Fast output buffer for function generator

Thread Starter

hrs

Joined Jun 13, 2014
520
Hi,

I'm working on a function generator that should produce up to 1MHz triangle and square wave outputs that I would like to be buffered. I've attached a discrete buffer. Can we do better than that without making it much more complex? Would some low value emitter resistors improve the buffer or is it not needed? Perhaps there are better topologies?

Alternatively there appear to exist current feedback amplifiers that I know little about. Would such an amplifier be a good buffer here? That would cut down the component count by a lot. There once was the CLC5633 but the only source I found has prohibitively expensive shipping. Are there others that might work and match the vintage of the other chips in the circuit? Is a video amplifier (e.g. NE592, TL592) just a CFA specifically designed for video?

Furthermore, for reference I've attached the unfinished function generator schematic as it stands today.
 

Attachments

Thread Starter

hrs

Joined Jun 13, 2014
520
What is the max voltage and output current?
The circuit will be power by 10V. The current, maybe up to a few mA? I don't have an intended application for it right now, just that it would be fun to make and could be a useful tool in the lab once in a while.

[edit]: I imagine to use it primarily as a voltage source.
 

schmitt trigger

Joined Jul 12, 2010
2,027
Could you have a bipolar supply instead? That would eliminate the output capacitor and allow very low frequencies to pass un-attenuated.

Also, how fast do you require it to be? What is the desired output impedance?
 

jjlarkin

Joined May 30, 2025
7
Hi,

I'm working on a function generator that should produce up to 1MHz triangle and square wave outputs that I would like to be buffered. I've attached a discrete buffer. Can we do better than that without making it much more complex? Would some low value emitter resistors improve the buffer or is it not needed? Perhaps there are better topologies?

Alternatively there appear to exist current feedback amplifiers that I know little about. Would such an amplifier be a good buffer here? That would cut down the component count by a lot. There once was the CLC5633 but the only source I found has prohibitively expensive shipping. Are there others that might work and match the vintage of the other chips in the circuit? Is a video amplifier (e.g. NE592, TL592) just a CFA specifically designed for video?

Furthermore, for reference I've attached the unfinished function generator schematic as it stands today.
Try BUF602. Nice part.
 

Thread Starter

hrs

Joined Jun 13, 2014
520
Could you have a bipolar supply instead? That would eliminate the output capacitor and allow very low frequencies to pass un-attenuated.
The lowest frequency of the generator is 10Hz. So if I change the output capacitor to 10uF that would be 1.6 ohms. I think that would be fine.
What would it entail to make +5V/-5V from a 12V wall wart?

Also, how fast do you require it to be? What is the desired output impedance?
When I make a triangle by adding cosines I need to add a lot of them to make the triangle look nice. From that I guestimate around 50MHz BW for a 1MHz triangle.

I don't really know what impedance, but if I'd drive lets say a 10k load then it shouldn't attenuate the signal too much. I think, perhaps incorrectly, that if I connect the output straight to the pot wiper that would be bad.

I also added the 47 ohm resistor at the end because for driving high impedance loads it won't matter much, but should I ever want to drive a 50 ohm load, it is matched. But then I guess I need more like 50mA output current.

Try BUF602. Nice part.
Thanks, I'll have a look.
 

Thread Starter

hrs

Joined Jun 13, 2014
520
BUF602, LMH6559, those are seriously fast. I see that TI calls LMH6559 "current feedback" buffers. Maybe they're a bit lipstick on a pig next to the LM318 and LM319. Or maybe I should find more important issues to obsess about.

Also thanks for that link, Ian. I'll play with that in LTSpice for a bit.
 
Last edited:
Top