Assembly of the components is fairly straightforward, but if you don't have experience with SMT parts it could take a little while. What you will need is:



Additionally have a look at my SMT PCB maker project ( if you plan to make your own double sided FR4 PCB. It is how I make the PCBs as well. The EAGLECAD PCB files are in the shared folder on the sidebar under the gallery pics.

The kit-of all-parts will include the drilled and de-burred end plates and the alum. enclosure. Contact me at for availability. As a convenience I can supply whatever component mix you prefer, including just the PCB or the connectors.


The first component to secure is the SMA launcher.

Mount it in the enclosure with just the SMA endplate and the bottom section of the enclosure assembled with two enclosure screws. Tighten the SMA nut, hand tight with the 'legs' horizontal and the centre pin oriented upward. Note the images supplied as a guide.

Slide in the PCB with the copper SMA pads pre-tinned on both sides of the PCB. You may have to file/sand the side edges of the PCB for a snug fit. Once the PCB SMA pads are aligned with the SMA launcher, solder it in place to the launcher. Now remove the bottom of the enclosure from the SMA plate and solder the two remaining ground legs of the launcher. The launcher is a 1.6mm launcher and the PCB is a 0.8mm PCB so you have a little clearance to align the fitment. Fill any gap with solder. remove the SMA nut and the PCB/launcher assembly from the end you can continue installing the SMT components.


There are a couple vias which require a bit of component lead (from the chokes) to ground both sides of the copper planes. These are essential, as is soldering both layers of the rear edge mount, thru hole, choke pads to ensure a low inductance ground plane.


When constructing the LEMO 304 connector end, you might need to widen the expandable gland (4 tines) thru which the cable passes, a bit of pet. jelly helps here as well as it is snug. Note the images supplied for the assembly and the color codes of the wiring. If you make a mistake here, the unit may always show unleveled on the SG 504 indicator. But no harm will come to the head or the SG-504 plugin.

The drain/shield wire needs to fold back on the cable so that the metal gland slides over it, removing some strands of the shield is usually required for the fit to be good.


I supply a BNC pigtail for use with the head, but if you prefer you can assemble a 3' RG58 or RG400 cable directly to the head to remove the need for a BNC barrel and BNC m2m patch cable to connect to the SG-504 BNC-f output. If you're using the pigtail, be careful when stripping the insulation from center core, it can pull right out of the pigtail, so grip it with a needle nose pliers to prevent that.


When assembling SMT parts with a soldering iron, the technique is to heavily tin one PCB pad for each component. Then holding the component in place on the pads, reflow the solder on the tinned pad onto the component. Then remove the heat and let the component set, apply solder to the remaining component pads afterward. Start from the centre of the PCB...with the schottky detectors and work your way outward.

As an easy guide, print the supplied top and bottom layout images, oversized on a sheet of paper to follow the component placements.

*NB* the 49.9Ω resistors are to be mounted with the markings facing down to the PCB to reduce inductance effects. It actually makes a difference at the upper freq. range.


Don't forget to slide the input cables through the aligned end plate holes BEFORE soldering to the PCB or you'll have to redo those connections and feel foolish as well.


Testing the build:

Set the SG 504 plugin to the 6Mhz range, set the amplitude to 4Vpp, and connect the head with the output directly connected to a 50Ω O'scope input or Spectrum analyzer (or even a microwatt power meter via a 10dB pad.) Upon power up the un leveled light should go off. If not you have a build error and need to check the part placement, the lemo wiring and look for solder shorts or wire bits from the RG58 shield floating on the PCB.

Once the un leveled light goes away all should be fine.

My experience is with a 12Ghz sensor micro-wattmeter and a 10dB minicircuits pad I observe about 5.88dB (16dB-10dB with connector losses) at the 4Vpp, level for the 6Mhz ref. signal. This looks like just about 4Vpp on a O'scope 50 ohm input with a smooth sine CW wave.

note that Vpp = 1.414 * 2 * Vrms

Power= SQ(Vrms)/Impedance => 40mW @ 4Vpp into a 50Ω load.

Secure the input cables with hot glue as pictured, close the enclosure, and you're done! Now you can reliably calibrate your VHF/UHF instruments or determine the basic performance of any RF amplifiers etc.


Going Further - More accuracy via custom calibration.

If you have access to a calibrated RF power meter good to 2Ghz and accurate to two decimal places of dB you can tweak your build to possibly achieve an accuracy (<2%) that exceeds the OEM Tek. head. Be aware though that all amplitude testing should take place with the enclosure assembled and not half open.

Note the Schottky detector nearest the RF input pad. It has a pin landing pad with stubs extending off it. A couple stubs are unconnected. To increase the upper end amplitude (800Mhz thru 1050Mhz), you can join an unconnected stub to the landing pad with a wire strand to increase the parasitic capacitance at that node. To decrease that amplitude you can cut the connected stub's copper bridge trace.

By adjusting this you can shift the average amplitude of the upper range to suit the +/- dB deviation that gives the best overall accuracy relative to the 6Mhz amplitude figure of merit. The lower range (245-550Mhz) is hardly affected by this.