GW-Instek GFG-8020H Teardown and Repair (Help)

[TimNJ]

Hi guys,

I bought this function generator off eBay for $75 shipped. It's not very special by today's standards, however, it didn't cost very much and I was looking for something that was more usable than the bare-board DDS function generator that I have. It was advertised as "works" and for the most part it does. However, it get's a little finicky in the higher frequencies, and I would expect it to perform correctly all the way up to its upper limit (2MHz).

Here it is


Four screws and you're in. It's an all analog design. LOTS of LM741s, some 74HC logic, and a scattering of discrete transistors. Pretty old fashioned, I guess. Quality of construction is above average. Soldering quality looks very good. Nothing caught my eye as being a horrible design flaw. The wires may look a little hectic but they're not too bad (looks worse in photos). The only real gripe I have is that they used Jamicon electrolytic capacitors. While there is worse out there, I don't think Jamicon are particularly good. It appears to have a 2004 date-code, so it's only around 10 years old.


Now to the problems...It seems to work pretty well all around, but once you start pushing the frequency, some weird stuff starts happening. For example, at its upper limit of ~2MHz, here's a sine wave.


Here is a negative trough. There is jitter both at the top and the bottom. Kind of looks like ringing?


And here's a "square wave" at 2MHz. Poor rise time AND fall time. Looks like some sort of RC problem to me. Perhaps the power supply is struggling, or some capacitor can't keep up?



The other problem is the offset adjustment. If you don't touch the offset adjustment pot, it creates good looking waves. However, if you pull out the pot (adjustment mode), it gets weird. In fact, the offset never really changes. It kind of just dances around, but the offset stays at ~0V. As you adjust the pot, The wave goes from completely flat (not shown), to a clipped waveform (1st picture), to a normal-ish looking wave (2nd picture), to another clipped waveform (3rd picture), and then flat again (not shown). What gives...

This first photo is with the offset knob pointed towards "- offset".


This one is with "0 offset".


And this is with "+ offset".



So I basically have two issues. One with the poor rise/fall time and jitter on the bottom. And the other issue is with the offset adjust. Any help is greatly appreciated. I don't have too many ideas on my own yet, though I feel like capacitors might be a good start with the first issue. The offset problem has me a little perplexed. Thanks!!

The manual is attached. I was looking at the distortion figures. It says <1dB between 100KHz and 2MHz. What would that look like? i.e. Is the distortion noted acceptable according to these specs. I still do not think the square wave is acceptable though.

[PA0PBZ]

I think you had your scope set to AC coupling when trying the offset (it is almost visible in one of the pictures). Try again with the scope on DC.
As far as the quality of the waveform, this is about what I would expect, except for the flattening when using the offset. Maybe you turned it too far and it just limits in the output amplifier.

[han]

The limit is there, i have see several unit, none of them capable of clean 2MHz square wave.
they using inadequate TTL converter, i will find the datasheet if i remember the type.
and for the unperfect sine wave, you can tune the trim pot in the board, but i can't remember witch one.


One trim pot can make the signal missing, so you should mark the original position before tuning the trim pot.

[TimNJ]

I think you had your scope set to AC coupling when trying the offset (it is almost visible in one of the pictures). Try again with the scope on DC.
As far as the quality of the waveform, this is about what I would expect, except for the flattening when using the offset. Maybe you turned it too far and it just limits in the output amplifier.

Wow derp  ...late night. I think you must be right. It was on AC coupling.

The flattening would happen when the waveform gets clipped because its greater than the +Vs or less than -Vs.

Thanks.

[TimNJ]

The limit is there, i have see several unit, none of them capable of clean 2MHz square wave.
they using inadequate TTL converter, i will find the datasheet if i remember the type.
and for the unperfect sine wave, you can tune the trim pot in the board, but i can't remember witch one.


One trim pot can make the signal missing, so you should mark the original position before tuning the trim pot.

Thank you for the input. There are two or three pots which look like candidates. The rest are small cheap one that don't look like they're meant to be adjusted. I'll see which one i think it is. Which part of the circuit would it be in?... if you had to a circuit block a name.

[Fraser]

The manual containing circuit theory and a block diagram is here:

http://www.ece.rice.edu/~jdw/data_sheets/gfg8020h.pdf

and a cleaner copy may be here:

http://www.artisantg.com/info/PDF__496E7374656B5F4746475F38303230485F4D616E75616C.pdf

Aurora

[TimNJ]

Thanks! Yeah putting the scope in DC coupling fixed the "problem". Turns out offset adjust is working just fine. As far as the other weird shapes, looks like that might just be a limitation of this generator. Still curious about the little glitch that appears on the trough of the sine waves.