I'm fooling around with a simulation package (don't tell my wife), hoping to scale the SOZ to headphone size. Even with the stock circuit, I noticed when I drive it from an unbalanced source the output waveforms aren't balanced. I thought I could just leave one input open (actually grounded through the input resistor R11), but it seems not to work.
Driving from a balanced source works fine. Does the SOZ require a balanced source? Or could it be the simulator?
Driving from a balanced source works fine. Does the SOZ require a balanced source? Or could it be the simulator?
Try taking it direct to ground with no resistor. The input resistor is there to appease the gate of the mosfet but you might not need it with it grounded . Also , try grounding the other one instead. The article states "ground the input of one mosfet for unbalanced operation. Good luck🙂
Having gone back and looked at the schematic ( imagine that ) leave the resistors in place and ground the input direct to ground . In a differrential circuit its possible to drive it bal or unbal but in unbal mode the unused transistor still does something and it must be given a voltage reference even if it is 0 ( ground=0volts ). Hope this helps🙂
Paul,
It's not the grounding, it's the resistance under the tail of the thing. For a differential to be a good phase splitter (i.e. taking an unbalanced input and creating a balanced output), it needs a lot of resistance underneath. The why we'll leave for another day when I've had more sleep.
If--purely for experimentation purposes--you increase the values of the resistors underneath the output devices, the waveform should become more symmetrical. However, you'll lose gain and such, so you're not going to want to actually build it that way.
You've got two choices:
1) Put a decent phase splitter ahead of the SOZ. The BOSOZ should do a fair job.
2) Put a current source under the SOZ (see the SOZ w/current sources thread). Current sources, at least in the ideal sense, have infinite resistance. They are very good at making differentials behave as phase splitters. Performance in the real world won't be perfect, but should be better. You'd probably still want to put a BOSOZ in front of the SOZ. And, yes, current sources in the BOSOZ would make it an even better phase splitter...but we'll leave it there for now.
Grey
It's not the grounding, it's the resistance under the tail of the thing. For a differential to be a good phase splitter (i.e. taking an unbalanced input and creating a balanced output), it needs a lot of resistance underneath. The why we'll leave for another day when I've had more sleep.
If--purely for experimentation purposes--you increase the values of the resistors underneath the output devices, the waveform should become more symmetrical. However, you'll lose gain and such, so you're not going to want to actually build it that way.
You've got two choices:
1) Put a decent phase splitter ahead of the SOZ. The BOSOZ should do a fair job.
2) Put a current source under the SOZ (see the SOZ w/current sources thread). Current sources, at least in the ideal sense, have infinite resistance. They are very good at making differentials behave as phase splitters. Performance in the real world won't be perfect, but should be better. You'd probably still want to put a BOSOZ in front of the SOZ. And, yes, current sources in the BOSOZ would make it an even better phase splitter...but we'll leave it there for now.
Grey
Thanks, Grey. It's somewhat ironic that I'm actually starting with a balanced source (from a preamp in another room), converting it to unbalanced so I can control the volume, and now it looks like it would be best to convert it back to balanced again.
I noticed your current source design requires 2 more FETs.
So much for simplicity...think I'll put this on the back burner for a while and go with something simpler.
I noticed your current source design requires 2 more FETs.
So much for simplicity...think I'll put this on the back burner for a while and go with something simpler.
- Status
- Not open for further replies.