KB9JJA’s Building of Scotty’s 0 - 3 GHz Spectrum Analyser

Collection of info on the “Filter Plus Group Buy” (Feb 2011)

Zip file provided by Jim with all the boards and schematics

PDF File provided by Jim with all the boards on the Panel

A group buy was put together by Jim W1JGH in Feb of 2011 for a group of Filter Circuit boards, plus a few other boards for use with the MSA.
The group buy is now over, and the boards are no longer being offered for sale.

This page is a collection of information that I gathered for my use from multiple sources. Most notably documentation that Jim Provided, Scotty’s and Sam’s web sites, as well as postings on the list reflector. I thought that some of you might also be interested in the info so I decided to make it a web page instead of a word document. I make no guarantee to the correctness of the info presented here. Use at your own risk, and remember that some of these boards that were offered in this buy were considered experimental.

If you have any info that you think would be good to add to these pages, or find any errors please send a message to me at mycall at gmail.com

73 de KB9JJA/Dale

Pasted Graphic

5 Pole Dishal Filter Board

All filters have provisions to utilize surface mounts trimmer capacitors. The series TZY2 trimmers from Murata (available at Digikey) has prices comparable to the good quality ceramics., but the boards have pads that also allow for the larger 1/4 inch sizes. The ceramic filter pcb is a variation of the one available in October's group buy, it is slightly larger and also includes pads for trimmers. This board is well suited for wider bandwidths.

4 Pole Filter Board

Ceramic Filters

Hi All, I have been playing with some Murata ceramic filters that Jerry gave me. They are 10.7 MHz, with input impedance approximately 330 Ohms. I put 2 directly in cascade with no coupling reactance. The input output matching consists of a series 1.8 uH and 100 pfd in shunt next to the ceramic's input (and output). No critical tuning, just standard components. The results are surprisingly good. I am amazed at the isolation for just two filters. Each single unit is actually a two pole (internally cascaded)filter. Therefore, two units create a 4 pole filter. The screen shots and one pic of the component is on the Photos page under my name.


Sam made mention that you could not install these types of filters on the bottom of the Crystal Ladder Filter PWB. He probably should have added, "unless you shield each one of the ceramic filters". A recent post explained a very good method to do this. "Desolder an old crystal and use its case as a shield placed over the ceramic filter. Then, solder its perimeter to the ground plane". Scotty


Upon further review, Scotty's ceramic filter appears to be  part number SFE10.7MA5-A (old part number).  The new part number is SFELA10M7FA00-B0.  See page 3 of the file:  Murata PartMarkingGuide.pdf.
The input/output impedance is 330 ohms, as Scotty said.  The typical bandwidth is "within 280kHz ±50kHz". See the files:Murata SFELA10M7F00-B00_1.jpg, Murata SFELA10M7F00-B00_2.jpg, Murata SFELA10M7F00-B00_3.jpg.

Jim McLucas


Resolution filters (Final Crystal Filters)

If you just want a filter that can let you determine whether everything else is basically functional, you can build a 60 kHz bandwidth filter very simply, by cascading two ceramic filters and adding impedance matching components. The results are shown here:

That image used resistors to match, but you would want to use 1.8 uH inductors (Q of at least 20) leading to the filters and 100 pF shunt to ground at the filter terminals. You likely could get by without any fine tuning, though you would probably want to tune it later.

The wide filter limits the resolution of your scans, but there is very little chance that its shape will come out totally wacko, so it at least lets you see whether the rest of the MSA works. And when you have a full set of RBW filters, this can be one of them. Sam W.

Filter Matching

I tried some additional filter matching with a 75 kHz ceramic filter, Murata SFVLF10M7LF00-B0. This is the one that Digi-Key lists as 25 kHz, which would be nice if it were true. But a combination of two of these makes a nice RBW filter of about 60 kHz bandwidth. I did the matching experiment with the MSA software, using S-parameters provided by Lin.

Ceramics never seem to be specified as requiring any reactive termination, though the data sheets show that they are measured with some capacitance. That capacitance is just part of their measurement setup; they don't specifically recommend that you use it.

The shape of the peaks of ceramics is often very ugly, like maybe a tree fell against one side of it. This particular filter, when matched to 330 ohms wasn't too ugly, but was very asymmetric and not very flat at the top. I found that putting 10 pF on one side and -15 pF on the other made a significant improvement, with a much flatter, symmetric peak, and a bit less insertion loss (less than 3 dB). (If the negative capacitance throws you, see the document in my prior post.)

My guess is that no specific terminating reactance is recommended in the data sheets because sometimes you need a capacitor and sometimes you need an inductor (which is what the -15 pF represents). The fact that no terminating reactance is recommended does not mean that a terminating reactance would not be helpful. It just means that you are on your own in figuring it out.

Let's say you are matching a ceramic filter to 330 ohms by using approx. 106 pF shunt on each side of the filter, with 1.8 uH inductors leading out from there. That combination in theory is presenting no net reactance to the filter at 10.7 MHz. If, based on my experiment, you figure you might need to add up to +/-20 pF to create some reactance, you might drop the fixed capacitor to 82 pF and add about 50 pF of variable capacitance. If you set the variable capacitance below 106 pF, you are effectively adding inductive reactance. If you set it above 106 pF, you add net capacitive reactance.

In addition, I found that the theoretical pure resistance of the low pass LC impedance match always seemed to have a residual capacitance of 5-10 pF, so I usually ended up with smaller capacitors than I would have predicted.

You could just make the entire capacitance variable, but that can make the tuning extremely sensitive. You might start with nothing but a large variable capacitor, tune, and measure the capacitance. Then replace it with a smaller variable capacitor plus a fixed capacitor. Or, if you're sure you won't have to do more tuning, replace the variable capacitor with a combination of fixed capacitors.

Sam W.

There is also so info in the files section on the list server

Hi/Lo Pass Filters


The amplifilter is Sam's new addition to the arsenal. It will house 4 crystals with a buffer in the middle. A very intriguing experimental design.

The “Amplifilter” (I'm using Scotty's term, though the gain is not enough to replace the IF amp) has two crystals, then a buffer amp, then two more crystals. Hopefully it will make a very narrow RBW filter (400-800 Hz?) that will be fairly easy to tune. Sam W.

The amplifilter (using Scotty's term ) is Sam's new addition to the arsenal and is a very intriguing experimental design.. Though the gain is not enough to replace the IF amp, it has two crystals, then a buffer amp, then two more crystals. Hopefully it will make a very narrow RBW filter (400-800 Hz?) that will be fairly easy to tune.
Other than resistors/capacitors, it uses
1 OPA842 op amp, SOT-23 package
1 BLM21AG102 ferrite bead
1 L4931CD80 8v LDO regulator
2 Pulse PE-0805CM681JTT 680 nH inductor
The inductors are for 10.7 MHz; for 10.1 MHz use the 750 nH equivalent. All these parts are available at Digi-Key, except you can get the op amp at Mouser. I was intending to find one that Digi-Key carries, but haven't done so. The key thing is that it be available in SOT-23 and can operate at 8V with reasonably flat gain to 11 MHz with gain of 4. This probably takes gain- bandwidth product of at least 200 or 300 MHz. Correction. Mouser is now out of OPA842. OPA695 would be a good substitute. Mouser has them; Digi-Key does not. OPA356, which Digi-Key has, would likely work, and with that one you could use the normal ua7805L regulator, which the PCB will also accommodate. (In fact, with the OPA356, you have to use the 5V regulator; 8V would blow it.).

It can use 4 of the trimmers that Jim mentioned, but I am hoping the tuning will be simple enough that trimmers aren't required. If adjustment is required, I'm hoping it can be done by one-time adding or removal of a parallel cap.

Updated Video Filter Switch Board

(Zip file from Sam W’s site)

There is an updated PCB for the video filter switch incorporating the latest refinements from Sam W.

The video filter board is the same as the one I am using now, with fixes to a couple of traces and pads. Sam W.

Updated W1JGH Power Meter

The power meter auxiliary board is designed for the MiniCircuits ZX47 series power sensors. It simplifies wiring by incorporating the calibrating pot, led and switch in the board itself. Dan Ellis (dhellis@dslextreme.com) has got a discount price from Mini Circuits for the sensor and he is offering to assemble complete kits. Just get your board to him.

High Isolation SPDT

The SPDT is Sam W design for a generic PE4251 filter that may be used as a band or a transmission/ reflection switch. It has very good isolation up to 1GHz (90 dB). The SPDT is a compact high-isolation solid-state switch. Sam has not built this exact one, but a similar switch performed very well. It uses 3 very cheap PE4251s.

VHF Transmission/Reflection Switch

The VHF transmission/ reflection bridge is Sam W implementation of an S-parameter test set that should be good through 300 MHz or so, and should be superb to 100 or 150 MHz. The set combines a bridge and buffer amp with a transmission/reflection switch. It's not really a full-blown S-parameter test set because it cannot reverse the DUT. It should be accurate without the use of 6-term or 12-term calibration, and any mechanism to reverse the DUT screws that up. To use it, the DUT is connected between the DUT port of the bridge and the input of the buffer amp. This lets the DUT see solid 50 ohms on each end regardless of the relay setting. The relay selects the bridge output for reflection and the buffer amp output for transmission. "VHF" is intended to convey that it does not cover the full range to 1 GHz, though it does also cover far below VHF, to 250 kHz or so, depending how big you make the capacitors. It could also do a nice job of displaying return loss from 250 kHz to 100 MHz with just a spectrum analyzer with TG. We will publish the schematics in a couple of days.
Sam wrote:
I have been fascinated by op-amp based bridges, in part because they can perform nearly perfectly below some frequency (somewhere in the 10-50 MHz range) and can perform very well above that. One thing this lets you do is display return loss graphs using just a spectrum analyzer + tracking generator. In fact, you don't even need a tracking generator, if you have a noise generator or plain-old swept-frequency-generator. Or, with the MSA/VNA, scans can be made with reference calibration, much simpler than OSL.
In some cases, it would be handy for the bridge to have a reference impedance different from 50 ohms. 75 ohms is an obvious possibility, but there are also situations where a much higher impedance would be appropriate. For example, you could get more accurate measurements of MCF filters by using higher impedances. Especially below 30 MHz, many circuits might have impedances much higher than 50 ohms. Likewise, there may be circumstances where it would be nice to have a very low impedance bridge.
I have designed a bridge whose reference impedance can be changed simply. Different high impedances could be selected by changing a single resistor. This single resistor approach will work for 50 ohms and higher. Lower impedances require changing a couple of resistors. The idea of changing resistors probably only works if you do so only occasionally, or if you create some simple switching method or allow plug-in resistors. The latter approach would work, but the goal of "perfect performance" probably limits the frequency range to <10 MHz or so.
Sam W.

The "VHF Transmission/Reflection Set" combines a bridge and buffer amp with a transmission/reflection switch. "VHF" is intended to convey that it does not cover the full range to 1 GHz, though it does also cover far below VHF, to 250 kHz or so, depending how big you make the capacitors. It could also do a nice job of displaying return loss from 250 kHz to 100 MHz with just a spectrum analyzer with TG. Sam W.

AADE Accessory Board To Measure SMDs

The AADE accessory board l fits in the terminals of the AADE L/C meter. It may be left permanently in place without interfering with normal operation. By simply placing the SMD device in the proper position and holding it in place with a transparent plastic strip (held down by scotch tape) will let you determine their values. It will allow for parallel and series combinations of components.

Diplexer PCB

Diplexer on Mixer 1 output to simplify switching between 1G and 2G modes. 1G is the normal 0-1 GHz mode. 2G mode is 1-2 GHz. 3G mode is 2-3 GHz and operates with the same hardware setup as 1G mode. (See Sams Site)

Zero response zapper PWB_RFA-1 by Scotty/SamW

(See Scotty’s Site for more info)

Hi All,
I have completed the Web Page for the Buffer Amplifier:
It can be used as the Buffer/Isolator between PLO 2 and Mixer 3. It is
designed to have 0 dB gain, with about -38 dB of reverse isolation.
Scotty pointed out that the "zero response" that actually extends up to 1 MHz or so in VNA modes, can be significantly reduced by adding a buffer amplifier in the path from PLO2 to Mixer 3. This prevents LO3 from leaking back into PLO2. Essentially, this requires 15 dB of attenuation, followed by 18 dB of gain from an ERA-33 or similar MMIC, followed by 3 dB attenuation on the output. I made up the attached layout for such an amp, which is about 1.2" x 0.9" and is designed to be placed "in-line", with its connectors sticking out each end. That way it can be placed in the cabling from PLO2 to Mixer 3. I haven't done the schematic yet, but it is similar to the MMIC amp on the Signal Router board--but that one was a little more complex because it was also designed to provide power for other boards. This one can be built as normal MMIC-plus-inductor (the inductor actually fits on the R6 pad), or it can be an inductorless buffer amp such as the one I use on my TG output.

Hi Jim and All, I would point out that the SLIM-IFA-33 is a specialized amplifier, but it is constructed on a general purpose printed wiring board called "PWB_RFA". You can build the 1024 MHz buffer amplifier on this pwb by replacing L1(L2) with a 68 nH or similar choke. Replace the filter circuit (C26,L27,C28) with an appropriate resistive pi type attenuator. You can add the 15 dB input attenuator (resistive pi type) to the pads at X21, X22 (R22), and X24.  Also, this is a dual amplifier board and you need not populate the U2 side.  I will update the web page for this pwb. Scotty


Yes, the PWB_RFA will work well for the PLO2 buffer amp. The board that is included on Jim's PCB has only one amp stage and is slightly down-sized to 0.9"x1.2". It is designed to have connectors sticking out opposite ends, which might be more convenient if you don't want to actually mount it in the grid, but want to have it "float" in the cabling. But if someone has an extra PWB_RFA, there is no strong reason not to use it.
The other thing this board will accommodate is a 1206 resistor in place of the inductor. This lets you build a broadband amp that extends to fairly low frequencies, so long as you accept lower gain, which is perfect for something such as a TG buffer amplifier. That resistor can effectively become the first leg of an output attenuator, which can also give the output very good return loss. Likewise, an attenuator on the input can provide excellent input return loss. Altogether, it can provide good in/out return loss and reverse isolation, with just a few dB of gain, which is why I refer to it as a buffer amp. The inductor less design does not work for the PLO2 buffer, because the minimum output attenuation is about 6 dB. So for that purpose the inductor would be used, which is fine because there is no need for broadband amplification. For that purpose the PWB_RFA board is just as good. Sam W.

Step Recovery Board

The step recovery board was part of the MSA and there are plans to resuscitate it as a low phase jitter alternative to PLO2. This board is for the experimenter who wants to test the technology. See the
http://www.scottyspectrumanalyzer.com/srd/srd.html page