Designing a Function Generator (ICL8038)

[EddieC]

Hi, I have just began to design of my first function generator, which I'm going to build around an ILC8038. I would really appreciate your help on this project, and I hope others can also use this thread as a guide to make their own function generators.

I've been looking for a design on the internet for a few weeks, but so far I can't find any that I like; none of the designs have all the features I'd like, and I could only find one design with a display for the frequency. So, as I can't design my own circuits -since I'm just a beginner- I decided to put together various parts from a few designs with the features I like... and, this is where the problems began.

So far I made a simple prototype in a breadboard with just the core of the function generator -which seems to works pretty well-, and now I'm trying to attach another circuit that will display the frequency. This last circuit is based on a CD4047 multivibrator and a series of 7-segments displays through their respective CD4026 decade counters... well, this circuit:



But the problem is that when I connect the circuit for the display to the function generator, the nice and clean signal that I got before, and which now goes to pin 1 of the first decade counter, gets interferences from the pulse that the CD4047 multivibrator generates to measure the frequency. Even if I follow that diagram exactly -including the function generator on the top-right corner- I will continue to get interferences on the output from the display circuit.

I would really appreciate if anyone could help me solve this problem. How can I avoid getting interferences between both circuits? I was thinking of using something like an optocoupler, but I'm not sure it's the best way of doing it or even if it will work at high frequencies. Which would be the proper way of doing it?

By the way, can anyone explain me the function of D1 in parallel with R7? I think I understand what C11 does -limit the reset pulse sent through Q not-. But why does it also have a diode and a resistor in parallel with ground?

Thanks,

I will post the other parts selected for the final design later on in this thread; in case anyone is interested.

[FrankBuss]

Add decoupling capacitors to all ICs, from +V to GND, near the ICs. There are even sockets with built-in capacitors. 100nF should do it.

If this is not sufficient, add a ferrite bead in series to +V of ICL chip, or a low pass filter with a series resistor (maybe 10 ohm, could be higher if the IC doesn't need much power) and a big capacitor to GND, or even a separate 7812 power supply just for this chip.

I guess the diode prevents negative voltage at the reset inputs. The negative side of the AC pulse after the capacitor is grounded. And the resistor is a standard pull down resistor, to avoid floating inputs.

[dannyf]

But the problem is that when I connect the circuit for the display to the function generator, the nice and clean signal that I got before, and which now goes to pin 1 of the first decade counter, gets interferences from the pulse that the CD4074 multivibrator generates to measure the frequency.

Sounds like it is bleeding through to the input - weird. You may want to make sure that it is not radio interference - pulling the input pin up / down will help identify the issue.

I would put a serial resistor, 1k - 10k, on the output of the signal generator so see if it helps address the issue.

By the way, can anyone explain me the function of D1 in parallel with R7?

Clamping on the negative side. Really, it is not needed as those gates have protection diodes that do that too.

[EddieC]

Thanks guys, I was planning on testing all your suggestions one by one until I solved the problem, but it seems to be solved with just the first decoupling capacitor next to the CD4047. I can't believe it was that easy; you have no idea how many hours I've been checking and rechecking everything.

Now I have another problem... the display only works with triangular waves; with sine waves it doesn't display anything but zeros, and with square waves it sometimes displays random numbers and sometimes zeros. I can bypass that problem connecting the display only to the triangular wave -while the other 2 would still be available on the output- but I would like to know if there's any way to solve that problem. I'm mainly asking this question because I'm planning on using a dual power supply with -15v and +15v, so that I can get an AC output, and I already kind of tried it and it gave me a lot of problems; the first being that the display doesn't detect anything... the second is that the negative side of the wave is cut in half by the multivibrator's pulse. Anyway, I'll looking into that later on; now I'm wondering why it doesn't detect all waves.

[Rasz]

You can check out this polish project for some tips
http://www.elektroda.pl/rtvforum/viewtopic.php?t=2649919
translated: http://translate.google.pl/translate?client=opera&hl=en&ie=UTF8&oe=UTF8&u=http://www.elektroda.pl/rtvforum/viewtopic.php%3Ft%3D2649919

[EddieC]

Thanks Rasz,

That's definitely going to be useful; as I already experienced some of the problems described there.

I found a design here http://electronicseverywhere.blogspot.com.es/2012/06/versatile-200khz-function-generator.html, which I'm planning on using for the main section of the function generator, and which accounts for many of the problems described on that website with the square wave and the output amplitud.

[dannyf]

with sine waves it doesn't display anything but zeros, and with square waves it sometimes displays random numbers and sometimes zeros.

The output from the signal generator may not have sufficient swing to trigger reliably 0 / 1 on the digital gates.

You can use a dedicated amplifier (a transistor for example), or a comparator (or even a ST gate) to trigger the frequency display unit.

[EddieC]

with sine waves it doesn't display anything but zeros, and with square waves it sometimes displays random numbers and sometimes zeros.

The output from the signal generator may not have sufficient swing to trigger reliably 0 / 1 on the digital gates.

You can use a dedicated amplifier (a transistor for example), or a comparator (or even a ST gate) to trigger the frequency display unit.

I thought that might be the reason, so I guess it will be solved once I add the output gain amplifiers.