Sometimes Things Make No Sense

Yes sometimes things make no sense at all. Typically because you stuffed up in the first place. In this post haste modern world I got my Ying mixed up with my Yang and well things go bonkers when you do that.

So I am still trying to work out why my inductors are not right, but I have finally started to get some results that make sense and fit with that I am seeing with my measurements. So my dilemma has been my low pass filter designed for 7mhz had a center frequency closer to 20mhz, obviously something is wrong. I checked all the cap values as I was building it, they were all correct, I wound the inductors with the correct number of turns for the inductance needed, check them with an LCR meter and they seemed to corroborate. There were no solder bridge’s or cold joints. So the obvious place to look was the inductors.

So i knocked up this test jig, the inductor is in series with a known value resistor. V1 is a signal generator, the frequency is changed until the voltage at V_2 is exactly 50% of the voltage at V_1 and the frequency of the signal generator noted.

So I did this 2 times, the first time above using a 99.7R resistor, that was the measured value of the resistance used and the above was measured on the oscilloscope. The blue trace is the input voltage V_1 and the yellow trace the 50% measured voltage V_2. The frequency on the signal generator was 6.3mhz. This scope if not very accurate.

I repeated the process using 199R resistor and noted its frequency, the measured inductance was 4.3uH and 4.5uH certainly near enough to the method was working as it is meant to and well within the margin of error for this type of ball park measurement technique.

So with the frequency of the 50% voltage we can use this formula to find the inductance. L= R*sqrt(3)/(2*pi*f) which works out to be 4.3uH or near enough

So with my new found inductance value, I simulated the filter this time using 4.4uH for the inductance value mid way between my 2 measured values, and the simulation plot started to look a lot like what I was I was seeing with the bode plots i was making of the filter. So it looks like I have found my culprit. The actual inductance was lower than what would have been expected for the number of turns I had on the toriod. A few more turns and it should be right to go.

Facebooktwittergoogle_plusredditpinterestlinkedin

Switch Mode PSU Noise

So a few weeks back I got making this nice looking lab power supply, but other than running my bench LED lights, its not been used. And thus begins my world of hurt, because the other day I had a project that needed some power and after connecting it up, things just were not doing what you would expect them to do.

 

A circuit that should give a nice sine wave looked a mess on the scope, a receiver I powered up had this god awful switch mode racket on just about everything. So get get the pixie wrangling gear out to take a look at what was going on and it was nice a pretty picture.

 

 

As you can see from the scope output, the 12v DC was not pretty and it was making my projects not happy. So i figure, It needs a filter and I do 30 seconds of math and figure 10 to 100uh of inductance with 100uf of capacitance should be close to enough to the business.

So I make a nice looking filter and connect it up to the power and yes it attenuated the noise and ripple but it really did nothing else, all the problems were still there.

 

When you know you are chasing your own tail, start talking to the smart people. In my case this is Brenton, and we got talking about all aspects of the design, what modules I used what SMPS i used and a bunch of other specifics.

My pretty Power Filter

Well it turns out, had a fundamental flaw in my PSU design, I had left the 0v DC floating and this is a trap for young players. Because in the words of someone smarter than me  “Without that connection the output is magnetically isolated and noise leaks out via capacitive coupling in the output transformer.” So after checking that 0v DC was not tied to Mains Earth, they were quickly coupled together and all my problems went away.

 

And as you can see from the first scope output above, the power is now rather clean, and in the 2nd scope output, is the DC after my new filter. The filter will be installed permanently in the next few days and will remain a permanent part of my lab power supply. And with that fixed, we can get onto building a receiver now for the 10w CW transmitter.

Facebooktwittergoogle_plusredditpinterestlinkedin

CW for Bitx 40m

So whats missing from the Bitx 40m SSB? CW of course. And there is a fix, a tone injection board to put a 700hz tone onto the audio. The board came from Fusion Radio on ebay. Board quality is good, build difficulty is easy. I just do not have any 1n4001 diodes in my parks box, I have 100 of just about everything else you can think of, just none of the required. I will pick some up from a local electronics retailer Jaycar tomorrow and then add power.

Facebooktwittergoogle_plusredditpinterestlinkedin

9mhz IF Filter

9mhz_if_schematic
What to do when bored? Experiment with some bits you have sitting about of course. Its been so long since i have experimented just for the sake of experimenting. So in many designs you see the 4 crystal ladder filter with various value capacitors of the same value. So i grabbed 4 9mhz crystals out the parts box and some 100pf capacitors and soldered them up in the ugly style. Connected up the signal generator and scope and began to plot out the shape of the filter.

img_20161021_120521
The basic setup.

9mhz-if-filter

With 100pf caps a filter with about 900Hz -3Db width, not sure why the right hand side is not a smoother curve, i probably need to match impedance’s and use better quality caps.

9mhz-if-filter-2k

Next i changed the capacitor values to 50pf, and plotted out the filter shape again, this time converting millivolts to Db to see what sort level of attenuation is being achieved, and after some stuffing about to get my math right, this is what i ended up with, well 3rd times a charm, to get the math right HIHI. Kind of looks like a filter now and not a Mr Squiggle drawing.

 

img_20161020_173841
The test setup.

Facebooktwittergoogle_plusredditpinterestlinkedin

K16 Keyer

I have had this K16 keyer sitting about for ages and thought it was about time i put it in a box and started to use it together with other projects. I have not yet added in any buttons for the auto call and memory functions, I do not really use them all that much for the types of operating i do. Maybe I will add them in at a later date.

img_20161008_112333

Working out the basic layout.

img_20161008_115444

Adding in a rainbow of wiring.

img_20161008_122014

The switch operates in both the on and off positions.

img_20161008_143941

Yes it all fits kind of.

img_20161008_144601

Pigeon pair, K16 keyer and Pixie Transceiver.

Facebooktwittergoogle_plusredditpinterestlinkedin

Measuring Amplifier Output Impedance

13533329_1340106839336356_6538344959771332536_n

Measuring the output impedance of an amplifier does not require technical gear like a VNA, though that will make for a nice quick and accurate measurement. You can also do it using your multi meter and a scope. Here is how:

Connect the output of the amp to the swiper of a 10K pot or better yet, a decade resistor box.

13528977_1340107616002945_8318913652801389010_n

And measure the resultant waveform on the scope.

13510895_1340107646002942_4297484531567183411_n

Starting with the pot set to 10K ohm, adjust the pot until you have 50% of the amplitude of the signal.

13483003_1340107769336263_7781115297360550022_o

Remove the pot from the circuit and measure its resistance, this value is equal to the output impedance of the amplifier.

New Bitmap Image

You can then use a transformer turns calculator to determine the turns ratio for transformer for your circuit. Not hard now is it.

Facebooktwittergoogle_plusredditpinterestlinkedin

12v 0.5A Solar Panel

Its been a while since i have updated this space. I have been waiting on parts and kind of lost interest with my home brew endeavours for a bit, but i still find some time to tinker. I finally got the rest of the solar panels I had on order and have wired them up, 2 panels paralleled 4 pairs in series. Theoretically we should have 12v 8watts or 0.6amps.

DSC_1928

So I got some sun today, being winter and 4pm in the late afternoon, to see 300ma peaking on a panel that should max out at 0.6amps is quite nice. With some sun intensity we should see close to the max being made, and certainly over 0.5 of an amp.

DSC_1927

Unloaded voltage is about 13.8v, so everything is looking nice, now i just have to mount it all up on some plastic sheet i have, add some plugs to the wire, series diodes so the battery does not flow back into the panel and finally a volt meter and ammeter as I will be the charge controller and over volt protection for the AGM deep cycle,

Facebooktwittergoogle_plusredditpinterestlinkedin

Solar Panels for Homebrew Projects.

I have been playing with some solar panels i got off ebay, they are 6v 1w. Its winter here currently and the sun is not as intense as it can be, so its a good time to see just what these panels can do and how close to their ratings they come. Surprisingly, pretty close to specs.

DSC_1896

Started with a single panel checking voltage and ampers. Voltage looks good, 7v on an open circuit, and close enough to 6 under load.

DSC_1895

Single panel was peaking about about 130ma, near enough is good enough.

 

DSC_1894

Put a single panel on some very dead NiZn 1.6v per cell, cells i had laying about, and in a few hours all were reading 1.5v per cell, which was nice because I will use those cells with 2 panels in parallel to keep them charged for an Arduino Weather Station im currently building.

 

DSC_1897

Next i paralleled all the panels up, not facing the sun in an ideal way or anything, voltage was close enough to 7v.

DSC_1900

The for panels in parallel gave a nice 0.5a of current, so i got to thinking about series and parellel up 8 of these and make a nice 0.5a panel for the HB1B YouKits cw transmitter.

 

DSC_1902

Series up 2 panels. 13v, perfect.

DSC_1903

130ma, peaking 150ma in winter sun, 8 panels should give a nice 0.5a and keep the 7ah deep cycle well and truly topped up all day.

Facebooktwittergoogle_plusredditpinterestlinkedin