Page 1 of 1

Oscillator with 50% duty

Posted: 08 May 2019, 08:08
by gesari
After receiving a Gigatron kit I finished assembling it yesterday late at night. I built a fully socketed one, and this allowed me to do some experimentation regarding its clock oscillator.
First, replacing U1 by a 74HC04 I got a near perfect square wave for the clock. With this clock, and without the delaying capacitor, C3, I could run the Gigatron at 8MHz with a 70ns SRAM (A cypress part, not the 55ns one provided with the kit).

So the easy way to get a square wave in the clock is to replace U1, but what can we do if a 74HCT04 is already soldered without a socket?
Well, a square wave can be achieved by changing the resistor network:
oscillators.gif (13.47 KiB) Viewed 2150 times
The idea is to desolder the 1Meg resistor and to connect it to the two terminals of the crystal, and to add another 1K2 resistor to ground.
These are the waveforms obtained:
osc2.jpg (80.35 KiB) Viewed 2150 times
osc1.jpg (81.25 KiB) Viewed 2150 times

Re: Oscillator with 50% duty

Posted: 09 May 2019, 18:26
by marcelk
Thanks that will come in handy! I never got the 74HC04 to start oscillating reliably BTW.

Re: Oscillator with 50% duty

Posted: 09 May 2019, 22:22
by klf
You should use 74HCU04 for an oscillator, not 74HC04. A 74HC04 actually consists of three inverters in a row, therefore it cannot be biased into linear mode (that's what the 1Meg resistor does).

Re: Oscillator with 50% duty

Posted: 10 May 2019, 21:43
by gesari
Yes, I know about the 74HCU04. It is well suited for linear amplifiers (as does the 4069U). But when it comes to oscillators the problem with the HC04 or the HCT04 is to oscillate at some other frequency than the crystal (a buffered inverter with a feedback resistor is a ring oscillator already). This is what looks strange to me: Marcel says its HC04 doesn't star, but I had no problem at all with an HC04 in the Gigatron oscillator. I made it oscillate with several crystals (2.048MHz, 4MHz, 6.0MHz, 6.144MHz, 6.25MHz, 8MHz, 9.04MHz, 10MHz), and even with some coils in the place of the crystal, and I always got a nice square wave. (For 6.0 to 8.0 MHz I also got my monitor in sync. At 9.04 or 10MHz the Gigatron LEDs stopped working)

Here, a photo with a homemade coil, running at about 7MHz...
coil1.jpg (124.24 KiB) Viewed 2096 times