Hi All,

I am working on a wave generator program that outputs 0-3.3V on the analog pin (p18) on the LP1768. The generator works reasonably well, but I would really like this signal to be between -3.3V to +3.3V. I could use a capacitor to shift the signal down to -1.6V to +1.6V, but the load impedance and the input impedance of the DAC seem to be greatly reducing the signal amplitude. The other thing to try is an OpAmp voltage follower (a buffer):

http://www.muzique.com/lab/buffers.htm

but I don't want to use any external power supplies for V+ and V- of the opamp. I noticed that there is a +5v out of the MBed. Is there any way to reverse the polarity so that it's BOTH +5V and -5V? I'm using this on a speaker, so I would like both the + and - reference.

I'm very much a novice with electronics, so this question might not be the right way to think about this.

Hello,

You cannot simply reverse the polarity. Aditionally the 5V output of the mbed is simply the USB power directly, so no USB, no 5V.

Also just an opamp buffer won't cut it, an opamp can't deliver enough current for a speaker. You can include a simple class-A amplifier, but then you might as well just buy an audio amplifier. There is also no real reason to have symmetric power supply, a capacitor will do the trick also. Or even better, an H-bridge setup can be used for maximum voltage swing over the speaker. So I would just search for an audio amplifier IC, preferably with H-bridge output.

Apologies, I wasn't clear in the first message. I'm using an audio amp to drive a number of speakers. My test rig works well with a function generator, so I'm just trying to control the speakers programmatically through the Mbed. I did use the capacitor trick, but found that when I connect the MBed analog out with a capacitor to the audio amp, the signal amplitude is greatly reduced. This doesn't happen when I just connect the analog out to the amp without the cap, but I do want the voltage swing in the full +/- range to properly drive the speaker. I assume the cap issue is because of the large impedance mismatch. Never thought about an H-bridge setup (don't know what that is to be honest) but I'll read up on it. Let me know if you have any other thoughts.

Thanks!

You could do this in hardware, use something like this to create a -5V rail and then use an op-amp to shift and amplify the 0V to 3.3V output from the mbed to -3.3V to +3.3V to drive the audio amp.

Which audio amp are you then using and with which supply? Generally they won't even accept input voltages outside their own supply, so shifting it to a negative supply doesn't exactly help. Also quite some of them have automatic input bias supression, or at the very least have an inverting input you can put at a bias voltage to compensate the bias on your signal.

And the capacitor should not have any influence, it is in series with the signal so as long as it is large enough it does not alter the signal, and definately does not appear as a load for the DAC.

Thanks for the feedback guys. I'll definitely order some of those converters Martin, they look quite useful.

Erik, the audio amp is a car audio amp (a Kicker ZX 400.1):

http://www.kicker.com/sites/default/files/2008%20ZX%20400%20500%20750-1%20f01%20web.pdf

The supply I'm using to power the amp is an external computer power supply (12v, 30A). The capacitor is a 0.1uF, which is in series wit the signal which is then connected to the amp. I don't see an inverting input on the amp I'm using (it's a mono bridged amp) but I'll investigate if it has auto input bias suppression.

Just checked up on this if anyone wants to know, the amp I'm using has automatic bias correction. Pretty handy. The Kicker has differential output by the way, so you can see both the positive and negative signals. I think Erik is right, most amps come with bias suppression, so just send in the +signal you want, and the amps should do the rest.