GW Instek SFG-1013 DC Offset Issue

[ampwizard]

I just bought a cheap Instek DDS function generator - the SFG-1013 - and I noticed that the output has "Default" DC Offset of -60mV (that is the DC offset when the DC Offset pot is pressed in: its default state). I can adjust the the DC Offset manually by pulling out the pot knob and turning it but it's obviously not very precise. I would have expected the default offset to be 0. Has anyone else got one of these units?

[ModemHead]

Yes I have one of those, and the "default" offset was always way too much on my unit as well.  It was bothersome enough that I added an internal 10-turn trimmer, accessible through a hole drilled in the front panel.  The push-pull switch now selects between the regular pot and the internal trimmer.  The tweaked "zero" offset still tends to wander a bit (especially until the unit warms up) but pushing the switch in usually gets within 5mV of zero.

[ampwizard]

Yes I have one of those, and the "default" offset was always way too much on my unit as well.

Thanks, I guess that means my unit is working normally 

[tron9000]

Yes I have one of those, and the "default" offset was always way too much on my unit as well.  It was bothersome enough that I added an internal 10-turn trimmer, accessible through a hole drilled in the front panel.  The push-pull switch now selects between the regular pot and the internal trimmer.  The tweaked "zero" offset still tends to wander a bit (especially until the unit warms up) but pushing the switch in usually gets within 5mV of zero.

one of mine has a -300mV Offset. Could you show/tell that mod? Much appriciated

[ModemHead]

one of mine has a -300mV Offset. Could you show/tell that mod? Much appriciated

The front-panel offset pot is connected as a voltage divider between +15V and -15V. When the switch is pulled out, the wiper voltage is injected into the final stage negative feedback loop via R922.  When the switch is pushed in, it simply isn't connected.

The existing switch is double-throw, so I cut tracks and added mod wires to have it select between the original pot and a trimmer pot that I glued on the board.  This allows you to adjust the offset to zero when the switch is pushed in.  I drilled a hole in the front panel for easy access in case it drifts and needs to be tweaked again.

If I recall correctly, I used a 20K trimmer since it was what I had available.  Anything around 50K will probably work fine.

[tron9000]

Fantastic! Much appreciated, thanks!

How often do you have to tweak the pot, just curious?

[ModemHead]

How often do you have to tweak the pot, just curious?

Actually I upgraded to a better generator a while back, so this unit doesn't see much use anymore.  I don't recall tweaking it but once since installing the trimmer 8-9 months ago.  The offset on this unit varies by several mV constantly, so I set the trimmer to average it out to zero, and it seems to hold this pretty well, as long as you accept the fact that it's still going to drift a little in the short term.

Other thoughts: Without this mod and with the front-panel offset control disabled, the R922 input is left flapping in the breeze, so the default offset will depend on the DC balance of the final stage.  300mV seems excessive to me, so maybe the complementary output transistors are mis-matched or something?  I dunno.  I do know the output stage is biased so that it runs pretty hot even at idle.  It needs to warm up and stabilize before anything can be tweaked.  I used to use this unit to troubleshoot the ACV function of DMMs, and it would need about an hour of on-time before the amplitude would be stable enough for that task.

[tron9000]

Other thoughts: Without this mod and with the front-panel offset control disabled, the R922 input is left flapping in the breeze, so the default offset will depend on the DC balance of the final stage.  300mV seems excessive to me, so maybe the complementary output transistors are mis-matched or something?  I dunno.  I do know the output stage is biased so that it runs pretty hot even at idle.  It needs to warm up and stabilize before anything can be tweaked.  I used to use this unit to troubleshoot the ACV function of DMMs, and it would need about an hour of on-time before the amplitude would be stable enough for that task.

Good to know, cheers!