DIY 8038 Function Generator Problem

[Rob11966]

Hi Folks,

I am normally found around the tube amp forums but have recently brached into the dreaded solid state to build a function generator.

I built this around the 8038 chip as described on this site -

8038 Audio Generator Project

The generator is up and running and does everything that it is supposed to do except that the signal (Sine, square and Triangle) is quite ...fat. On my analogue CRO, the trace is about 4mm thick and looks as if the CRO is out of focus (similar but not quite as bad with my digital scope). The CRO is not the problem as I have borrowed an off-the-shelf frunction generator which displays a very precise and clear line.

I can't imagine that this output would be considered acceptable. I have fiddled with all the trimpots with no improvements. Is there anyone out there with any ideas or potential solutions? I have attached the schematic and built exactly as per the description on the website.

Thanks in advance,

Rob

[MJL21193]

Could it be oscillating? Do you have supply decoupling / bypass caps close to the opamps? Not shown in your schematic.

Also, your link to the project doesn't work for me.

[Rob11966]

Hi John,

Thanks for the reply. The layout of the components is per the printed circuit board on the link. It looks like a pretty well thought out design. I am not sure what is wrong with the link, it works on my computer but I have included the URL without the http: (so you can view the link directly) - maybe this may help

//mypeoplepc.com/members/scottnoanh/birthofasynth/id32.html

I gues it could be oscillating but I to be honest I am not sure how to tell. When I look closely at the trace it looks like it is made up of lots of individual lines. In other words the thickness of the line seems to be a result of lots of sine (square or triangle) waves all slightly out of sync by a few microseconds.

Rob

[ikoflexer]

Check pins 2, 3, and 9 of the 8083, see if you have a clean trace there. If so, then at least that you get a clean function out of the IC. If not clean, then you need to check the operating parameters of the 8083, power supply, etc.

There's a test circuit for the 8083 in the datasheet, see attachment. If I were you I'd breadboard it and power it with two batteries. Then I'd change one thing at a time and continue to measure it, until you make one change that gets that thick trace again.

[DigitalJunkie]

Are there bypass near/on the pins of the output buffer op-amps?

[Conrad Hoffman]

Has to be oscillating. As said above, check that you have bypass caps near each chip. Also check your wire/trace routing to be sure you haven't created a positive feedback path to an opamp input by daisy chaining grounds. Reminds me of my first sig gen, a SWTP kit from back in the '70s. The main difference was that it used the dreaded LM318 for an output amp- without good bypassing those were absolutely guaranteed to oscillate!

[Rob11966]

Originally Posted by DigitalJunkie

Are there bypass near/on the pins of the output buffer op-amps?

Sorry, DigitalJunkie, I don't think I understand the question. The layout is extactly as per the PCB board on the website.

Hi Ikoflexor,

Good advice with respect to the output of the pins. Makes sense. At least then I can narrow it down to the power supply or rest of the circuit. It's not the IC - I have 2 8038's and have swapped them over without any change. Thanks also for the test cicuit - I will try and get one of those breadboards this afternoon.

Cheers,

Rob

[Rob11966]

Originally Posted by Conrad Hoffman

Has to be oscillating. As said above, check that you have bypass caps near each chip. Also check your wire/trace routing to be sure you haven't created a positive feedback path to an opamp input by daisy chaining grounds. Reminds me of my first sig gen, a SWTP kit from back in the '70s. The main difference was that it used the dreaded LM318 for an output amp- without good bypassing those were absolutely guaranteed to oscillate!

Howdy Conrad,

Now I am sitting bolt upright and taking note. I grounded the whole thing with a daisychain loop - a great description. I hesitate to ask the question but - is this bad? It certainly makes for a neat layout.

I would never do this with a valve amp (big copper earth bus always) but I figured that it wouldn't matter with IC's.

If you think this might be a problem, let me know!! More importantly how should it be grounded? Star?

Cheers,

Rob

[ikoflexer]

Always star ground. You may want to power it with batteries to eliminate the psu from the picture, while testing.

[Rob11966]

OK,

Here is a quick update before my wife kills me if I don't get up to bed.

The 8038 has the dirty trace on the pins so I am presuming my problem may be in the power supply.

The only unusual thing that I can see is that there is a 1KHz ripple on both 'in' pins on the VR IC3 and IC4 (PDF 2) but the ripple is only present on the 'out' pin of IC4. I have changed the elecrolytic cap (C9) associated with IC4 - no change. Also changed C6 - no change.

Is this ripple normal or is it abnormal and if so could it account for the 8038 distortion? Could a leaky C8 cause this?

Thanks again,

Rob