Homebrew 3 Channel Pure SIN Generator ( 0 ... 60MHz )

[Yucca]

Hi Peter,

I knew about the synching pulse supplied by micro, but the hardware clocks are a bit of a pain as you point out.

I wonder if micro coax transmission lines could be cut to length to get exact input clock phases from one shared clock?

I know it'll probably just result in chaos but before I remove the clocks and make a shared clock I will try (probably in vain) to tie the 3 seperate clocks together somehow.

I've just been writing hibernation procedure in the GUI. Across the 10V input supply I have a big electrolytic. I sample this electrolytic through a pot, tuned so its output is just above logic 1 level. As soon as the cap sags I have a few hundred ms to save the complete system state to EEPROM and shut down, I want to make a routine to write in different EEPROM positions each time so I don't wear out the first EEPROM positions.

[dankie]

The TPU is in the khz range .

You dont need high Mhz speeds .The 74HC series can go up to 20 mhz . Why not just go with some 555 regular square waves for such a simple task ? You are bringing a 100 pound sledgehammer for this one little nail sticking out ...

It will be alot less complicated and expensive to start from scratch and make this with cheap IC components . The most expensive IC is 2$ , most are like 50 cents . You can sim most of it with a free sim , but theres always a limit to what you can sim with free unregistered .

Bringing a PC into this is not necessary .

And everybody experimenting on a tpu needs an oldschool analog scope , any old piece of crap is way good enough , not computer scopes .

[Yucca]

The TPU is in the khz range .

You dont need high Mhz speeds .The 74HC series can go up to 20 mhz

It will be alot less complicated and expensive to start from scratch and make this with cheap IC components . The most expensive IC is 2$ , most are like 50 cents . You can sim most of it with a free sim , but theres always a limit to what you can sim with fre unregistered .

I'm not building the unit solely for TPU research. I want to experiment with RF intermodulation in small coils amongst other things. I understand this unit is not for everybody, many will see it as folly and so be it, I will still build it.

I have a full complement of 74 chips, hundreds of them, probably enough to build a small CPU. I know 74HC are fast logic, but you would only have square output. Also to get rock solid freq you would need to drive 74 from xtal clock and it would only be easy to divide that clock by 2,3,4 etc using n bit ripple counter. To design an accurate square sig source even to 500kHz capable of 1Hz resolution would be near impossible with 74.

I can guarantee to design a clean sin generator with 1Hz resolution even up to 1MHz would cost alot of time and expense and to get all harmonic content 40dB under fundamental as these DDS are capable of would not be an easy task.

[dankie]

Well I happen to have much experience with this , There is much much effort to be done to make a good sine wave , the wave is a stepped wave and needs post filtering .

A 1 mhz pure  "sine wave" is not possible to do , you should forget that idea , a word of advice , stick with those square waves ,

A very fast clock is indeed necessary but that clock must be "variable"  haha ! , good luck with finding a 100 mhz simple square wave oscillator thats not too exotic , something simple ...

[Yucca]

Well I happen to have much experience with this , There is much much effort to be done to make a good sine wave , the wave is a stepped wave and needs post filtering .

A 1 mhz pure  "sine wave" is not possible to do , you should forget that idea , a word of advice , stick with those square waves ,

A very fast clock is indeed necessary but that clock must be "variable"  haha ! , good luck with finding a 100 mhz simple square wave oscillator thats not too exotic , something simple ...

Using DDS a "pure" sin wave is quite possible. DDS is a digital synthesizer producing a 10 bit linear approximation of a sin. So yes it contains 2^10 = 1024 discreet steps vertically but these are so fine that the output elliptical filter can pretty much take them all out. As I say all harmonic content guaranteed to be under 40dB relative to fundamental.

To test the stability of an oscillator it is best to listen to a heterodyne signal against another accurate oscillator. I have a DDS based small Yaesu VR-500 receiver 200kHz...1GHz all mode all band, and when I tune it on sideband to say 6MHz and then make an oscillator with an LC and listen to it, it is all over the place, very unstable and muddy and wide, I wave my hand near it and it zips away to another freq, I change the load impedance and it again loses freq. DDS would produce a rock solid tone with very narrowband regardless of conditions.

Don't get me wrong I'm not saying this is the perfect TPU tool. It's what it is, a 3 channel sin source with spectral purity as good as can be achieved and 3D scanning capability, I plan to keep it for life and use it in many experiments to gain a better understanding of frequency mixing in various systems.

You may well be right, square waves may be the way to go with TPU research, who's to say?