I'm trying to persuade the attached circuit to oscillate (ultimately at 100MHz but any frequency will do for now).
https://drive.google.com/file/d/0B1ZlKa1Knz2cQTA0TUlxMEpKQWc/view?usp=sharing
The MRF136 is supposed to be operating as a Colpitts oscillator but everything is just DC - I am applying a drain voltage of 20 and a bias of 5-7 volts to the gate and I'm measuring 150mA going through the device but however I tune the two capacitors, no oscillation.

Advice please!

 

What does your circuit layout look like? At these frequencies it can make a _lot_ of difference.

 

Here's the layout, partially built but the matching components not yet in. The 20V comes in on the red wire and case is ground.
The flux is messy but I'm scared to try to remove it because of the trimmers.


https://drive.google.com/file/d/0B1ZlKa1Knz2caVJXSnBCcmlWbVk/view?usp=sharing

https://drive.google.com/file/d/0B1ZlKa1Knz2cNUk0Y3YweVI0NTQ/view?usp=sharing

 

That is a beautiful layout and printed circuit board.

It appears as though you are hoping an electrolytic capacitor will suffice as a bypass capacitor. Not a good idea because the equivalent series inductance in the hundreds of MHz makes electrolytics "invisible". I don't know whether that is what is keeping your oscillator from oscillating, but putting a good quality ceramic capacitor or two across the electrolytic will assure that you really do have a decoupling capacitor.

 

Hello,

As Dick said, the electrolitic capacitor will not help in the place you put it.
There is a long line between C7 and C9, wich will act as an inductor.
Also you did not have the intended inductor placed at C9, assuming TP6 is the power point.
A couple of ceramic capacitors near C7 will help to decouple the circuit.

Have a look at the following PDF for RF prototyping:
rf_proto-1.pdf

Bertus

 

That is a beautiful layout and printed circuit board.

It appears as though you are hoping an electrolytic capacitor will suffice as a bypass capacitor. Not a good idea because the equivalent series inductance in the hundreds of MHz makes electrolytics "invisible". I don't know whether that is what is keeping your oscillator from oscillating, but putting a good quality ceramic capacitor or two across the electrolytic will assure that you really do have a decoupling capacitor.

Thanks for the kind words and very sensible advice which I have immediately implemented - please see revised arrangement with "your" capacitor inserted at ">":

https://drive.google.com/file/d/0B1ZlKa1Knz2cdWtJUFFGRlFVQ0k/view?usp=sharing

I also moved the two trimmers to the reverse of the board and cleaned away as much flux as I could.

Unfortunately, there is still no oscillation ...

Hello,

As Dick said, the electrolitic capacitor will not help in the place you put it.
There is a long line between C7 and C9, wich will act as an inductor.
Also you did not have the intended inductor placed at C9, assuming TP6 is the power point.
A couple of ceramic capacitors near C7 will help to decouple the circuit.

Have a look at the following PDF for RF prototyping:
rf_proto-1.pdf

Bertus

Thanks for responding Bertus - I made the changes suggested by Dick and also cleaned up the board arrangement somewhat. I shall leave no stone un-turned and your advice is appreciated - I will go and get the choke you pointed out as missing and try again. Thanks also for the documentation - that is very specific and to-the-point for me and for this project - I am reading it.

My ideal frequency here is 100MHz but right now I'll take anything except DC!

 

Inductor now in place but still nothing. Very puzzling!

 

Hello,

I am getting confused with the partnumbering in the schematic and PCB.
The partnumbering does NOT match.

Also the soldering on the transistor tabs is very little on your PCB picture.
Have a look at this PCB how the RF transistor is soldered:



Bertus

 

Sorry, Bertus, for boring reasons I didn't match the two - here is the schematic with the proper part-numbers:

https://drive.google.com/file/d/0B1ZlKa1Knz2cLUF3Yjg0akhHV1k/view?usp=sharing

Together with its layout:

https://drive.google.com/file/d/0B1ZlKa1Knz2caVJXSnBCcmlWbVk/view?usp=sharing

I appreciate that you took the time.

 

Hello,

I am getting confused with the partnumbering in the schematic and PCB.
The partnumbering does NOT match.

Also the soldering on the transistor tabs is very little on your PCB picture.
Have a look at this PCB how the RF transistor is soldered:

View attachment 95968

Bertus

Sure, I used only a little solder because that is an expensive transistor and I might have to junk the circuit! If the circuit works I shall not spare the solder, just like that other fella.

 

I discovered two things. (1) my inductor is around 6nH which, with a maximum capacitance of perhaps 10pF will give a frequency that is around 650MHz (way higher than I want) - I miscalculated the inductance and have now replaced this with a 330nH inductance. However, still no oscillation.

And (2) I took a look at the transistor.

While there is almost a linear variation of Isd with Vgs (Vds = 20V) with a cutoff at Vgs=4V and a maximum of 205mA for Vgs=8V - there is virtually no change of current with Vds. In other words, if I set Vgs=8V, Isd remains almost constant at 200mA for all Vds >10V, regardless of the voltage applied to the rail.

I have treated this expensive transistor with kid gloves but, unless I am doing something really dumb (which is all too possible) I believe I have a bad transistor here.

I'd be grateful for any confirmation - or otherwise.

 

The MRF136 has a huge amount of capacitance. It may be fine for a power amp where it is driven by a low impedance source but you will have a hard time making a stable oscillator in the 100 MHz range. The transistor's capacitance will overwhelm the capacitance in the tuned circuit.

 

RichardO said: ↑
The MRF136 has a huge amount of capacitance. It may be fine for a power amp where it is driven by a low impedance source but you will have a hard time making a stable oscillator in the 100 MHz range. The transistor's capacitance will overwhelm the capacitance in the tuned circuit.

Richard - thanks for responding. Your explanation fits what I was seeing and I have accordingly dumped that idea. Since the demise of the 2N3375 I haven't been able to find an equivalent and so lashed out to try a power FET.