simple 6W DC to AC converter

[Pjoudislav]

Hi,
I need to design 6W 50Hz DC->AC converter 9-12V RMS preferably with least THD as possible.  I tried using pwm from arduino (around 30kHz), filtered, buffered and amplified. However, I needed to use quite a complicated filter to get rid off the 30kHz switching frequency and its harmonics.

Do you think it would be possible to build some sort of small and simple 50Hz 6W sine generator with adjustable output voltage?
Or do you have any other ideas?

Thanks for replies
Pjoudislav

[IanB]

If you want a pure sine wave output I think you could use a standard sine wave oscillator configuration and feed the output through a power amplifier stage, maybe even an ordinary audio amplifier.

[Psi]

Yeah, that's a pretty normal way to make AC power signals.

It really isn't any different to an audio amplifier, in fact it's easier because it only has to work at one frequency and the audio quality doesn't matter as much.

You can make a real easy AC powersupply using just a high current opamp with a sign wave input.
The disadvantage is that high current opamps can be expensive.

[onlooker]

I recently bought off ebay a (class D AB) "LM3886TF Amplifier Assembled Board" to toy with.

According to the datasheet, LM3886TF can reach THD<0.003%. But, I am more interested to see, from a hobbyist point of view, if it can be made stable enough to be used as an AC source for calibrating DMM (against a known good DMM).

The inputs are very stable sine waves from an LCR meter. The power supply is a Power Designs 5020 plus  a less speced old Ondor (that is all I have for now). As to the output stability, no conclusion yet. Has anyone tried similar setup (using classD amp) ?

The cost of this LM3886 board was just under $10 with free shipping. Interesting enough, the 5020 was also bought from ebay after reading an eevblog post about Power Designs PSU. It costed me just under $100 with shipping included. From the reading, the 5020 was supposed stable enough to calibrate DMM's DCV and DCI against a known good DMM.

[David_AVD]

I recently bought off ebay a (class D) "LM3886TF Amplifier Assembled Board" to toy with.

The LM3886 is not class D.  It's definitely class AB.

[Pjoudislav]

Thank you for your aswers.

Based on your suggestions, I'm going to try Wein bridge oscillator amplified with some integrated amplifier (these things are cheap and it will save me time and nerves:-))

In the attachments are Wein bridge oscillator examples from TI Application Note 263 and the whole app note as well, however, it seems a bit outdated to me.

Does anyone knows a better/simpler/cheaper circuit?

Thanks for your replies
Pjoudislav

[onlooker]

The LM3886 is not class D.  It's definitely class AB.

Thanks for the correction. It is a quasi-AB according to the datasheet. I got stuck with the wrong impression probably due to mix-up with other ebay items I looked. Actually,maybe,  this explains why it was running hotter (with heatsink added) than expected.

[Zero999]

A few questions:

1) Why? I can't think of anything which runs off 6VAC and can't tolerate a modified sine wave or even a square.

2) How stable does the frequency need to be? The Wien bridge won't be very good compared to the mains or a crystal oscillator.

3) Is efficiency important? A class AB Amplifier won't be any better than 70.7% efficient, even in a perfect world.

4) what's the input voltage? A perfect half bridge design will require at least 17V and a bridge will need 8.5V. In reality, you'll need more