monitor the refrigerator temperature and compare it to the set-point, also power-down the fridge if the battery voltage drops too low as well other useful functions.

My experience is nil on PWM. Can anyone give me some guidance on how to generate PWM in the PIC and get something close to a sine wave out of the step-up transformer.

Thanks!

User avatar
reloadron
Posts: 519
Joined: Sat Jun 28, 2008 8:57 am
Location: Cleveland, Ohio
Contact:

Re: 12VDC to 24VAC inverter

Post by reloadron » Sat Mar 07, 2009 1:12 pm

You may want to read THIS LINK.

I am guessing you have something like that and the suggested remedy may be the best approach.

Ron

rf burns
Posts: 6
Joined: Thu Nov 22, 2007 2:47 pm
Contact:

Re: 12VDC to 24VAC inverter

Post by rf burns » Sat Mar 07, 2009 1:53 pm

Thats exactly what I have, mine is a 1977 GMC.

I have both modified and pure sine wave inverters lying around my office after trying to make them work at remote tower sites. They do work, but they have some shortcomings. I also have this thing about converting power up and back down, it just rubs me the wrong way.

The fridge's power supply actually had a transformer with a 120v, 12v and 22v winding. 22Vac ran the swing motor in either 120v or 12V mode. In 12V mode the transformer stepped up the voltage to 22V using a multi-vibrator circuit driving some power transistors. The 120V winding was disconnected via the power relay and was not used in the 12VDC mode.

I'd like to use a PIC to monitor the fridge temperature and control the set point. The PIC would drive a pair of power MOSFETS into the existing transformer to operate the swing motor.

I would also have the PIC monitor battery voltage and display the current temperature. It would also drive a Piezo alarm to warn of low voltage or high temperature in the fridge. So there is method to my madness here.

I can do everything else, but PWM is not something I have done before.

I am not sure where he came up with 4amps at 12 volts, the swing motor is rated at 60watts. He might be using a modified sine wave inverter which isn't passing full power through the transformer.

Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Re: 12VDC to 24VAC inverter

Post by Bigglez » Sat Mar 07, 2009 5:59 pm

rf burns wrote: I want to replace the old circuitry with a PIC driving some MOSFETS into the step-up transformer from the old inverter. I'd like to get a 60 hz sine wave (or something close) out of the transformer, not the funky looking square wave (ish), using PWM or something.
If you're interested in the system level design and only
want to solve the immediate problem, then a commercial
DC to AC inverter is probably the easiest approach. Radio
Shack and others make several models.

OTOH, if you're interested in making your own inverter,
there are several options and many pitfalls.

The basic idea is to convert DC to AC, and use a transformer
to get the required output voltage. Most likely 12V DC in and
120V AC out.

For sinusoidal output (appropriate for an induction motor
load) the transformer can operate at high frequency, be
very small, and use PWM techniques for higher efficiency.

For quasi-sinusoidal output there may be serveral transformers
in series, to create a stepped waveform; better than
simple squarewaves but not as good oe efficient as PWM
techniques. These are usually large iron coil-core types.

A dead simple squarewave inverter may suffice. That is what
was used with the transistors and vibrator in your original
design (and many tube car radios after WW-II).

To complete the project you will need to write and 'burn'
code into a uC chip, design and wind transformer(s), and
do extensive bench testing to get a reliable solution.

That trip to Radio Shack is looking more attractive...

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Re: 12VDC to 24VAC inverter

Post by MrAl » Sat Mar 07, 2009 8:39 pm

Hi there,


I have worked with synthesized sine converters in the past that put out almost
perfect sine waves, three phase and single phase. Efficiencies approaching 90 percent.

It sounds like you already have a transformer from the old unit? That could help a lot.
I am not sure how it is wound however, so you make need an H bridge instead of
just two MOSFETs.

What do you need to know about PWM? PWM really just means that the time off
and the time on are calculated to produce a given output level at a given time.
If the transistor stays on longer the output ramps up, shorter and the output
ramps down. If you PWM at 50 percent duty cycle you end up with about half
the output level that you get with 100 percent. To get a sine wave you program
in a small lookup table of sine values from 0 to 90 degrees and use that for the
entire 360 degree wave as appropriate. This means your pulse is shorter near
0 degrees and longer near 90 degrees. The opposite polarity is used for 180
to 360 but the pulses are the same. An H bridge does the polarity reversal.

To drive a motor however you probably dont need many different sine steps.
A few will be good enough because the motor isnt that picky. I wouldnt be
surprised if a rectangular pulse calculated to take out the 3rd harmonic
alone worked good enough for a motor.

It's very simple to monitor temperature and voltage with a PIC chip and provide
feedback for the whole process. Even a 50 cent thermistor works pretty well
for temperature detection.

Also, just wondering, how much is a new refrigerator instead of rebuilding the
old one?
LEDs vs Bulbs, LEDs are winning.

rf burns
Posts: 6
Joined: Thu Nov 22, 2007 2:47 pm
Contact:

Re: 12VDC to 24VAC inverter

Post by rf burns » Sun Mar 08, 2009 5:08 am

The existing transformer is your basic iron core 60hz transformer, so I can't put a high frequency switching through it. The 12V side is center-tapped so I need only 2 FETS to drive it.

I have written a number of programs for the PIC processors using PIC C from CCS so the rest of the program will not be a problem. A new fridge is 800-1000, so its not cheap, but this is more of a "I can do this" project rather than a least cost approach. I've got the PCB laid out so I want to be sure I have everything covered before I send it off to be etched.

The original 12Vdc operation simply put a square wave out the 22V winding, so that is easy to do. I could add a pause to make a "modified sine wave" and maybe that will work fine too. A "swing motor" is more or less a vibration rather than rotation so I don't think it cares too much about the waveform, but I don't want it to make much noise from a square wave (buzz)

I'll read up on the PWM function under PIC C. Maybe just one more step at 50% would work fine.

Any input on that would be great.

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Re: 12VDC to 24VAC inverter

Post by MrAl » Sun Mar 08, 2009 7:14 am

Hi again,


Here is one way to do the pattern. If you have never worked with PWM before though
perhaps you should look into some DC applications first, but anyway here are some
notes that might help with the sine synthesis...

Image

Image
LEDs vs Bulbs, LEDs are winning.

rf burns
Posts: 6
Joined: Thu Nov 22, 2007 2:47 pm
Contact:

Re: 12VDC to 24VAC inverter

Post by rf burns » Sun Mar 08, 2009 7:40 am

Ok, that looks like something I can do without even getting into the PWM function, but I'll never learn it without doing.

Thanks for the guidance

Robert Reed
Posts: 2276
Joined: Wed Nov 24, 2004 1:01 am
Location: ASHTABULA,OHIO
Contact:

Re: 12VDC to 24VAC inverter

Post by Robert Reed » Sun Mar 08, 2009 11:18 am

MrAl
Nice presentation! So it looks like with the crude resolution at that point, you obtain a sort of truncated triangular wave. Now to clean it up even further, would it require more resolving (more pulses of variable widths per sine wave cycle)? What would it take to produce a relatively clean sine wave of say less than 10% distortion?
I am not in the market to build one of these at this time, but I find your replys interesting and who knows what may be needed down the road. A little uneeded knowledge (for me at least) never hurts. :smile:

User avatar
reloadron
Posts: 519
Joined: Sat Jun 28, 2008 8:57 am
Location: Cleveland, Ohio
Contact:

Re: 12VDC to 24VAC inverter

Post by reloadron » Sun Mar 08, 2009 2:20 pm

This article may be of some interest. They use a PIC generating a square wave and run it through a Maxim 292 Bessel filter. The circuit is pretty straight forward.

Ron

rf burns
Posts: 6
Joined: Thu Nov 22, 2007 2:47 pm
Contact:

Re: 12VDC to 24VAC inverter

Post by rf burns » Sun Mar 08, 2009 2:33 pm

If you guys are just looking to generate a nice sine wave try this chip ML2036. Its a serial loaded sine wave generator, easy to use, but requires + & - voltages.

It will not work for me in this case as I am looking for a power supply circuit with good efficiency. Using this would make it an analog circuit require more power and a big heat sink.

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Re: 12VDC to 24VAC inverter

Post by MrAl » Mon Mar 09, 2009 9:54 am

Robert Reed wrote:MrAl
Nice presentation! So it looks like with the crude resolution at that point, you obtain a sort of truncated triangular wave. Now to clean it up even further, would it require more resolving (more pulses of variable widths per sine wave cycle)? What would it take to produce a relatively clean sine wave of say less than 10% distortion?
I am not in the market to build one of these at this time, but I find your replys interesting and who knows what may be needed down the road. A little uneeded knowledge (for me at least) never hurts. :smile:

Well thank you Robert :smile:

Trying to reduce the distortion is part of the problem, and what happens is we end up trying to
eliminate (or nearly so) the lower order ones like 3,5,7 and allowing the higher ones to get
through. The higher ones have to be done with passive filtering because with n pulses we
end up with those high order harmonics in the pattern itself.

Here is an example of an 11 pulse per half cycle sine synth pattern at full output.
The diagram shows the top half of the synth pattern, with the lower half being
exactly the same except peaking at -1 instead of +1.
There are 22 pulses total, so there will be harmonics around the 21st and 23rd, but
as far as the 3rd, 7th, 9th, 11th, 13th, 15th, 17th, 19th go they are extremely
low, less than 1 percent (using floating point pulse positions).
The pattern is cut back for voltage regulation in a way that keeps the ratios of all
the pulses the same...ie as the 90 degree pulse gets narrower by say 10 percent
all of the other pulses get narrower by 10 percent also.
Using integer pulse positions (simple uC type single chip controller) the distortions
may increase due to the round off errors in the pulse positioning.

This would be something cool to try out for a variable voltage sine wave
generator running at line frequency (50 to 60 Hz) for testing other things.
To get the negative pulses either an H bridge is used (four transistors) or
a center tapped transformer is used (two transistors).
The nice thing about the H bridge is that we dont need a transformer if
the dc input voltage is high enough to get the required output ac voltage.
If it is too high, the pattern can be cut back.
Some inductive and capacitive filtering would be a good idea though if the
higher order harmonics will be a problem.

Here's the drawing...

Image
LEDs vs Bulbs, LEDs are winning.

cherala
Posts: 1
Joined: Wed Sep 02, 2009 11:19 pm
Contact:

Re: 12VDC to 24VAC inverter

Post by cherala » Mon Sep 07, 2009 2:14 am

My Hotpoint refrigerator is making a loud buzzing noise when the ccoling cycle begins. Any ideas?? Two days ago, my Hotpoint refrigerator began making a loud buzzing noise - similar to a hockey buzzer - when the cooling cycle begins. It lasts about 3 to 4 seconds and then quits. The refrigerator and freezer are cooling fine. What could it be? Thanks for your time.

User avatar
Bob Scott
Posts: 1192
Joined: Wed Nov 20, 2002 1:01 am
Location: Vancouver, BC
Contact:

Re: 12VDC to 24VAC inverter

Post by Bob Scott » Mon Sep 07, 2009 2:52 am

cherala wrote:My Hotpoint refrigerator is making a loud buzzing noise when the ccoling cycle begins. Any ideas?? Two days ago, my Hotpoint refrigerator began making a loud buzzing noise - similar to a hockey buzzer - when the cooling cycle begins. It lasts about 3 to 4 seconds and then quits. The refrigerator and freezer are cooling fine. What could it be? Thanks for your time.
This is common. The fridge compressor vibrates the most at startup. Find the part that is rattling and tighten it down. In my old Admiral fridge, it was the drain tube from the automatic defrost system rattling on a loose clamp, near the evaporation pan on the bottom. It is accessable by removing the clipped-on panel on the front of the fridge down at the bottom. The evap pan should slide out the front for cleaning.

Bob
-=VA7KOR=- My solar system includes Pluto.

Post Reply

Who is online

Users browsing this forum: No registered users and 28 guests