Friday, June 29, 2018

2N7002 2 Stage AF amp

This was so simple even I could do it. No math just pick some values to limit the current and set the bias. I'm looking at about 1 ma current draw.

 I mv in 500 mv out. 1 mv will rattle my earbud 1 volt is loud enough with my ceramic head phone.


The frequency response is well beyond my hearing range. I put it on a bread board and put a diode, coil and antenna on the input. It plays my local station.
I made this circuit using 2SK669's  before and it was a good build also.

Thursday, June 28, 2018

lacing a parallel cable run

Usually when we lace a bundle we just wrap the string around the wires but what if we want the wires to lay flat?
I bent a hook in the end of a piece of wire to make it easier to feed the ribbon through
Lay the ribbon on the cables with a long enough tag to tie it off when finished.
Take the ribbon under and over to the first gap between wires.
Wrap under the main tag and back up then down through the second gap.
Wrap the tag and back up.
Keep spiraling down - over - up and down the next gap until you reach the far side.
Now bring the first tag under and up the right side. Tie the ends off and trim the ribbon.
Repeat at the other points needing tying.
If you use dental floss you can get a needle with a big eye and sew the wires in place. Using a needle will aid in separating the wire too. If they get a twist you can put the needle between them and iron the twist out by running it down the wire run.

Sunday, June 24, 2018

A MW RF amp using the data from the previous post

500KHz to 1500KHz
Expanded range to show more roll-off
The yellow is the half power points. The red is the MW band.
 Increased range to show roll-off.
Different transistor. You can try others to see which are useful in the desired application.

Saturday, June 23, 2018

amp input characteristic vs bandwidth


Setting the parameters for a stage with a 600 Ohm generator, 10ufd coupling cap, 2.4k Ohm input Z, and 1 nfd input capacitance.  10 Hz to 300 KHz bandpass. Now change some values and see what happens.
Reducing the input capacitance to 10p raises the high end to 30 MHz. What determines this value? The transistor inter-electrode capacitance, Beta, and strays. How would we reduce it?
Raising the input Z made the response worse.
Reducing the coupling cap value and the input cap gives a better high end response.

So the question becomes how can the designer control the circuit reactance? Maybe the question is how to work with what he has to get max benefit?

So build the circuit in SPICE and try different values. Look at the transistor datasheet and see what your real world values will be. Above all else have fun.



high-low corner and frequency response

The half power point is .707 signal out or when Xc = Rl in a RC coupled circuit. I will show some RC circuits and some transistor amps to examine.
50 Ohm response rolls off below 90MHz
Looking at the low corner.
Here I have 10 mv in and 5 mv out

My 2N3904 roll off is less than 50 MHz

Changing the caps shifted it slightly above 50 Mhz
usiong 50 Ohm I/O

600 in 50 out

50 in 600 out

600 in 50 out and smaller cap

50 in 50 out and small cap. observe the scaling.


You can see the RC circuit response is predictable but adding a transistor will cause the curve to shift. The transistor parameters come into play and must be accounted for. Some things are beyond the designers control. So we need to use the manufactures data sheet to see if out transistor will serve our purpose.

It's a simple circuit you can type into SPICE is a minute or two but playing with it can reveal a lot about how the coupling system responds.

Friday, June 15, 2018

Lorne MacDonald books on circuit analysis



I would recommend the MacDonald series. These two are what I will be posting screen shots from. If you want paper books you can purchase them at Amazon. A search may find a PDF copy. These are used in tech school with lab exercises to demonstrate the circuits. First the theory then the lab to demonstrate. They cover the theory so you know what to expect when you do the test. If you do not have the test equipment you can run the sim in LT spice with good results.
The current project is a 118.5MHz amp.  Better to build a low frequency amp and they increase the frequency. Here are a couple of screen shots to show what will be found in the books and to answer a couple of questions asked recently.


Interesting he points out the probe capacitance. The act of running the test invalidates the test results if you don't allow for the interaction. (Dr. Heisingberg's principle at work). Note the 10:1 probe is 7-10p.

How to determine input Z.
Miller effect.

Sooner or later you need to add a second stage. He cover cascade and cascode amps. After laying the foundation he uses it to explain RF circuit limitations. Using the datasheet to determine the component limits could save some grief in failed construction. I generally play around with Short Wave which is 30 MHz or less. I have a bin labeled RF amps. My go to transistor are rated 50MHz so for this project I must use a different transistor.
Well time for the adventure to begin.

Tuesday, June 12, 2018

The Miller effect and more

The scans are not the greatest but maybe you can read them. The inter-electrode capacitance and stray capacitance are the enemy. And then comes the Miller Effect.
I will post a link to a good book with this information in it if I can find one.






 The half power point is where Xco = Rl. Increases C causes Xc to decrease. Adding parallel capacitance causes C to increase. Stray capacitance and inter element capacitance are at play.

Monday, June 11, 2018

Large signal or small signal detection

If you amplify an AM signal with a non-linear amp it will demodulate the signal. Some AM detectors operate in the non linear region and are square law or small signal detectors.
If you rectify an AM signal it will be detected. A detector operating in this region is a large signal or linear detector.
A simple crystal detector is a small signal detector and operates square law. This is not an advantage it is just the way it is. As the signal increases the output increases at a faster rate because of the non linear detector but it is still weak.
A receiver with an RF amplifier will boost the signal to the point it operates large scale or linear. It will have a higher output because it was amplified not because of the detector operation.
This is a simple explanation of how a detector works. Nothing wrong here unless we assume it is a crystal set operating square law. The signal is rectified. In order to be rectified it must be strong enough to over come the knee of the diode. What is the knee? It is sometimes called the bend. What it is in a solid state diode is the voltage required to overcome the built-in potential. The semiconductor doping causes barriers in the structure which must be overcome. The easiest way to define the knee or bend is that point where the slope of the curve equals 45 degrees. Bend detection was used in tubes too.

 Square law, weak signal plate detector.



Linear,strong signal plate detector.

If you build an amp attach an antenna to the input and hear audio on the output you have a non linear amp or you or over driving it.
non linear = square law detector
over driving = linear detector.
If you connect the antenna and hear AC it is just line noise. When adding a diode across the input produces audio it is high gain enough to trace signals in a radio and linear enough to give valid test.

If your amp passes these test it is ready to serve as a signal tracer.

signal tracing as a troubleshooting technique

Looking through the Health manual will tell more than I can in just a few words. The simple AF amp with a probe can be a good addition to a test bench. Especially for the bottom feeder without deep pockets.

signal tracer






 This looks like the one I had as a kid. Notice it had a RF and AF input. The amp was strong enough I could clip an antenna to the RF input and receive our local station.

So you can build a high gain AF amp and a RF probe and be in business. The probe would be like this one.


This one is designed to be used with a voltmeter to read RF signals. A search will find more this is the idea. C1 would be small enough to block AF and large enough to pass RF.

Now how would you use it?

If you made this radio and it failed to produce output you could check the speaker with the AF probe. If you hear signal the speaker is bad otherwise check the input to the power amp. If you hear signal the power amp is bad otherwise check the input to the AF amp. If you hear signal the AF amp is bad otherwise attach the RF probe and check the detector input. If you have signal the detector is bad. Etc...

If you started at the first RF amp input and had signal and moved to its output you would expect to hear a louder signal. As you progress through you would be turning the volume down to a good listening level and expecting and increase. When you step over a stage and the signal drops you have found the problem area. Then you troubleshoot that stage......

Here they show the signals you would find in a superhet receiver. You just need to know when to look for AF or RF and use the correct probe.

So build an amp and probe and you are set to troubleshoot radios.

What difference can the transistor make?

I built the circuit in the simulator and the changed the transistor.




The J112 is a good low frequency transistor but not so good at 100 MHz. It is a 5 cent part so a second stage could be used to pick up the gain.



The 2N3819 is another good low frequency amp. All but a total failure here though.




The good news and the bad news (for me). The best performer is a surface mount. For the difference in performance it is worth a little mounting issue. The good news is Arrow sells a 3K reel for $15 with free shipping.

This was an exercise in making the cascode amp. I would use a BF998 or BF2040w when available. Sadly they are both reaching end of life. Last time I checked Arrow still had some but when they are gone they are gone.

The simple approach used in the above circuit was to set the bias on J2 at 1 volt and the bias on J1 at 4 volts. I wanted about 1 ma current so I adjusted the other resistors to suit.