Tuesday, July 31, 2018

K2539 Amp rebuild

This is my new go to amp. ( I may remove the first stage.) It is almost to high gain. The one I made last night was great BUT I added a switch and rearranged the circuit in the build I did tonight.
 The lower one was a 1-1/2 volt Germanium transistor version of the hearing aid amp. The smaller board on top is the 9 volt JFET version. It worked so well I decided to do another and add a switch and battery tie points. To test it I connected 2 DVM's set on the 200 mv range and a signal generator. I have the ear buds plugged in so the reading are loaded values.
 Let's see .03mv = 30 uv input and 11 mv out, very good.

 The amp was to sensitive. It would hum when i put my hand close to the input. I placed a 1nfd cap and 510 Ohm resistor across the input. This helped bypass noise to ground. The readings were all taken with the input bypassed.

The next step will be to place an attenuator on the input and a demodulator to feed it. I thinking Gilbert Cell.
I'm still considering removing the first stage. I clipped a diode with a antenna across the second stage input and hear my local station. It may need the first stage for weak signal reception????? Time will tell.

Sunday, July 29, 2018

Poor man's double sided circuit build of a 3 JFET amp.

The ground plane is very important and a good practice is to use a double sided board. Making one foil +Vcc and the other -Vcc. So what do you do if you have no double sided circuit board? Order some from a supply house or maybe just what you have available and make your own.
This old rough cut 2 X 4 is what I was looking for. It was used to chock some material shipped to a construction site. It would be land fill had it not found its way into my shed.
I chopped a short piece from the end of one to make it easy to handle. The sliced some thin pieces from it. I did not measure just something around 1/4".
I laid it on a magazine so I don't make a mess on my desk top.
I applied a coat of contact cement and get some aluminum foil for the laminate. What I'm using is not as thick as you might like but it was in the shop.
After the glue set for a couple of minutes I applied the foil.
I turned the page in the magazine to get a clean surface and laid the blocks facing up. Apply a coat of cement.
After a couple of minutes fold the foil over on the blocks and let it set a while so the glue can cure.
A sharp knife or razor blade can be used to cut the laminated blocks out.
 And we now have three double laminated bread boards. You could use flashing or a can just whatever is available. I like the coffee can bottoms. They are large enough and flat. Hair spray cans are good too.
As you can see it is a little ragged. You can sand or file the edge smooth but as long as it doesn't short out a little fuzz will not hurt.

This one is a little crinkled. You can use a putty knife or old playing card to work out the wrinkles I just was sloppy I guess? It will work wrinkles and all just not as pretty as it could be.
The large breadboard is poster board. It is the amp I posted 2 or 3 post ago. The little one is the double sided board version. I left the leads full length on the poster board version so i can reuse the components.

I used the copper foil glass workers use to solder stained glass with to make my solder points. I mounted all the transistors on little insulated islands and then added the grounded components. The +Vcc being on the bottom foil I just drilled holes sized to the wire and the used a 1/8" drill to drill through the foil on top. this allows the leads to pass through without shorting. The little amp is a good one. I would think you could use any JFET available to you. The biasing may vary a little but the worst case would call for adjusting the source resistor.

I used pasta maker manhatten pads to mount the transistors.
manhatten-pads-with-pasta-maker.html

If you don't have the board for the pads you can make them using copper foil or old tin can strips laminated to craft paper. Just use what's on hand and some imagination.



2SK2539 three stage AF amp

I built the circuit from 2 post earlier and it worked very well. I thought I would try it in JFET. This is what I came up with.






Saturday, July 21, 2018

Pulsing and ringing a coil (or earbud)

The question is what would it sound like. My best guess is it would be good since the circuit was presented in magazines and data books but the debate is on going. Let's look at at couple of factors. My first question is why use the unijunction? If we drove the speaker directly from multivibrator how would that work?
The data I present is for an earbud but would apply to any coil. Remember the coil has resistance, inductance and capacitance and therefore has self resonance.
If you pulse with a single spike the coil it will 'ring'. It will oscillate at its resonant frequency and the signal strength will fall off as show above.
If you pulse it with a square wave it will ring as above.
The coil is being pulsed at a low frequency and rings at its resonant frequency.
The pulse frequency has been raised so the ringing forms the sine wave and re-enforces it before it drops off. So you should be able to see the coil will 'shape' the wave and the self resonance of the coil will be the controlling factor. Now the question is 'what is the resonant frequency?'.
This is from a headphone spec sheet. It show the phase angle and impedance. Remember the phase angle is 0 at resonance. so we see resonance at 100 Hertz. If we pulse at a rate less than 100 hertz we have ringing above 100 hertz the next pulse arrives before the coil produces a full cycle and thus no ringing.

This is the response of the unijunction transistor. The 47K resistor and 22nfd cap trigger the unijunction. In the circuit above the supply is DC. In the circuit being discussed the supply is a square wave.  It could be said the multivibrator is unnecessary and redundant. Connecting the battery directly to the top of the 47K resistor will produce a tone. Using the square wave supply will produce a tone control by controlling the charge rate of the 22nfd cap. Does the cap charges in one cycle by the multibrator or 10 cycles? That is the question!

Saturday, July 14, 2018

3 stage AF amp using the Fair radio interstage transformers

 The third line down is the transformer I wanted but I didn't want the $66 price tag, ouch!!
I shopped at Fair radio a little and found this one.
At $2 it was a deal. This is an older one with the same spec's.


This is the circuit I used for my hearing aid amp. It would be a good place to use a couple of the transformers. T1 and T2 are interstage and T3 is an out put transformer. I have a box of them I picked up for 30 cents each at a surplus sale a while back.

The output transformer is not necessary but I don't like putting my headphone in the DC circuit.
I put one on a piece of poster board to see if it was as good as I thought it might be. It was better! It needs the volume pot. Touching the input will buzz your ear.
I put my audio generator on the input and it sang with it barely off minimum output.
Time to find a box and a board for a permanent build.

This is the circuit I used before. I used GS109 and MP40 transistor with good results.


Thursday, July 12, 2018

2 transistor radio in sim

The circuit asked about
The sim. I added C3 and R4 so I could read the output with a zero reference. .1mv in 5mv out. The transformer is step down so the output current level is higher with the lower voltage. Chances are the loading produced by a mismatch would not produce the same output level without the transformer. food for thought? (maybe another sim should be ran with RC coupling?)


Wednesday, July 11, 2018

building a JFET amp no math just logic

I'm going to piece together an audio amp to drive my Low Z earbud and my Hi Z earbud without any math other than in my head (guessing).
I selected the 2SK2539 simply because Arrow.com sells a 3000 reel for $15 dollars with free shipping. A reel should be a lifetime supply for an amateur. You can use whatever JFET you have on hand. The first step is to determine the Q point. I put a 1K resistor in the source circuit to set the bias and a 100k to tie the gate to the circuit. I wanted about 1 ma. This gave me just over 1 ma so far so good. The drain needs room to swing so I select a drain resistor that will drop about half the supply at 1 ma. 4.7K should work nicely. so ready for the next step.
Hows that for dumb luck? The drains Q point is about 4 volts. Now to try some signals to see how it functions.
1mv in 800 mv out.
3 mv in 2 volts out, starting to distort.
5 mv in flat topping. Now to look at some loads. I have standard earbuds which are 60 ohms and a high Z earbud which is 2K. I will make and output port for each.
Yea that will drive my earbud to hard. I am hard of hearing but I can still hear thunder;).
Yes the Hi Z phone will work too. I have some transformers and will probably put one here but it will work without it.
Time for a build!

Saturday, July 7, 2018

Probing the colpitts circuit in SPICE

If you load the oscillator in LT spice and run the sim it can help understand how the circuit works.
 I set the coil resistance as 1 ohm and the reference at the bottom of the tank.

I adjusted the coil resistance until it killed oscillations. You can see the signals levels falling as you come down the sequence. It was a good exercise in SPICE. I will have to build one on board and see if it will sing for me.