Monthly Archives: July 2020

Bias-Tee Module with DC Switching Logic

2 Replies

The DG8 RXtee Board for 28-432MHz

RXtee board complete

I designed this board because I needed a few compact bias-tee modules with built-in logic for remote switching of preamplifiers, transverters etc. Now that the design is working well, I’ve decided to make the board available in the GM3SEK Odd-Boards series for home constructors.

What’s a Bias Tee?

A “bias tee” is a simple way to feed DC power up the coax to remote devices such as preamplifiers and switching relays. At its very simplest, it consists of a series capacitor to pass the RF, and an RF choke to introduce the DC.

Generic schematic
(To build a high power 1.8-50MHz bias tee, click here .)

  • CAUTION: whenever you build a bias-tee, LABEL THE CONNECTORS! 
    Connecting a bias-tee the wrong way round can cause expensive damage because one of the two RF ports can deliver significant DC voltage and current into whatever equipment is connected.  I have labeled the ports correctly above, and also on the PC board – “RF” and “RF+DC” – but the rest is up to you.

Features

The DG8 RXtee is several steps smarter than the basic bias tee above (although you can also wire it as a simple basic bias tee if that is all you need).

  • Covers all RF frequencies from 28MHz to 432MHz (at least) with low loss and low VSWR (<1.05)
  • Low-power applications only – not rated for handling high RF or DC power ( click here for details of a 1.8-50MHz high power version)
  • DC supply typically 12-15V, currents up to ~1A with short-circuit protection
  • Versatile TX/RX switching of the DC power (separate logic inputs for ground-to-transmit and TTL)
  • Output for an external indicator LED
  • Designed to fit inside a standard 20x20x37mm tinplate box (optional; user provides box and connectors – see parts list below).
  • Supplied as a bare board only, for home constructors with sufficient experience to source the parts and assemble the board.

BOARDS NOW AVAILABLE – order here

Schematic and Notes

DG8-RX-Tee schematicClick here to download as a printable .jpg file

Construction and connector options

You have choices! Get used to the idea of building things as you want them – not as someone else wants to sell them.

    • As noted above, you can leave out the DC switching options if you don’t need them.
    • This description features SMA female connectors soldered onto a screened tinplate box – a good general purpose option.
    • If you wish, you can solder flying leads of RG174 or RG316 coax directly onto the box (or use a connector on one side, a flying lead on the other).
    • If you don’t need screening, you can leave the box out completely and solder connectors and/or flying leads directly to the board.

Performance

All measurements were made on the prototype shown in the photographs,  using the recommended Epcos 2.2uH RF choke.

DG8 RXtee v096

Click here to download as a printable .jpg file

Marker data from the above traces:
                    Freq        Ins loss    Rtn Loss    VSWR
Marker 1   28MHz     0.05dB       33dB          1.05
Marker 2   50MHz     0.05dB       39dB          1.02
Marker 3   144MHz   0.06dB       42dB          1.02
Marker 4   432MHz   0.14dB       33dB          1.04

Parts List

C1, C5, C7: 100pF C0G dielectric, 0805 SMD (various sources)
C2, C4, C6, C9: 1nF (1000pF) C0G dielectric, 0805 SMD  (various sources)
D1: 1N4148, 0805 SMD  (various sources)
FB: Ferrite bead RF filter, 0805 SMD (various sources)
PTC: ‘Resettable Fuse’ rated to carry 1.0 A (various sources)
Q1: NDT2955, P-channel MOSFET, SMD (eg Farnell 9846271 )
Q2: 2N7002, TO-236AB SMD (eg Farnell 3439577 )
R1, R4: 10K, 0805 SMD (various sources)
R2: 220K, 0805 SMD (various sources)
R3: 1K0 0.25W, 1210 SMD (various sources)
RFC recommended: Epcos B78108S1222K000, 2.2uH 1A (eg Farnell 3806231 )

Optional:
RFC1, RFC2, C3 – 0805 pads are provided in case you need a 2-stage filter.
See notes on the Schematic above.
Tinplate box: Schubert type FG1,  20 x 20 x 37mm (un-punched) from G3NYK
External LED.

Construction Notes

  • For a high-resolution view of the completed DG8 RXtee, click here or click on the lead photo.
  • For most applications you can use a single wire-ended RF choke, as shown in the prototype. If you really do need two-stage RF filtering, 0805 SMD pads are provided for RFC1, C3 and RFC2.
  • If you are intending to mount the board in a tinplate box, see the detailed notes below.  Your order of assembly must be as follows:
    • file the board to be a snug fit inside the box with no gaps (board is supplied slightly over-width to allow this)
    • check that the RF striplines end about 0.5-1.0mm away from the walls of the box.
    • assemble the board and test the DC switching first (the underside of the board will be inaccessible when soldered into the box)
    • install the RF choke last of all, after the board has been soldered inside the box.
  • Here are close-ups of the top and bottom of the board with parts assembled ready for DC testing.  (Later batches of the boards will have printed part numbers as shown on the right.)

Topside
C4 is soldered on top of C5; C6 is soldered on top of C7.

Underside

Installation in a tinplate box

Many small RF construction projects involve a PC board inside a tinplate shielding box, with accurately drilled holes for connectors. The detailed instructions below will be a useful checklist for other projects too.

1. Mark the box very carefully for drilling a pair of holes on opposite sides, as shown below (left, dimensions in mm)).

Drilling

2. Accurately centre-punch the correct locations and then accurately drill a 1mm pilot hole in each side. Tip: this thin sheet steel isn’t easy to drill accurately. For the best possible start, use a short 1 x 3.2mm centre drill (eg Farnell 378756 ) to prevent the drill wandering off-centre.Centre drill3. Check that the two holes are exactly opposite each other, square and level across the box, by feeding a 1.0mm drill through both holes. If all is well, use the same centre-drill to open out the holes to 3.2mm.
4. If necessary, line up the two holes using a round needle file. To avoid this horrible job, be careful and accurate when marking and drilling the pilot holes!
Tip: when drilling thin sheet metal, always clamp it down onto a piece of hardwood. (For this particular job, a scrap piece of 19mm floorboard fits nicely between the two sides of the box.)
5. If you are installing SMA connectors as shown above, open out the holes further to 4.0mm – again being careful not to go off-centre.
If you are installing flying leads of RG174 or RG316 coax, leave the holes at 3.2mm diameter.
6. Carefully de-burr the holes, inside and out.

7. Position the board accurately inside the box, and solder it into place. This is much easier If you are using SMA connectors that have a flat rear face. More detailed steps:

    1. Trim the PTFE insulation off the rear face and trim the remaining centre pin down to about 2.5mm. You should now find that the connector is self-locating inside its 4.0mm hole. Solder each  connector into place, making sure the solder covers the entire rear of the flange (not easy to do neatly – see above).
    2. With the connectors in place, feed the pre-assembled and tested PC board into the box, underneath the connectors. Position the board straight and level within the box, with the two connector pins accurately centred along the RF striplines on the PC board.
      Take your time to get this right.
      Check that the ends of the RF striplines are not touching the side wall of the box! If in doubt, remove the board and use a sharp knife to chamfer the ends of the striplines (but don’t remove more than is really necessary).
    3. Tack-solder the connector pins into place on both striplines. Check again that the board is level, and then tack-solder the far end of the board to the side wall of the box to hold it.
    4. Now seam-solder the PCB ground-plane to the wall of the box, at each side of each connector.
    5. Underneath the board, also seam-solder the accessible ends of the ground-plane as well.8. Install the RF choke as described above.

9. Find the other part of the box (it must be on the bench somewhere, right?) and trim one end to 8.0mm as shown above. This will leave room to connect wires to the row of pads on the board.
10. Last of all, assemble the whole box and test for correct operation.

As I said above, these “European tinplate box” techniques
will prove useful for a wide variety of small RF projects.

Page, circuit design and PC board,
all © 2020 IFWtech (Ian White, GM3SEK)

The Private Life of Coaxial Cable

3 Replies

This is a common theme from several of my talks over the years about Ferrite Chokes and Baluns. By expanding the topic into a blog format, I hope to be able to explain it in more detail, and add some FAQs as time goes on.

This well-known image originated with Walt Maxwell W2DU, the author of many articles about transmission lines. It brings out several important points.

  • The inside of a coaxial cable truly is “private” – completely screened from the outside world. This is one of the rare statements that can be handed to absolute beginners as a simple fact, and still remains valid all the way through to graduate-level EM physics.
  • The private interior of coax is due to the skin effect which means that, at RF, current flows only on the surfaces of conductors.

    There are many confusing illustrations in textbooks that make some people imagine the skin effect only applies to electrical conductors of some particular shape and size, or in some particular type of circuit. That isn’t true! The best and most generalized derivation of the skin effect I have been able to find is reproduced here. All the high-level EM physics and maths leads to a simple and very powerful conclusion:
    Whenever and wherever an RF current is flowing (regardless of the circuit configuration, regardless of the reason) the skin effect will always be present.
    RF current will always be forced to flow very close to the outer surface of the conductor. There is no net current flow into or through the thickness of the material. 
  • Because we know the skin effect will always be present, we can use that as a starting point when tracing RF current pathways on complex shapes such as coaxial cables and shielded loops. Suddenly, the whole topic of shielding (and what you need to do to make it work correctly) will begin to make sense!

    There is one condition, though: you have to hold onto that idea consistently. Many of the myths in RF engineering arise because people aren’t being logically consistent. They imagine they can switch these basic ideas on-and-off to suit their own preconceived notions… and physics just ain’t like that.  
  • In coaxial cables, the skin effect in coaxial cables means that the outside surface of the shield is a completely separate conductor from the inside surface.
  • The outside surface of the shield is part of the outside world. This is where the so-called “common-mode” currents flow.

Private Life of Coax

Keeping Count of Currents

(This section contains detail to prove the overall conclusions. If you’re reading this for the first time, skip down to the Summary below. Then come back when you’re ready.)

There are five different current labels in the diagram above, so let’s simplify this:

  • On the antenna, I4 and I5 are the currents at the two terminals of the feedpoint. Textbooks often show these as equal – but when we connect a feedline, I4 and I5 may not be equal any more.
  • Inside the coax, I1 is the current flowing on the surface of the centre conductor. I2 is the current flowing on the inside surface of the coax shield. The centre conductor is entirely surrounded by the shield, leading to very strong electromagnetic coupling between these two conductors. The consequence is that I1 and I2 are exactly equal and opposite. (Strictly speaking this is an approximation, as distinct from a true ‘law of physics’, but in real-life coax I1 and I2 will be equal within a small fraction of a percent.) This means we can write: I2 = I1 (no minus sign because the arrows point in pposite directions).
  • The junction point X is where the inside and outside surfaces of the shield meet together. Kirchhoff’s Current Law says that all current arriving at and departing from this point must sum to zero. Noting the directions of the arrows for I2, I3 and I5, this means that: I2 (arriving) = I3 + I5 (both departing).
    We can rearrange this equation to say: I3 = I2 – I5.
  • Because we already know that I2 = I1, and I4 = I1, it follows that: I3 = I4 – I5.
  • So, any difference between the two currents I4 and I5 at the antenna feedpoint will be routed onto the outside surface of the coax shield to become the common-mode current I3.
  • Those are a lot of observations to keep in mind, but they all lead to one simple conclusion.

Summary

  • Because the currents I4 and I5 in the antenna are never quite equal, they can be mathematically resolved into a differential-mode component and a separate common-mode component.
  • However, a coaxial feedline will physically do the same thing! At RF, where the skin effect creates a third conductor on the outside of the shield, the pure differential-mode component of the incoming two-wire connection will automatically be routed into the private, shielded interior of the coax to become I1 and I2. Meanwhile the common-mode component is routed separately onto the third conductor, the outside of the shield, to become I3.
  • If you can block the common-mode current I3 by using a common-mode choke, you will force the antenna feedpoint currents to become much more equally balanced.

And that’s the private life of coax.