Saturday, August 29, 2015

Calibration Loads for the Sark-110

I finally got around to creating some calibration loads for the complete alignment and OSL calibration of the Sark-110 analyzer.

I was mainly waiting on the 0.1% 0603 resistors to show up from DigiKey and the weekend to actually work on it. After all was said and done, I ended up with a compact single piece unit to keep track of instead of 5 little pieces.

Back to the junk drawer for the edge launch female SMAs and 0.062" board. It all went together fairly well but did have to use the drill press, Dremel tool, hand knife, nibbler, and an awl to carve out the board near the loads. The short was a round piece of copper with a hole drilled in the center and soldered down to the connector.

After calibrating my AA-600 to some know good OSL standards I checked it against this little fixture and all seemed okay. The 50-Ω load was 2 parallel 100-Ω loads and it turned out better than I expected. The 200-Ω load is a single 200-Ω load and the 100-Ω load is a parallel combination of 2-100 Ω 0.1% 0603s. The 50-Ω termination measured at least 45dB return loss at the highest frequency of 230MHz. Agilent's (or Keysight rather) threshold for CAL loads is 42dB return loss. Check.

After performing a full detector cal, all seemed to be functioning as expected. 


Friday, August 28, 2015

Good Cap, Bad Cap

So I recently purchased the Sark-110 antenna analyzer but really wanted to explore the world below 100-kHz and more specifically the world from 10-100kHz. The reason is because Electrolytic capacitors often exhibit some interesting properties in that frequency range. So, I downloaded the "b" version of the latest firmware from the Sark website.

The firmware updates are just the easiest thing to write into the unit as it's never more than a few button pushes and "beep", you're done! With it installed I recalibrated the detector complete with it's full detector alignment mode using my newly created 100 and 200-Ω loads. This by the way, this is very easy. If I, as a somewhat computer illiterate person, can do this, anybody can.

Enter the ESR measurements. I was really hoping I could find a bad cap in my junk drawer and I did. First, I set the unit to display series resistance on the (green) left vertical axis and series reactance on the right vertical axis (Xs). Later changing the right axis to the |Zs| impedance magnitude because it tells you what the total impedance is taking the capacitive reactance into account as well.

The first photo is what a good cap looks like. It's a similar vintage electrolytic 7.5μF cap. The bad cap was marked 4.7μF and looked like it had seen better days.

The maximum values (yellow marker) Rs and Zs occurs at 10kHz and is 0.81 and 2.48Ω respectively. Minimum values (blue marker 2) of 0.49Ω and 0.5Ω  at 923kHz.

Then there's the bad cap. The Rs is 1740Ω at 10kHz, and the impedance is a whopping 15kΩ!

All that's left is to recalibrate my OSL memories to the new, gladly welcomed, extended frequency range that includes the 10-kHz low end.


Saturday, August 22, 2015

Too Busy Playing

Pretty much ever since I got my Sark-110, I've been busy playing with it. Just getting to know it. One thing I found is the left/right jog switches also depress straight down to activate the current selection. 

Another thing I've noticed is that when in d/s mode (double/slow) OSL speed is reduced. Could it be that extra info is stored? Dunno.

Here is a data screen for the Transmission Line function located in Mode/Single Frequency/Cmodel/Transmission Line. I had 93 feet of old Belden 8210 laying around that seemed to have more loss than I wanted. After measuring it with my AA-600 awhile back I thought I'd sweep it again. Yep, same loss numbers.

More Later.


Sunday, August 16, 2015

Sark-110 and TL

The Sark-110 has a few neat features that make it easier to measure a few things. For one, the Transmission Line function in the Single Frequency/CMOD menu. It allows a person to come up with the maximum and average matched line loss and at the frequency it occurred, velocity factor and the frequency it's valid, and the characteristic impedance of the TL, in about 2 seconds! My Rigexpert AA-600, AC6LA Zpolts, and a computer can do the same thing but taking quit a bit longer.

Enter balanced transmission line Belden 8210, 72Ω transmitting line, or so it's called:

Notice I ran the frequency up to 30 MHz in one sweep and 14MHz in the other. This is because the loss slope is reported at the max values only. So if you want to see what the characteristics are at your freuency of interest, you can set the stop frequency at say 20m, as I did here.

You do have to know the TL length, so get your tape measure out.
BTW, this line must be bad. The loss is high and when checked with my AA-600 and Zplots, it revealed  the same thing, 3-4dB of loss! Yikes.

On another note, when measuring in the Scalar Mode, selecting the tracking function of the markers is pretty handy. One feature that stands out is the ability of that tracking to search to the "next peak" by highlighting the "Marker 1" (or Marker 2) field on the left side in the main window, then using the leftmost rocker switch, you can go to the right or left with a flick of that control. This control also affects change to other modes or fields as well. Like if you were on the frequency field, flicking that switch will increment or decrement the center frequency. Likewise while in the span mode, and even scroll through the available measurement parameters when highlighting over the Left Y or Right Y axis fields.


Saturday, August 15, 2015

SARK-110 First Impressions

Before my first impressions are gone I've decided to let these subsequent blog posts become more a stream of conciousness than literary works of art. This way I can convey more important distinctions that jump out at me as they happen to preserve the fidelity of how they presented and why they are worth blogging.

The Sark-110 is not for the faint of heart, or eyes. This thing is TINY! If you do not have good eyesight you should stop reading here and forget about investing $400 on this unit. Even with computer control you still have to navigate through a menu with micro-point font on the display to get there. But if you can work with 0402 components with relative ease then this analyzer can revolutionize how you perform S11analysis. Yes, that's a credit card sized overlay of an iTunes card laying on top of the analyzer. That's a stock Kenwood TH-F6A rubber duck antenna at the top of this photo.

My burning question answered first off was, "Can the OSL calibration "stick" to the frequency range and span and not cancel when you QSY, even one step?" (A feature of the Rigexpert and HP/Agilent/Keysight series). Yes, a resounding, YES! In fact, Melchor has improved they way it ought to be done by making the cal stick to the frequencies cal'd no matter what span or range you dial in. That means if you perform an OSL cal to a frequency range of 0.1-30MHz and dial in a display range of 3-50.5MHz, you will see calibrated readings from 3 to 30MHz and uncalibrated readings from 30 to 50MHz. The CAL won't disappear when you jump around, you can't make it go away unless you delete it through the menu somehow. There are 8 CAL memories, wahoo! Finally, a device that can store different CAL configurations to allow a person to have a calibration for different lengths of coax, adapters, or frequency ranges. And that's another thing there are 400 local calibration frequency points, no matter what frequency or span you've selected, as opposed to only 80 on the Rigexpert. I say local because you can store any number of OSL or span points with the RE when using the computer.

The next thing that I wanted to see is was how the antenna connector was implemented. Is that MCX connector up to the task of day to day use? Well, it seems that the choice was sound because of the engineering trade space involved on such a small device, you just had to go to a smaller implementation of getting RF squirted into and out of the box somehow. Most ratings of RF connectors are based on how many times you can connect and disconnect (one cycle) the connection without causing the DC contact resistance to exceed some specified limit such as MCX's 1 milli-ohm for the outer contact and 5 mili-ohms for the center contact. It has a rating of 500 cycles minimum. Now keep in mind there are no RF connectors out there that have any ratings greater than 500 so just think of what that means to how we look at Type N, BNC, or UHF connectors. It doesn't matter in the end because if we did exceed the cycle count, it doesn't mean we can't use the connectors, it just means that there is a chance that the contact resistance can exceed the mΩ specification limit. That won't bother us because we can calibrate this out or it becomes part of the reading. An additional 20 mΩ of series R, or whatever added to our readings. Just get a cable assembly with an end you can use and forget about it. The RF accessories supplied with the analyzer are a straight MCX male to SMA female adapter (an optical reflection illusion on the website creates an impression of a RA adapter) and a 6 inch SMA male to SMA female RG-316 cable assembly. I did purchase the BNC adapter, the neat little Aluminium stand, and the MCX cable with the small clip pig-tail ends. I haven't broken out the BNC adapter yet.

Can you run right angled connectors to the -110? No, there isn't enough clearance between the housing and the RA connector body to allow a fully inserted connection to happen. Straight MCX males need only apply. But don't worry, eBay and Amazon are full of adapter and cable assembly options, I chose to use a 12-inch MCX male to SMA male RG-174 cable assembly because it's what I found in my junk drawer.

So how does it work? Well, so far so good. Below is a photo of my 20m W3NQN mini filter and the VSWR (green) and |Z| (red) response. These candid photos are not going to blog well.

Another feature of the RE that bugged me, the inability to adjust the VSWR scale higher than 10:1 locally or 6:1 computerally. With the -110, that limitation is gone. The vertical axis scaling is accomplished in 4 ways or levels of scaling accessed through the menu:

In the VSWR mode:
"Normal" 1:1 min / 25:1 max vertical axis,
"High" 1:1 min / 99:1 max,
"Low" 1:1 min / 10:1 max, and 
"Auto" an auto-scale display of the current sweep that I've seen range up past 100:1. I've selected "Auto" ranging for my activity so far just to keep things centered until I get used to the analyzer. And, I believe, these can all be changed with the computer if you want.

How does that accuracy compare to the Rigexpert AA-600? I haven't performed any comparisons yet but "A man with two network analyzers knows no impedance. -M. Schaffer".


Saturday, June 13, 2015

Field 2015 Day Dry Run with the PPD-EXT-FD

"Wow"! That's the first thing I said after setting up my Field Day 2015 Park Portable Doublet and getting the first impedance reading from my AA-600 attached to the feedpoint, just wow.

I calibrated the AA-600 the night before with the Micro-Balun set, at first, to 4:1 and using a 200Ω carbon composition resistor but had some reservations about the extra capacitance found in these "Lil Devils". It was significant enough to warrant loosing the high Z accuracy and going back down to a balun setting of 1:1 and suffer the measurement inaccuracies using the well behaved HP 50Ω load resistor. That way all of the impedance measurements are displayed directly with no need to multiply by four.

My friend Gary, KØIMJ and I went off to what my kids call "The Purple Park" because some of the playground equipment is purple, set up the 28' Jackite in the center and the 2-20' Jackites as end supports, using the Park Portable Doublet construction method described in earlier blog posts. BTW, the lawn at the park was so lush because of the recent rains and very squishy and sloppy with flip-flops on. :(

I drove an Army tent stake down and Velcro'd the 28' Jackite base to the stake, while stringing 95# paracord guys. (Obtained from Hobby Lobby).

I took readings with the calibrated AA-600 center frequency at 14.1 and span of 28.0MHz making it sweep from 0.1-28.1MHz, with 80 data points (the native data point number for the AA-600 not tethered to a computer). The measured results (cloumns E and F) were very close to EZNEC predicted values as shown in columns A through D. 

Column G is the settings on the BLT where the clips are attached to the coil turn number from the center tap. I made a nemonic of the reporting of tap position with C-A-P representing where the tap points are with the (C) of the air variable capacitor tap point, (A) meaning Antenna tap point, and (P) being the switch position on the Elecraft BL2. The (+HC) is the addition of the Helper Cap in parallel with the air variable to help it out when it ran out of capacitance. The HC is an ARCO 427 set to about 75pF yielding roughly 175pF of capacitance needed to resonate the tuner on 80 and 60 meters at their respective tap points. The air variable was about 3/4 meshed.

In days prior to field deployment, I created discrete loads representative of the antenna's input impedance for each band, I used an air variable cap, a pot and some ferrite cores to create these discrete loads. I measured each load and attached it to the tuner achieving a match for each band, noting the C-A-P positions. These were my presets I used to get the tuner close when I set up out at the park. I reported only the final settings in the spreadsheet. The 'wow' factor came when the lowest SWR reading was dead on or only one turn different from the preset.

The RX graph above shows the periodic nature of the impedance (displayed in resistive and reactive componenets) as a function of frequency. Column E in the spreadsheet is color coded to show that the R or X (or both) can change quickly by varying the frequncy slighly over areas where the slope is the greatest. This accounts for the error between the calculated and measured impedances shown in the spreadsheet because the reading taken didn't line up exactly with the calculated value.

Results from modeling a Park Portable Doublet set up for Field Day 2015 are shown and align well with measured results taken with realatively inexpensive test equipment.

Carefully constructed EZNEC models can show a high degree of fidelity with the physical antenna characteristics. The model is validated by actual measurements that are in good agreement with predicted values.

Myron WVØH
Printed on Recycled Data

Friday, May 29, 2015

Back to Back BLT Test (Again)

Because I couldn't leave well enough alone, I had to retest the Elecraft BL2 BLT back to back with the Mini-Balun BLT. Same results. Low. How low? Read on.

I am planning for 2015 Field Day and there is a chance that it may rain. So I decided to enclose the remaining BLT in a Lexan weatherproof case. It was waterproof, but since I drill holes, it now "weatherproof". Now I will take two along for redundancy. My main antenna is the Park Portable Doublet, and it provides a high impedance at 20m, so I will tune the matchers for 14.050MHz. Here is when it is matching a 15-Ω load. An input impedance seen on 80m with the Park Portable Doublet. That's an ARCO 427 helper parallel capacitor in there to aid the air variable in achieving a match at these lower frequencies, set to 125pF here.

My back to back setup involves calibrating a through path to establish a reference power level with the Kenwood TS-2000 to a Bird 4311 (the older peak reading 43), and a 100W dummy load. 98-Watts, fine enough.

The procedure for getting the tuners aligned is as before, adjust each one separately with a 4.7kΩ load and then attach them back to back. After doing so, it is necessary to tweak the tuners just a little to dial-in the SWR. 

Once that is completed, we are now ready to attach the 100-W rig and play arcy-sparky! Let's hope not.

Good, it's not arcing anywhere. What's the power coming out the other end? Um, 83 Watts. Okay, what's the loss through the entire contraption? 

IL = 10*Log (98/83) = 0.721dB.

Well now, that's not bad. But that means that's for two BLTs. Yes, it is and it would be unfair to say that is an equal split because of the larger balun core and air-dux coil of the Elecraft BL2 BLT versus the MiniBalun BLT and T106-6 core. So what should the ratio be? 60/40? 70/30? 

So looking at the Smith chart S11 response of the matchers when open circuited at the output, one can see how well the response "hugs" the outer circle, indicating high Q and low loss. The BL2-BLT edges out the MB-BLT by a neck, 70/30...okay, so it's not that precise, but it's not a bad guess.

Okay so 70/30 passes the sniff test. So that yields an insertion loss of 0.5dB for the MB-BLT and 0.22dB for the Elecraft BL2-BLT. Neat!