Saturday, December 30, 2017

Mounting some dual gate Mosfets

Moving on to the amplifiers. The good news and the bad. They are available at reasonable prices but they are surface mount. The old components were more user friendly but are harder to find. The JFET for example is available for about a dollar each, I recently bought a reel of surface mount JFETs (3000) for $15. The auto industry uses a lot of components that are suited to hobby work, they are low voltage and low power. So the chief designer addresses the issue of small surface mount components. Here is his solution.
He obtained these chad from a circuit board shop. You can uses a hole punch and make your own.
Stick the chad to a board using double sided tape and saw the cross pattern with a small hacksaw.
Sand the protective coat off the copper.
Tin the copper.
A drop of super glue will hold the chip in place while you solder it. Now the question is shall the chad be glued on a board or shall leads be attached and the chad embedded in a drop of epoxy?





Thursday, December 28, 2017

Logic Oscillator options.

The chief designer said he would like to use some logic in the radio design. The second detector oscillator will be crystal controlled and he wants to use a logic gate to make it. This is and interesting approach. What are the requirements for an oscillator?
1. an amplifier overall loop gain greater than 1
2. a feedback circuit providing positive feedback

So the question is can a logic gate be used as an amp? Let's take a quick look at a couple of gate circuits.

For our oscillator the amp does not have to be high gain. It simply has to replace circuit losses.
The CMOS inverter has possibilities.
It seems we have some possible circuits let's look at some oscillators using them.
This one looks good as far as simplicity and component number count. Notice the RC circuit adjust the frequency so it would need the variable capacitor. They are still available so this is a good option.
This one might be better because the two gates provide the phase shift required for oscillation. In the first circuit the RC circuit had to be set to provide the phase shift. In this one the capacitor is trimming or pulling the crystal frequency.
This one is possibly better. It will oscillate without the crystal. The crystal establishes the frequency. (R2 and C2 shunt spurious frequency) Let's look at another version of this one.
The best one I could find. As it says will oscillate without the crystal and has a buffer. If I could just remember where I put those chips I would have to try one.
I made this one and it oscillates with 1-1/2 volt to 4-1/2 volt. It locks to the crystal frequency. I can adjust the duty cycle and shape the wave a little. It seems the pulse train is locked to the crystal frequency. The duty cycles swings from 37/63 to 48/52 but the frequency counter holds steady?
I tried 2.0, 4.0, 6.0, 8.0, 12.0 mhz crystals and it worked with them all. I could adjust the duty cycle but not the frequency with them all. The Texas Instrument App Notes shows the circuit and gives info on adjusting the operation. They show sine waves being produced. I am seeing square waves. The chip I'm working with is a 2 input NOR gate,single 5 pin device. Maybe I'll try again with the proper IC. It will be next year before I get some.

Sunday, December 24, 2017

how can one amp be used for AF and RF? or Reflexing

When you look at a broadband amp such as the antenna amp on your TV do you understand that the amp is amplifying channel 1 and channel 10 and channel 100 and all in between all at once? If so it should not be to difficult to accept one amp can amplify AF and RF so I will simply show the circuit and offer a brief, hopefully logical theory of operation.
First the basis for circuit analysis. First step define DC biasing. More often than not getting the DC established is the biggest job the rest falls into place. Now the circuit in question.

Mr Andersen says it works so we will accept his word for it and try to see how it works. Again the DC bias is the first step.
In order for Q1 to conduct it requires an emitter base current. Starting at ground up through the 1K resistor and out the base. Now how to get to the battery? Um, into the center tap and out the bottom of the coil through the diode. One comment was the diode adds a needless extra 0.7volt drop because the emitter base junction is a diode. Hopefully Mr Andersen is smarter than that! Anyway on through the 33mh choke and the 3.3k resistor and switch to complete our path. Now we have established a DC bias and the transistor is switched on.
What do we know about transistors? One thing is they amplify current. A small current through the base produces a large collector current. The amount of current gain is the transistor Beta factor. If ß is 100 and we apply 1 micro amp through the base we get 100 micro amp out the collector. Be aware the RF transformer is in series with the base current and look at the signal on the transformer.
If you can accept that the diode acts as a switch it become apparent the RF is chopping the transistor current and producing a larger version on its collector. So if the antenna can supply a 1 micro amp current to the diode it is being amplified to produce a 100 micro amp collector current. BUT only on the positive half cycle. The collector sees the center waveform and the choke/cap filter it to produce the bottom waveform.
Does that make since?
Just imagine they placed a switch in the bias circuit which is switching at an RF rate. HEY! That might be why they call them switching diodes.
The transistor selection could effect this circuit a lot. I have some with a ß of 1000. Um 1 micro amp to 1 milli amp would be a giant leap in signal?!?

EDIT: As base current increases collector voltage decreases reducing the diode bias. In case you did not see that.That would be the feedback or reflex function.

Thursday, December 14, 2017

Filter after Xtal filter

In my post on the Hi/Lo filters I posted one with the capacitor in series and one with the inductor in series. We could use the one to block DC and to other to pass DC. It would be a circuit design issue that would determine the filter to use. Looking at my last post you may see the filter in the Xtal output circuit?

Look at L6 and the 47pfd and 2000pf cap. Can you see a PI filter? The way it is drawn you may see a tank with center tapped capacitors as in a colpitts oscillator too? This is not the case.

This is what I wanted you to see. I used 20p and 2000p. I wasn't looking at the schematic just building a circuit to look at. Now what is the response of this filter?

Tom said his inductor was 100 -200 uh so I used 100 uh and 200 uh and made two plots.

 I should do some more with 47pfd as Tom's circuit did. You can see if the filter is a low pass circuit it would work as we had planned our. It could pass 80 meter and block 40 meter for example. In this circuit it is modulating the screen grid voltage. We could use a DGMOSFET and feed G2 with the filter output and have the same effect.
If a fellow had Elsie (LC) installed on his machine it might be an interesting exercise to run the numbers on the circuit feeding the Xtal filter. It just may be a bypass filter on the Xtal filter input?

Hey that was our plan a high pass on the input shunting it to ground and a low pass feeding the Xtal!!! They stole our idea 65 years ago!!!

Seriously if you look at the circuit as having a PI filter feeding the Xtal and a PI filter exiting the filter I think you will be on the right track.


Xtal filter question sparked by TRB post

My first receiver was a Hammarlund HQ-140 the manual can be downloaded here:

Hammarlund HQ140

The set that caused the question can be found here:

Simplex Super

They use a Xtal filter with only one Xtal.
Here is the Xtal filter and brief description from the linked manual.




The circuit from the schematic.


The Hammarlund predated the Simplex. The Simplex was published in a ham handbook so the simple statement would be they modified the commercial product for their project. (Hammarlund held the patent)
The question was about L5 and L6. They do not resonant at the operating frequency.

No, they are filters not tank circuits. See this post filter circuits. If you download the manual and read the alignment section it will explain the circuit operation.

These filters shape the overall filter response and keep undesired signal from passing. (The crystal will pass harmonics and overtones which are undesired.)

Tuesday, December 12, 2017

Hi/Lo filters

The chief designer said we would save the VFO to last. Build the test bed and give it a trial first. Using the scheme we can use any available Xtals for the filter and second mixer then adjust the VFO to fit.


 The basic scheme relies on the image being removed by the input filters. So we feed one filter to ground to bypass the image and one to the receiver. (RF amp or 1st mixer)
The high pass filter passing 40 Meters to its destination according to which band is selected.
The low pass filter passing 80 Meters to its destination according to which band is selected.
So one band is rejected and one is received.

 I left the credits for the program I'm using. It would be worthwhile to download it if you have need for filters.
You may have noticed it says manual entry. I ran the analysis and then manually adjusted the components to see the response with values available.
We plan to use the BF2040W RF amps and they are 50 Ohm devices so I ran the numbers at 50 Ohm.
The BF2040W is a high gain stable amp rated up to 1 Ghz. It could be used for the RF amp,IF amp and mixer.


Monday, December 11, 2017

Some oscillators to consider

The original Oscillator that we were looking at. PNP version.
The NPN version.

This one is actually a FM transmitter. It could be used as a starting point.I have a pile of the transistors. They were made for UHF Osc/Mixer in TV so should be good for our purpose.

Dual Band 80 meter and 40 meter Receiver


Sunday, December 10, 2017

Block diagram of the receiver

This is one possible layout. It has only one variable, that's the VFO. By using a band selector filter we get image rejection. The band select consist of a high pass or low pass filter. The second mixer/detector uses the same frequency Xtal as the filter. 


I have seen a set that used a NE602/612 for the second mixer/detector.

The tuning response will be determined by the VFO so a bandset and bandspread would be a good thing.

 This is a simplified diagram so it would be good only to understand the frequency conversions. I left the AF amp out and we may need an RF amp and an IF amp.

Now it is time to determine the receiver tuning range and the VFO frequency. Building a stable VFO will be the main project. 

How broad a tuning range do we want for the VFO?
How many bands?


commercial filter question?

I was asked a couple of questions that I hope this will answer.

This is the filter in a kit sold by BG Micro. Dave White's "beginner's and Experimenters" it uses 10 MHz
Xtals.
This is a partial schematic for a now obsolete commercial radio. It used the same basic circuit for the different bands and The 6.xxx,8, 11, and 14.xxxxx filters. The chart is for band selection.
I found this one on E-bay for $30 US. I think we can make our own for less. If the junk box has no crystals in it you can buy some for less than a dollar each.
When you get your feet under the bench it would be worth a little time to lash one up. A simple circuit with the generator feeding the crystal and a 50 Ohm resistor in series and measuring across the resistor. If you don't have a meter to read the signal with you can do as I do and put a diode feeding an AF amp and use a modulated signal to drive the loop. It is quite touchy as the crystal has a sharp resonant response. When you try 1 Xtal add another and see how it responds. If your generator has any back lash in the adjustment you may find the adjustment is very touchy by the time you have 4 Xtals in the circuit.

Edit: I saw some 16 Mhz Xtals on E-bay 5 cents each. (20 for 99 cents)

Saturday, December 9, 2017

Xtal filter and pre-selector

I did some rather unscientific testing and found another option which maybe even better. If you want practice winding toroids it is for you.
This will be the Xtal filter pre-selector. It has a response like this.
The previous post had a pre-selector to keep the LO out of the RF input. This pre-selector will keep lower frequencies from the Xtal filter.
With this circuit the Xtal filter can be 2 Xtals in a TEE arrangement with 3 capacitors. So we can wind 3 coils and use 2 Xtals.
When I hear from the designer I see how this idea flies.

Doing a little filter eval

I played with the filter programs a little and here is what I found to be one possibility. The main problem with the direct conversion receiver is the LO signal goes in the antenna and overloads the front end. A dual conversion would eliminate this problem. I have a box full of 25.125Mhz crystals so that is what I'll be playing with. The receiver will be from 0 - 20 Mhz. First a filter for the front end.
I told Elsie what I wanted and presses the calculate button and here we are. Click another button and see the frequency response.
So it is down 20 db at 20Mhz and I'm looking at a 25Mhz IF. Think that should keep my LO out of the frontend? Now for the IF filter.
I told the ladder Filter Program (LFP) I wanted to use 4 Xtals and an little about them. Now for the filter.
My post from yesterday had a page of possible filters to design. I copied this from that page. N=4 four Xtals. The component labels ID their location.
The print out list the components values. So moving the values to the schematic gives this.
The values are in picofarad. Now it would be interesting to put the circuit together and test it. I think I have some 1nf and 900pf caps so I may go with 2nf and 1.9nf for a test.

Friday, December 8, 2017

Made a tool for scribing the CB pattern

I mentioned how to make a tool when I made the last board.The groove will be wider with a hacksaw blade tool. When soldering the board I decided the wider groove is desirable so I made one today.
This is the board I made yesterday. It is ok but the lines are thin.
I simply took half a used hacksaw blade and ground the teeth off, rounded the end and put a small notch on the end. I just took a pencil and made the layout. there is no need for measuring because the dgmosfet will set on the cross. I pulled the tool on the lines and made the grooves. Then took my ohmmeter and checked for the lands being isolated.
This is the board I made tonight with the tool. Holding the tool like a pencil with the board standing on edge at about a 45 degree angle I lightly follow the lines and then go over them again. After a few strokes the copper is gone and the groove is formed.
Booth boards and another look at the tool. It take a light stroke to layout the groove in the copper. Then a few more light strokes clears the copper away. Once I hit the board I hear and feel it. a couple more strokes to clean the bottom of the groove.
So I now have a board with 2 dgmosfets and a board with 4 dgmosfets. When the design is decided on maybe we can put them to good use.

Maybe an RF amp and a mixer with a LO buffer and an IF on the one board. We'll need an AF amp and LO. The crystal filter for the IF and a bandpass filter for the input to the RF amp?

Looking at Xtal filters

Xtal Filter Calculator

This little program is for designing Xtal filters.

This is a list of filters to chose from.


It could be like this.

Or a bit more perhaps like this one.

Once you "design" it the program will produce a response curve such as this.
It is only 640Kbits so not to resource hungry.