el cheapo oscillator

[dannyf]

Anything over 1Ghz is impressive. Nice job.

[Isad]

Well i tryed the 2 transistor one
and it never worked with a short coil
and it would go up in mhz when i would higher the
ground resistor used a 10k 10 turn pot.

[Zero999]

From left to right, my earliest to my latest attempt.

Here's a video of the latest one running over 13GHz after trimming. 

[url]https://www.youtube.com/watch?v=8tbRkoq3HBI]https://www.youtube.com/watch?v=8tbRkoq3HBI] [url]https://www.youtube.com/watch?v=8tbRkoq3HBI[/url]

Which circuit did you use and what was the transistor?

I suppose the physical size of the components is the limiting factor 13GHz is a wavelength of just 23mm.

Have you tried using a piece of transmission line as a resonator?

Unfortunately I don't have any transistors which can go that fast or a good enough frequency counter.

[paulie]

Well this might not be the fastest but maybe smallest and el-cheapo-est. So current  record holder according to thread title. PF949 9ghz transistor at only 2 cents ea and 1mm size would be hard to beat. It could be made half again smaller with 0402 instead of 0805 and coiling the inductor.

Only 2 resistors and no tuning caps. Squeeze the coil to adjust. The supply bypass cap seems to be pretty much optional but I keep it in there to hold it all together. Very stable compared to some of the circuits and more than just a science fair project because 2.414ghz happens to be exactly what's required for generic spy cam tx. Pumping signals into the base does show expected image on a tv.

[dannyf]

The boards are made from FR-4 ...

Nice boards. You could start a business selling high frequency oscillators,

[G0HZU]

This subject popped up on another forum a few years ago titled the world's simplest VHF oscillator.

http://forums.qrz.com/showthread.php?335985-World-s-simplest-VHF-oscillator!/page2

I posted up a much simpler oscillator using just a JFET, some wire and a battery. It oscillates at 128MHz.

So no PCB or resistors or caps needed or special construction skills.

So current  record holder according to thread title.

Not any more

I could make something faster and smaller but that oscillator had to work in the VHF band to qualify as a contender for world's simplest VHF oscillator.

[Zero999]

Very stable compared to some of the circuits and more than just a science fair project because 24.14ghz happens to be exactly what's required for generic spy cam tx.

I think you've got the decimal place wrong there. The 2.4GHz region is normally used for spy cams.

[paulie]

LOL. Yes, 24ghz spy cam would never get out of the bedroom. At least the GOOIT display and photo label were correct. So maybe not fastest or simplest but so far definitely the cheapest.

Speaking of simplest, G0HZU, what part did you use? Also any chance for 5v instead of 9v? I tried with PN4117 but so far no hint of oscillation.

[paulie]

Not any more

I could make something faster and smaller but that oscillator had to work in the VHF band to qualify as a contender for world's simplest VHF oscillator.

I think yes. This is "el cheapo" thread so smaller and faster don't count.  Although I do think there is no smaller so far. Definitely cheapest because JFETs cost at least 10x more than PF949 and 0805s fraction of a cent. You might have record for simplest with proof it actually oscillates.

[G0HZU]

I think yes. This is "el cheapo" thread so smaller and faster don't count.  Although I do think there is no smaller so far. Definitely cheapest because JFETs cost at least 10x more than PF949 and 0805s fraction of a cent. You might have record for simplest with proof it actually oscillates.

It's effectively a Hartley oscillator biased at Vgs = 0V. The reason it oscillates is because it exploits the internal capacitance of the JFET.

I seriously doubt that anyone would deny that this type of circuit will oscillate. I designed it on a simulator and it is easy to demonstrate negative resistance occurs up at VHF. Anyone can do the same or they can just build it. You do have to get the lengths of each of the wires right though. I built this oscillator a long time ago and I don't think I need to build it again just to prove it will oscillate.

Speaking of simplest, G0HZU, what part did you use? Also any chance for 5v instead of 9v? I tried with PN4117 but so far no hint of oscillation.

I used a J310 JFET. To give you some incentive I'll offer you $500 via Paypal gift if you can prove my oscillator design below is fake and can't oscillate up at VHF

Just build one... 

[paulie]

I seriously doubt that anyone would deny that this type of circuit will oscillate.

IMO just the opposite, most would be very surprised such a simple circuit does anything at all.

Anyway no luck with 2n3819 ether. J310 too pricey to risk buying for something that may only work under very peculiar circumstances. I think in situations like this it's reasonable to expect some demonstration of functionality even if just current change touching coil. I could post a stock photo of a canadian goose saying it's oscillating at 300ghz and claim the speed trophy. But seriously I think it is possible this works and would love to get it working someday.

[paulie]

To give you some incentive I'll offer you $500 via Paypal gift if you can prove my oscillator design below is fake and can't oscillate up at VHF

Just build one... 

I've built several just in the last half hour. And proving it's fake is virtually impossible. It would be like trying to prove there is no god. Proving it WORKs would be more reasonable.

[G0HZU]

To give you some incentive I'll offer you $500 via Paypal gift if you can prove my oscillator design below is fake and can't oscillate up at VHF

Just build one... 

I've built several just in the last half hour. And proving it's fake is virtually impossible. It would be like trying to prove there is no god. Proving it WORKs would be more reasonable.

Jeez... just build it as per the picture with the correct part.

I dug out a J310 and did a youtube video just for you. The video is not fake. There is no conspiracy here. There are no aliens or holograms or illuminati involved in the making of this video. Just a J310 JFET and some wire and a 9V PSU

 http://youtu.be/WAKzprRu_T8

[dom0]

I'm smelling a reichsflugtransistor conspiracy here...

[paulie]

I dug out a J310 and did a youtube video just for you.

That appears to be one of those rare Adobe only videos which are unable to view on this PC but I'll assume you demonstrated functionality. Congratulations, forum record holder for simplest oscillator. I'll have to be satisfied with smallest and el-cheapoest ATM.

Interestingly no hint of that circuit in first couple hundred images with google "simplest oscillator" so looks like you are the world record holder too. Or maybe not because "reichsflugtransistor" came up with dozens of references to either nazi oppression or n-channel this morning, then an hour ago couple for mouser/digikey, and now... only this post. So something funny is afoot with my cookies or the search engine.

I'm impressed and almost tempted to order some J310 parts. But first further experiments with the PN4117. I'm wondering if it might have something to do with slight difference in Vgs(f). Mine draws almost nothing with gate grounded but J310 probably some microamps. Most other specs similar including picofarads. Not much info on top frequency for either.

Anyway thanks for the effort. You have definitely aroused my interest in this amazing circuit.

[G0HZU]

Here is the same video in MP4 format. Hopefully this will play OK.

http://youtu.be/PerBSUTi7Qk

I can do a version using a 2N3819 if that helps? Then you can copy it. It will run up at a much higher frequency than the J310. It should work down to 5V supply voltage possibly less although it might not be as reliable.

I can only assume that you have got the pinout wrong in your version so far?

[G0HZU]

Here's a quick and dirty SPICE simulation showing the same J310 JFET circuit oscillating in the FFT plot.

You can replicate this yourself? In the background is a basic linear simulation showing negative resistance in the same frequency band for the same circuit.

[dom0]

That appears to be one of those rare Adobe only videos which are unable to view on this PC but I'll assume you demonstrated functionality.

HTML5 player works fine here ; Firefox thirty-something.

[paulie]

Here is the same video in MP4 format. Hopefully this will play OK.

That second video played fine so no doubt now you deserve the title of king of minimal oscillators. This PC has FF 34.0.5 and your first definitely needs another codec. Fortunately only 1 out of a hundred or so are like that. A year ago it was 1:10 so things are getting better for HTML5.

Being obsessed with simplicity i found it the most interesting circuit to come along in months. I spent a lot more time on it. First I tried putting a cap on the loop and biasing the PN4117 so it's well into linear (few ma). No go. Then I pulled out an rf test jig that works with every type of transistor in the world: bipolar, mosfet, dual gate, etc.. Guess what... the PN4117 failed there too. So no doubt a problem with this particular part. They've been working great for super high impedance apps like ion chamber radiation detector and pico amp meter but apparently not for RF. Looking at specs I don't know why.

I was recently given dozens of these PN4117 but only had one 2N3819 and checking with a meter I see it is now blown. Probably during the biasing tests cranking up voltage with no current limit except that wire loop. I really should stock up on a few different JFETs because it looks like they are quite useful for other things like active scope probes too. Anyway I'm impressed with that circuit and may give another try some day.

[G0HZU]

I tried it with a 2N3819 and minimised the size and I could get it to run over 300MHz. If it was deliberately made to be microphonic then it could make a very simple wireless bug. Obviously it won't be very stable but you could make it very thin if you used a SOT23 package JFET and then hid it in a small wall or desk picture (where the floppy nature of the picture would assist the microphony) although you would still have to power it. The 2N3819 is a better choice because this 300MHz version runs at just 3mA.

The J310 has an Idss of 60mA or more and so is very inefficient. The one in the video got quite warm when running so the J310 isn't the ideal device to use.

[paulie]

OK, that did it. Inspired by your video and 2N3819 comments I've decided to order some. I resolved to stop ordering frivolous parts but this "worlds simplest oscillator" is so fascinating couldn't resist.

[paulie]

2n3819 finally arrived but still no luck. Built as in pic below no sign of oscillation. Also a buddy radio club member using 2n3819 from another source failed too. No ten thousand dollar spectrum analyzer and lab supply here but super sensitive Gooit counter detects nothing and no changes in supply current touching the coil. This same JFET runs fine up to hundreds of mhz in a simple Colpitts though. I will try longer leads and play with different inductor and 2n3819 or voltages. Hopefully I don't have to resort to ordering J310.

Strange there's no hint of anybody else ever doing this because it's an incredibly interesting circuit. It would be helpful to see a photo of another 2n3819 version. IDK.



ps. OK, problem solved. One kind soul sent me a photo of his 2n3819 build and it was immediately obvious to me the gate and source were switched. It was a mistake to wire it exactly like your j310 photo and video. I was aware of the pinout difference at one point and wired correctly but something else must have gotten screwed up and then I backtracked and forgot.

Note that the 2n4117 still does not work but that is to be expected because it didn't in the alternate Colpitts jig either. Not surprising based on comments from Bernice in the low noise measurement thread. Apparently something to do with the extremely small gate junction. Strange because I would think small area would allow faster operation.

In any case the one from my photo is now oscillating at 167mhz and passing the Gooit and touch test.. I'm absolutely thrilled and can tell this is going to take up many hours playing with variations and limits. Thanks a bunch (curse or credit? LOL).

[G0HZU]

Strange there's no hint of anybody else ever doing this because it's an incredibly interesting circuit.

I'm glad you finally got it going but I can't really understand why you find this crude circuit so fascinating? 

For example, the 2N3819 will self bias into the saturation region if you run at Vgs = 0V. So that's the DC operating point set at several mA if you short gate and source.

Then you configure it as a Hartley oscillator by arranging the wires a certain way. i.e. the circuit exploits the JFET's own internal capacitance (and the self capacitance of the wires) to achieve gain around the loop at a frequency where there is unity phase shift around the loop. So no need for external bias resistors or blocking/feedback caps anywhere.

If you look back at the QRZ.com link where I first suggested this circuit I didn't even bother to build it for the first post on that forum. To me, it's so obvious this circuit MUST work as an oscillator up at VHF that I just designed the concept in my head for the first post