GW Instek SFG-2004 synthesised function gen

[BloodyCactus]

Got a SFG-2004 off ebay for cheap. Hooked it to my rigol to see how it goes and hmm.. TTL output gives a nice square wave... output of the sine/triangle etc on the 50Ohm output, the rigol gets all confused and says "<15hz", it cant seem to find the non-ttl output.

I went through the manual. Set it to sine, 1Mhz, and nothing. took it apart, didnt see anything obviously wrong, thought Id take some pics while I'm inside its guts.

It is interesting so much is unpopulated (being the base model), I assume same motherboard for higher end models, just populated. Even the case has all the cutouts etc.

[BloodyCactus]

man still cant figure out how to get anything other than the square wave to come up on the scope

[saturation]

Some thoughts.  These units are built similar to my SFG-1003.  A common fault with function generators is user's overloading or damaging the output amplifier.

Do you get your set output frequency from the TTL out, ie.., DC- 4MHz?  TTL out is a separate pathway from the 50 ohms output; if the square waves are stable and respond well to controls it means the DDS module is functioning.  The TTL voltage output is fixed ~ 5V; the DC offset controls, and amplitude work on the 50 ohm output only.

If TTL works, then its likely the 50 ohm output amp.  Trace the BNC jack to the board and follow it and it should lead you the amp. 

If the TTL output is random or uncontrollable, its a much more involved troubleshoot.

[BloodyCactus]

cheers. the ttl square wave plays fine it seems. I can change the freq, spin the dials and watch it change on the scope fine. Guess I need to trace the 50ohm line back to wherever it dies.

[Mechatrommer]

an lower ISO and an tripod could help a lil'bit with the pictures. just saying.

[saturation]

Great!  That means there is a high chance of a successful repair.  Also as you trace it back keep an eye out for simple things, like broken connections, shorts, or faulty BNC jacks or joints in the 50 ohms output, its unlikely in an Instek device, but its easy to find and fix. 

cheers. the ttl square wave plays fine it seems. I can change the freq, spin the dials and watch it change on the scope fine. Guess I need to trace the 50ohm line back to wherever it dies.

[BloodyCactus]

So that heatsinked fellow towards the top has continuity across all pins, the heatsinked fellow below has no continuity on any of its pits to itself, so I'm guessing the top bloke is the dead one.

and of course, no schematic to be found anywhere guess I'll have to desolder it and try and get the heatsink off without destroying it and hope for the best.

[SeanB]

Likely a 2N2219A and a 2N2905A. Replace both at the same time, and check the diodes and resistors around them as well. For better life on the replacements space them off the board on 3 short insulating sleeves ( or get the right PCB spacers for the package) so they can be cooled better.

[saturation]

Yes, no schematic is expected.  But it appears you are on track for a high probability successful fix.  I would guess one of those amps is the TTL, and the other is the 50 ohms output.  Before going through all the work of desoldering etc., I would confirm that which one is indeed the fault unless you've tracked it to the BNC port.  Its best to ID the precise chip involved and probe the input and output pins to be quite sure because if the fault is also beyond the amps, it could be a more involved repair.

You can also load the TTL output and see which of those heatsinks warms up more, and that's the amp that probably isn't broken.   

and of course, no schematic to be found anywhere guess I'll have to desolder it and try and get the heatsink off without destroying it and hope for the best.

[cheeseit]

I know that this thread died long ago and that necromancy maybe frowned upon here but it helped me repair an ISO-TECH GFG-2004 I bought a while ago. I don't know if the GFG is an earlier model than the SFG, a rebadge or they just changed the model name but RS once sold it under the ISO-TECH name. Date codes suggests that my unit is from 2005.
Anyway, my unit had a working TTL output too but the other output was bad with a strange distortion and offset. Nothing apparently burned or broken on the PCB so I searched and found this thread. Decided to desolder the output transistors and they were indeed dead, and were a pair of 2219A and 2905A as SeanB suggested. Replaced them and the unit is now working as expected.

Figured that I'd confirm that it is 2219A/2905A in case anyone else finds this thread in the future, and thank the OP and SeanB for this easy repair.

[saturation]

Good to hear.  Reviving old threads isn't bad on electronics forums since the gear mentioned could be many decades old and the discussion is still relevant to that item.  Enjoy.

[singapol]

All these function generators have their roots from this Tektronix CFG-280 especially the output amp.
Notice that the drawing's title block says Tektronix Hong kong.

http://elektrotanya.com/tektronix_cfg-280_function_generatir_schematic.pdf/download.html