Function generator - which one to choose?

[Ilya]

I know, I know. This has been discussed here at some time. However, I couldn't find any meaningfull info, so I'd like your input, gents.

I'm looking for a basic function generator. It should have the following:

• CLEAN square wave; triangle and sine don't matter
• Frequency: 20Hz-20kHz
• Level: from 0.1V p-p
• Variable output impedance (50R min)

I believe that the level can be adjusted with a simple voltage divider, and output impedance can be set with external series resistor. Please confirm this.

Old units is not an option because they are too heavy (and expensive to ship) and bulky. I was considering something like this:

http://www.ebay.com/itm/25MHz-Dual-channel-DDS-Arbitrary-Waveform-Function-Signal-Generator-Counter-Kit-/400976855749?hash=item5d5c1542c5:g:-UEAAOSwd0BVzDYs

Does anybody have any experience with these type of generators? Any suggestions?

P. S. I've been playing with the NE555 chip for almost two full days trying to get nice and clean square wave. No way. This thing has awful overshoot and I can't get straight horisontal lines of the square wave no matter what I tried. I'm fed up with that.

 

[Bo Deadly]

If that function generator does what it claims, that looks pretty good. My only concern would be ground noise. But for $57 USD you could just buy it and find out.

 

[Ilya]

Actually, here's even better one (from the same company) that claims to have overshoot <=0.5% 

http://www.ebay.com/itm/Digital-Dual-channel-Function-Arbitrary-Waveform-Generator-200MSa-s-2Mhz-P5E7-/291566385361?hash=item43e2b5f8d1:g:VlgAAOSwQPlV-iqD

The main objection is time. It takes approximately 30 days to arrive.

 

[benb]

I can't think of why a NE555 would have overshoot, but I can see where a too-long ground connection would have too much inductance and cause ringing and such. Is your oscilloscope probe ground wire connected to the ground AT the 555?

 

[Bo Deadly]

Actually, here's even better one (from the same company) that claims to have overshoot <=0.5% 

I cannot help but wonder if putting a bunch of digital stuff in a little box like this would limit noise performance. Specifically, if it's powered by USB from a PC and you connect that ground to the circuit your analyzing, I would bet 256 bytes that you'll get digital noise.

Ringing has to do with the source resistance and inductance and capacitance of the load. So a 555 circuit could be made to work very well if you add some parts like a choke, cap and resistor in the right place. Where and what exactly I would not know without analyzing the circuit.

Oscilloscope probes usually have a little trimmer capacitor builtin to the base of the BNC connector. There's a square wave test connection on the scope and then you adjust that trimmer for minimal ringing.

 

[Ilya]

I can't think of why a NE555 would have overshoot, but I can see where a too-long ground connection would have too much inductance and cause ringing and such. Is your oscilloscope probe ground wire connected to the ground AT the 555?

I tried grounding the probe at the PS and 5 cm away from the chip - no difference. I'm on proto board, but this shouldn't screw the circuit THAT much.

I cannot help but wonder if putting a bunch of digital stuff in a little box like this would limit noise performance. Specifically, if it's powered by USB from a PC and you connect that ground to the circuit your analyzing, I would bet 256 bytes that you'll get digital noise.

Yeah, not going to USB power it. External wall-wart only.

Ringing has to do with the source resistance and inductance and capacitance of the load. So a 555 circuit could be made to work very well if you add some parts like a choke, cap and resistor in the right place. Where and what exactly I would not know without analyzing the circuit.

I agree, but that square wave doesn't look right at all. Maybe I'm doing something stupidly wrong though...

Oscilloscope probes usually have a little trimmer capacitor builtin to the base of the BNC connector. There's a square wave test connection on the scope and then you adjust that trimmer for minimal ringing.

Of course probes are adjusted. That's the first thing I did.

The whole thread moved from the function generators topic to the 555 square wave generator quirks. I'll probably open a new thread for the 555 stuff...

 

[JohnRoberts]

If you want to make a clean 50% duty cycle square wave, run the 555 up at 2x and then divide by 2 with D type flip flop.

JR

 

[Martin Griffith]

What about the Exar 2206, if its still available? http://www.vwlowen.co.uk/arduino/xr2206/XR2206-function-generator.htm

And Analog's audio DSP stuff, but it looks a bit OTT for a simple function generator