I've built a 3 stage FM transmitter with NPN 2N3904 Transistors and I did the diagram in Eagle Cad and the board and then used the toner-transfer method for the traces. I have installed all components and the 3rd stage transistor(RF output) is getting really hot when you touch it. I scraped the enamel off the coil(5 turn/3mm) and thought that might have been the initial problem(didn't scrape enamel off initially cause it was the copper colored magnetic wire) but after re-installing the coil and putting a new transistor in thinking the other was damaged, it is still getting very hot, hot enough that after a few minutes hooked up to the 9V source, when I put my finger on it I have to remove it within seconds cause it's so hot. I am somewhat of a novice at creating my own circuits but i know what I'm doing as far as putting them together. Here is the link to the circuit I built. http://www.youtube.com/watch?feature=player_detailpage&v=ft7-bxCA0y4. Please help me figure out what's wrong or if it's normal. Do not have a frequency counter so can't test it that way as of yet but I am getting voltage across the coil and from B-C, B-E & C-E. I'm at a loss . I'll gladly do anything you ask to further help you help me with this regarding providing more information. Thank you in advance. JT

 

Hello all,

I've posted the schematic, the board layout both done in Eagle and the back of the circuit after soldering. To show the back, the front is simply flipped 180° left. Hopefully this will help someone figure it out or if it's normal for the rf out trans to get hot enough to where you burn your finger when you touch it for just a second at least tell me that, please. Thank you very much. Your replies are much appreciated. BTW, the transistors are 2N3904 NPN.

 

No, the output transistor should not get hot enough to burn your finger. If you are satisfied that the board layout follows the schematic, the problem may have to do with a lack of a ground plane under the components. In general, RF circuits are built with lots of copper left on the PCB. Such circuits are often built using the "Manhattan" construction technique. Google that for some representative photos.

ETA: I just started looking at the video. He is using the Manhattan style of construction, which is much better for RF.

 

Hello,

The long lines on the PCB might give you troubles.
As said, the manhattan style is a better way to build RF circuits.

Have a look at this post for more info:
http://forum.allaboutcircuits.com/showpost.php?p=152539&postcount=30

Bertus

 

Thank you guys for the input. I thought of doing it that way but couldn't find any 1 sided copper board. I guess I could cover the copper on the other side with an insulating material before gluing down but I'll probably just order some single sided copper clad boards. The reason I decided to use the pcb technique I used is because there's another video where he builds almost the same circuit and that is the kit I mentioned with a professionally made board with vias and green coating and all the fixins. I'll research the Manhattan style and do it that way. If I wanted to try a toner transfer would you suggest a ground plane and larger mil traces with larger vias and shorter routes to the important parts?

Thank you again so much for your input. I appreciate greatly.

JT

 

You can use double sided PCB material. Just be sure that there is no electrical connection between the top copper and the bottom copper on the small pieces.

As to your second question, layout of RF circuits is a very specialized field. I don't know the rules.

There is a FM transmitter that was designed and built by a former member here. He built his on a stripboard, and published his design. You might want to try it. Google audioguru fm transmitter.

 

Thank you tracecom and bertus for your valued information and suggestions. I am very appreciative and grateful. Hopefully I can be of assistance on anything mechanical involving circuits obviously; this is an "All About Circuits" forum after all.

 

I think you guys are missing the fact that, if the beta of T3 is the specified maximum (300), then the base current will be about 90mA. That translates to about 800mW dissipated by T3, enough to burn your finger if you touch it, and possibly destroy the transistor.
If the beta is "only" 200 the dissipation will still be over 500mW.
The beta of BJTs increases with temperature, resulting in possible thermal runaway if operated with no resistance in the collector or emitter.
I would add a 100Ω resistor between the battery and the 1uH inductor, with 10nF to ground between them, and the base bias resistor connected to the same node. The resistor needs to be rated for at least 1/4 Watt. See attachment.

EDIT: I believe the circuit needs preemphasis in front of the oscillator, because FM receivers have deemphasis built into them.

 

Have you considered the possibility that if your antenna is a poor match a great deal of the forward power produced by the output stage will be reflected back causing excessive dissipation and potential destruction?

Get an antenna analyzer and construct an antenna that is a match to your output stage.

 

No, the output transistor should not get hot enough to burn your finger. If you are satisfied that the board layout follows the schematic, the problem may have to do with a lack of a ground plane under the components. In general, RF circuits are built with lots of copper left on the PCB. Such circuits are often built using the "Manhattan" construction technique. Google that for some representative photos.

ETA: I just started looking at the video. He is using the Manhattan style of construction, which is much better for RF.

Never heard of the Manhattan style.. but found this nice pictorial and description on the google. Thanks!

 

Never heard of the Manhattan style.. but found this nice pictorial and description on the google. Thanks!

You are welcome and thanks for the pdf, which has lots of info I didn't know.

Here is part 1 of the article.

 

Have you considered the possibility that if your antenna is a poor match a great deal of the forward power produced by the output stage will be reflected back causing excessive dissipation and potential destruction?

Get an antenna analyzer and construct an antenna that is a match to your output stage.

That could be, but, as I said in post #8, the transistor doesn't need any help from the antenna. It can overheat with the antenna disconnected.

 

That could be, but, as I said in post #8, the transistor doesn't need any help from the antenna. It can overheat with the antenna disconnected.

No antenna is one case where no power is radiated and most of it reflected.

 

No antenna is one case where no power is radiated and most of it reflected.

I am certainly no RF expert, but I believe that reflections occur when the load (antenna, in this case) does not match the transmission line impedance.
No transmission line, no reflections.

 

I can tell you for certain sure that an RF output stage is not very happy driving an open load (infinite impedance). In AC terms it is roughly the same as driving a short. True there is no transmission line, but there is also no place for the energy to go.

 

I can tell you for certain sure that an RF output stage is not very happy driving an open load (infinite impedance). In AC terms it is roughly the same as driving a short. True there is no transmission line, but there is also no place for the energy to go.

I ran a bunch of sims on the output stage, with varying drive levels and varying (cap-coupled) load resistances, from 50Ω to ∞. The transistor dissipation varied between ≈ 700mW and 800mW, which is too much for a 2N3904. I recommended a 100Ω resistor in the collector circuit, but after running some sims, I think it would be better to put it in the emitter circuit, with a 10nF - 100nF bypass cap.

Actually, I think this is basically a crappy circuit, and there are much better ones out there. You probably could do much better with one of Audioguru's designs.