GHz oscillator

Not a useful answer for you, but back in the 60s I was doing research on tunnel diodes. I was impressed on how, if we were not careful, one plus one two passive components would oscillate at a high (for us, at the time!) frequency. Remarkable for a two terminal device, because of its negative impedance.

I loved TDs when I was a kid. My college senior paper was "The Tunnel Diode Slideback Sampling Oscilloscope."

A few people still make germanium back diodes, but I think TDs are gone. The fab process was insane. TD's were hard to curve trace because they really wanted to oscillate in the negative resistance region.

Here's an oscillator with a phemt. I don't really like it because it seems to be tricky in simulation. It can do weird things like go chaotic with some part values.

Version 4 SHEET 1 880 680 WIRE 176 -256 0 -256 WIRE 0 -192 0 -256 WIRE 176 -192 176 -256 WIRE -240 -48 -336 -48 WIRE -48 -48 -144 -48 WIRE 0 -48 0 -112 WIRE 0 -48 -48 -48 WIRE 176 -48 176 -112 WIRE -240 -16 -272 -16 WIRE -96 -16 -144 -16 WIRE -272 32 -272 -16 WIRE -96 32 -96 -16 WIRE 0 32 0 -48 WIRE -336 112 -336 -48 WIRE -96 112 -336 112 WIRE -48 112 -96 112 WIRE 0 176 0 128 FLAG 0 176 0 FLAG 176 -48 0 FLAG -272 32 0 FLAG -96 32 0 FLAG -48 -48 D FLAG -96 112 G SYMBOL mesfet -48 32 R0 WINDOW 0 86 29 Left 2 WINDOW 3 61 56 Left 2 SYMATTR InstName Z1 SYMATTR Value SAV551 SYMBOL res -16 -208 R0 WINDOW 0 54 44 Left 2 WINDOW 3 53 73 Left 2 SYMATTR InstName R1 SYMATTR Value 50 SYMBOL voltage 176 -208 R0 WINDOW 0 88 48 Left 2 WINDOW 3 59 93 Left 2 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V1 SYMATTR Value PULSE(0 3.3 1n 1n 0 1) SYMBOL tline -192 -32 R0 WINDOW 0 4 -65 Bottom 2 WINDOW 3 2 -55 Top 2 SYMATTR InstName T1 SYMATTR Value Td=100p Z0=70 TEXT 144 88 Left 2 !.MODEL SAV551 NMF( vto=0.08, Beta=0.3,\n+ Lambda=0.07, Alpha=4 B=0.8, Pb=0.7,\n+ Cgs=0.997E-12, Cgd=0.176E-12, Rd=0.084,\n+ Rs=0.054, Kf=5e-11, Af=2) TEXT 262 34 Left 2 !.tran 0 20n 0 1p TEXT -328 -240 Left 2 ;3 GHz Phemt Oscillator TEXT -280 -200 Left 2 ;JL Apr 27 2021

I might dremel breadboard it, and a MMIC version.

tirsdag den 27. april 2021 kl. 21.35.27 UTC+2 skrev John Larkin:

$330 development board for a $7 part, it must be plated with platinum

$7 is cheap for anything Hittite!

MAX2750 is a 2.5 GHz VCO, but it's Maxim.

What ever happened to the Analog Devices acquisition of Maxim?

Being Maxim, of course most of the versions are obsolete.

onsdag den 28. april 2021 kl. 01.52.32 UTC+2 skrev John Larkin:

sure, but the $330 board is basically just the part and two sma connectors

afaiu they are just waiting for EU antitrust approval

Any chip or module with wifi or bluetooth. ESP32 or ESP8266 with wifi ought to do, or one of the Nordic ones with bluetooth. I think you need

As to whether it is a sine wave, it must be close, to get FCC approval without fancy filters before the antenna.

This is neat. I wonder if it would actually work.

Version 4 SHEET 1 880 680 WIRE 176 -256 0 -256 WIRE 0 -192 0 -256 WIRE 176 -192 176 -256 WIRE -224 -48 -272 -48 WIRE -112 -48 -144 -48 WIRE -48 -48 -112 -48 WIRE 0 -48 0 -112 WIRE 0 -48 -48 -48 WIRE 176 -48 176 -112 WIRE -112 -16 -112 -48 WIRE 0 32 0 -48 WIRE -112 64 -112 48 WIRE -272 112 -272 -48 WIRE -64 112 -272 112 WIRE -48 112 -64 112 WIRE 0 176 0 128 FLAG 0 176 0 FLAG 176 -48 0 FLAG -48 -48 D FLAG -64 112 G FLAG -112 64 0 SYMBOL mesfet -48 32 R0 WINDOW 0 -40 105 Left 2 WINDOW 3 -70 134 Left 2 SYMATTR InstName Z1 SYMATTR Value SAV551 SYMBOL res -16 -208 R0 WINDOW 0 54 44 Left 2 WINDOW 3 55 75 Left 2 SYMATTR InstName R1 SYMATTR Value 50 SYMBOL voltage 176 -208 R0 WINDOW 0 56 36 Left 2 WINDOW 3 47 82 Left 2 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V1 SYMATTR Value PULSE(0 3.3 1n 1n 0 1) SYMBOL ind -240 -32 R270 WINDOW 0 73 57 VTop 2 WINDOW 3 69 60 VBottom 2 SYMATTR InstName L1 SYMATTR Value 3.9n SYMBOL cap -128 -16 R0 WINDOW 0 -60 23 Left 2 WINDOW 3 -67 52 Left 2 SYMATTR InstName C1 SYMATTR Value 1.5p TEXT 40 72 Left 2 !.MODEL SAV551 NMF(vto=0.08, Beta=0.3,\n+ Lambda=0.07, Alpha=4 B=0.8, Pb=0.7,\n+ Cgs=0.997E-12, Cgd=0.176E-12, Rd=0.084,\n+ Rs=0.054, Kf=5e-11, Af=2) TEXT 248 -8 Left 2 !.tran 0 20n 0 100f TEXT -320 -224 Left 2 ;3 GHz Phemt Oscillator #3 TEXT -272 -176 Left 2 ;JL Apr 27 2021

Accidentally?

:)

I just posted an accidental 3 GHz phemt oscillator.

Purely fiddled in Spice in a few minutes; all instinct, no theory at all. Reasonably good sine wave.

Amplifiers will oscillate. Oscillators won't.

Pricing of evaluation boards, and even more of "educational" boards, is sometimes very interesting. They can be very expensive or ridiculously cheap.

Check the ADALM-PLUTO... it can probably do this function and some of the functions required in addition to that oscillator as well. But it costs much less than that, even though it has many more components.

Why not a synthetizer/VCO chip, like the ADF4351 and a small MCU to initialize it ? If you prefer a ready-made version, you'll find boards with the synth chip and a nice TFT for 40$ or so on ebay... Output power is a couple of mW.

Robert Lacoste ALCIOM - The mixed signal experts -

I don't have a specific application, but I was hoping for something simple. It's just a shower problem at this point.

The general problem is to push a DC-coupled baseband signal of bandwidth B, over a typical comm channel, like a twisted pair with magnetics on both ends, or a fiber with cheap tosa/rosa things on the ends, ac coupled with bandwidth 2 or 3 times B and basically a comparator as the receiver, maybe a poorly defined AGC.

One idea is to start with a 2 or 3 GHz oscillator and PWM it. That should be easy to generate and easy to detect. FM is more work. PCM much more.

Just thinking.

A little phemt or MMIC oscillator won't cost much more than a dollar.

Ungodly amount of gate current, though--abs max is 2 mA.

Cheers

Phil Hobbs

Good point; that's fixable.

I'd prefer to use a MMIC, but there seem to be no spice models for mmics. Mini-Circuits was adamant that they would never have spice models for theirs.

I guess I've complained about that enough times, so I should make some. The model will be easy but validating will be a pain.

I'm thinking of a BFT25 darlington, then tweaked to match reality. ERA-1 to start maybe.

They specify the S-parameters so you can figure out how much more phase shift is needed on the desired frequency to satisfy the Barkhausen criterion. Of course you have to check that the forward gain is larger than feedback network losses to have a sustained oscillation at desired frequency.

MMICs are basically Darlingtons. I suppose a Colpitts arrangement should work.

Jeroen Belleman

It's on a 52-week lead time, but ... wait, where are you going? We have this great new company coming out that you can buy...

-- john, KE5FX