[Solved/Measured] 74LVC1G04 highest frequency?

[szoftveres]

I'm planning to build a synthesizer using 74LVC1G04 inverters to triple the frequency at each stage. The datasheet says "for applications up to 100MHz" however I'm wondering if these devices will run up to 300MHz and will be able to produce harmonics up to the 900MHz range. (In the past I've clocked the 74LVC1G74 at 200MHz in a frequency divider (f/2) mode)
Any hands-on experience with these devices?

[tggzzz]

W.r.t. harmonics, see https://www.eevblog.com/forum/testgear/show-us-your-square-wave/msg1902941/#msg1902941

[szoftveres]

Digikey package arrived, measured with a VNA. There's a prominent roll-off which seems to be the function of Vdd, (i.e. not because of wrong input matching), starting at around 325MHz at Vdd=5V, all the way down to 150MHz at Vdd=1.6V.
I think I'll use something else.

[tggzzz]

The transition times are a function of the supply voltage, and the frequency domain mirrors that.

With a little effort, you can get harmonics to 1GHz (and beyond). See the post I referenced earlier, and the FFT below.

Note that the breakpoint between the 20dB/decade and 40dB/decade rolloff occurs at about 1GHz, which does correspond to Howard Johnson's rule of thumb noted in https://entertaininghacks.wordpress.com/2015/08/11/measuring-digital-signal-edge-rates-without-an-oscilloscope/#more-652

[szoftveres]

Yep, I saw your previous link, it obviously points out that the spectral content is there and one can easily make 900MHz by e.g. picking out the 9th harmonic and amplifying it after a 100MHz gate, and probably that's what I'm going to do. I was just hoping to be able to take advantage of the relaxed filter Q requirements and inherent high signal level at 900MHz when the gate is clocked at 300MHz. Thanks for the great info anyways!

[Gerhard_dk4xp]

This is the spectrum of a 100 MHz LVCMOS wave passed through 2 pcs.
1G lvc125, each with 82 Ohms and connected to get somewhere near 50 Ohms.
1G lvc04 is probably the same.
The B3728 are Siemens/EPCOS/TDK/... whoever owns them now, Digikey @ $1.50
The output is pretty clean. I use these filters also for my 1st IF, 900-928 MHz.
The 900 MHz is just on the lower pass band edge, but still OK.
It should fit to a LMX2594/95 clock input, I have tested only 100 & 300 MHz.
The LC filter@300 is not good enough for clean 300 MHz.
I did not find any 300 MHz SAW filters that are still available.

In a neighbor thread "why is the xyzzy synthesizer so good..." is a photo
of the multiplier.
And  remember, if you drown a harmonic in thermal noise, you will never
be able to make up for it again. 30 dB loss is a lot. There goes the phase noise.

cheers, Gerhard  DK4XP

[szoftveres]

Thanks for the spectrogram and schematics!
Just out of curiousity, what was the thought behind deciding to use two saw filters back-to-back? Did the strong fundamental or some lower harmoics leak through?

[Gerhard_dk4xp]

The idea was to get more far-off attenuation. The data sheet
promises only 40  dB and only if the ground connections are
very short, but 50 Ohm I/O and no-tune.  At 93 Euro-Cent per
filter @ 10 there's no need to skimp on them.

<   https://www.digikey.de/de/products/detail/qualcomm-rf-front-end-rffe-filters/B39921B3728U410/3492653?s=N4IgTCBcDaIEIGYDsYAcIC6BfIA      >