Tuesday, October 30, 2018


A simple coax switch...



Sooner or later every ham station will need a way to switch coax connections. There are plenty of commercial switches out there, MFJ offers a decent one that is easy to find. And you can always spend an afternoon collecting the parts and assembling your own. However, both options wind up being a little expensive when all you need is a switch to temporarily evaluate a new radio or some array of antennas.

Wanting another switch, I searched ebay and ran across a 3 position switch made by Workman. It claims 1000W, which I didn’t believe but it was only $12, something I could afford to lose when packing up after a field day. I remember the Workman brand from the 1970's. They mostly had tools and accessories for the TV repairman but also a few reasonably priced items for the hobbyist. I bought three of the switches and they arrived in the predicted 8 days by China post.

As a test the switch, was put in place between my radio and a 20m dipole on position 1 and a 40m dipole on position 2. Sadly, there was a noticeable decrease in signal and without re-tuning, the SWR was also worse. The MFJ switch didn’t exhibit either problem.

So, what's up? It is just a stupid switch. Popped the back cover (two screws) and surprise, it is packed with wires!! WTF! You would expect one wire from each connector to the switch and not much else. It is a cheap switch so you would expect the grounding rely on the screws to the case.

A little closer look and it seems that someone decided that the grounds should also be switched. It makes no sense because they are all tied together anyway! Whoever sent the order to the manufacturer probably wasn’t all too clear on the actual requirements. The good news is that the switch itself is a reasonably good one, and the SO-259 connectors look fine.

Ok fine. Grab a soldering iron and pull out all of the wires other than the four wires from the center of each connector. And even those were un-soldered from the switch.

That's better, now put the four wires back on the switch (on this particular switch I elected to put the input on the side with 3 connectors, your choice). Keep the lengths as short as practical.

Now the grounds. An ohmmeter suggests that the connections are pretty good between the case and the connectors but it looks more accidental than intentional. So, we could use some of that left over wire to tie each connector shell together, or try something better. I happened to have a bit of copper foil so it took a few minutes with a large soldering gun to tack it to each connector and to the steel box after scratching off the paint (good paint BTW). Some bits of copper clad PC board will also work

In this case I had to re-label some the switch positions, but otherwise it looks like the original. A quick test and it now behaves as well as the MFJ version at least under the above conditions. I didn't bother to test it on 2m, but it is probably only useful for HF.

Conclusion: You cant beat this as an option for an occasional use switch. The parts are worth more than the purchase price and it takes barely 20 minutes to do the rework. In spite of the nice statement on the label, this is not the switch to use if you are running more than a couple hundred watts, but it is great for field days and they make welcome gifts for your ham buddies.

QRM-180



QRM-180





Yet another Noise Reduction

 System







 
Background:
Unless you live far out in the countryside you likely have your share of interference from TVs, LED lamps, power lines, computers, solar panels, etc. There are a number of QRM solutions available today, and even a few kits, so, not happy with my noise situation, I have purchased or built any of the versions I could find that looked promising. A common theme (popular since the days of vacuum tube TVs) is to use a second antenna (used to collect a suitable sample of the noise), invert that signal and add it back to the main signal. It sounds simple but like most things, putting it into practice requires some effort.

The good news is that they can be reasonably effective. Don’t expect miracles but one of these can get rid of just enough noise to make them worth having. I own several of the commercial versions, including the MFJ-1026, and ANC-4, as well as the X-phase kit from Russia. They all work well enough, with some subtle differences. The ANC-4 seems the easiest to operate, and is arguably the better made, the MFJ has a useful preamp on the noise line. The X-Phase is pretty basic, but with some additional components is still useful. These units are designed with amateur radio transceivers in mind, but it will work equally well on receiver only applications. Again, they can do a good job canceling a nearby noise source, but not atmospheric noise. There are a lot of reviews on line as well as a bunch of Youtube videos that give some useful information.

The QRM-180 came about because the commercial units actually worked and I couldn’t resist trying to make one myself, so after a lot of iterations on several breadboards, I spent the time to lay out a decent PCB so that I could put a separate one on each of my three transceivers. The board (with ground-plane top and bottom) is much better than my breadboard and managed to fix some of the shortcomings (my subjective opinion) of the commercial units. The circuitry is not all that original, it is a combination of what I deemed to be the best of the existing designs. The board is all through-hole components which makes it much, much easier than surface mount to build by hand and also makes it serviceable in the future.

The case is a compromise between being roomy enough for old, fumble-fingered assembly and not so big as to take up valuable space next to the transceiver. It sits comfortably on top of most external speakers, or on top of the transceiver itself. It is thick aluminum so it is well shielded and heavy enough not to slide about during use.

To minimize confusion, all of the connectors on the rear panel are different. And rather than another SO259 connector, the transceiver is connected with a fixed pigtail/PL259 that should comfortably reach the back of the transceiver.

Power requirement is 12V to 13.8V. The included wall wart will work (it is a switching type and potentially generates noise, but I haven’t been able to detect it) or any other good 12V source. Most transceivers have a 12V output that could be utilized, but you should have at least 400ma available to power the QRM-180. Check your transceiver manual first.

Function:
The circuit is much like the others. A separate noise antenna gathers what is hoped to be a good sample of the offending noise without a lot of your coveted DX signal. That sample is run through a broad band amplifier before being inverted by transformer T1 and fed to a phase shifter. I tried a number of schemes for phase adjustment, and this one is not only the simplest, it introduced less distortion. Two J310 FETs then act to buffer and combine the two basic signals.

The combined signal is then buffered by Q4 and sent to the receiver through a relay. In transmit mode, the relays serve to isolate the device and allow the transceiver to be connected directly to it’s antenna. The relay is large enough to handle 200w, but make sure that you put the QRM-180 between your transceiver and your linear if you use one. It would probably be toast if you try to install it following the amp. The PTT (really just the transmit line not the actual PTT from the mike) input grounds the gate of Q2 which in turn turns on Q4 the power tab transistor. This not only energizes the relays, it supplies power to the rest of the circuitry. The relays energize on PTT low so that the default state is safe (circuitry bypassed) when the power is off.

Like the MFJ-1026 there are fuses and protection diodes on the inputs (although MFJ appears to have defeated some of them, likely to minimize service calls). In spite of some suggestions on the Internet, don’t remove them. They are useful and do not degrade the performance at all (unless they are blown, duh). The pi filters are a little debatable but for the bands where the QRM-180 is useful they are a good thing. I have tried the MFJ-1026 with and without, no difference. If you own the MFJ and want to improve it, add an LNA.

The provided amplifier for the noise antenna is a commercially available pre-amp (LNA) that provides a nice flat, low distortion 30dB of gain that can be adjusted down using RV1. In very rare cases where the noise is very large and the noise antenna is also large, the LNA may not be required and could be bypassed.

The QRM-180 does have transmit-detect circuitry to disable the relays to try and protect things if the PTT is not connected when you transmit (C17 coupling some RF and eventually turning on Q1). Like the other devices with this type of circuitry, it barely works. It is pretty good with AM but anything else is intermittent at best and will quickly blow a fuse. You have to connect the control wire from the transceiver.

Setup:
Start by checking the manual on your transceiver. On any transceiver built in the last 25 years there should be a connection on the back of the transceiver designated for operation of a linear amp. Look for something labeled TX GND, PTT or similar. If you are lucky it is a separate dedicated connector. That is probably what you want to connect to the PTT input of the QRM-180. But check it first: operate the transceiver while monitoring that output with a voltmeter. It may be just relay contacts to ground so if you read no voltage, try measuring continuity. Whatever the case, it must go to ground when in transmit mode. Depending on your transceiver, you may have to obtain a suitable connector to reach this signal.

If your transceiver has a built in tuner, check the output when the tuner operates. The tuner should also pull this output to ground when it transmits. (while it is possible to use the actual PTT line, the same one that comes from the microphone, if you have a built in tuner you will have problems since that line doesn’t generally change when you tune).

At the worst, you may have to install a small relay to provide the low-on-transmit signal. Check the Internet for instructions on how to connect your transceiver to a linear amplifier, the connection is the same.

Operation:
The trick to success is to get just enough of the noise signal to match the level of the noise component on the primary signal path. The kind of noise will dictate the requirements, but generally you will need an antenna and some time to experiment with settings at different times. Don’t expect to reduce atmospheric noise-- you cant. But for a specific noise source you should expect a significant if not complete reduction. Be sure to read the reviews and manuals of the other devices (all available on-line) to get the full picture.

Step one in deciding upon the noise antenna is to see if you can characterize the noise. What time of day is it present, is it near or far, where in the band is it strongest? If you have multiple antennas, is the noise the same on each? Try to tune it in. Start on an easy band, say 20 or 40M. And connect something to the noise antenna (an old TV antenna is great if you have one). Some people swear by a long piece of wire laying on the floor. If that works, the noise is probably in the shack, and should be fixed with some grounding and ferrite beads not a phase inverter.

  • The GAIN knob is just that, you are trying to find a level that just matches the noise so that the result is zeroish. It is not to find the best listening level. Use the radio knobs to do that.
  • The PHASE knobs jointly adjust the phase of the noise signal. You should adjust each for a minimum. Unfortunately they also mess with the level, so you have to go back and forth a bit between all three.
  • The NOISE Attenuator control is rarely used and is on the rear panel. When the LNA is installed it is used to get the noise signal in a range that can be matched by the GAIN control. Without the LNA it will probably just be set to max. It will depend on the size, type, and location of the noise antennas, as well as the strength of the noise source.

Start adjustment by turning the gain down a little on the transceiver, if it has a preamp turn it off (with the gain of the QRM-180 it is unlikely that you will need it). Turn on the QRM-180 and turn it’s gain to about half, then adjust each of the two phase controls for minimum noise. Keep adjusting things until the noise seems to be minimized. If the LNA is installed, start with the Noise Level control set to about 70%.

You have to use your receiver controls as well as the controls on the QRM-180 to get the best results. The signal will likely be different than what you are used to. Try with and without the transceiver’s notch filter, NB, and different width settings. It will take a lot of practice to get things right quickly. I have solar panel noise (from the neighbor’s inverter) that can be killed completely in 3-4 seconds of knob tweaking. Changing frequency within a band should require minimal readjustment.

Try with and without the QRM-180 frequently to be sure that you are helping things. Try to ignore the level change and listen just for the change in the noise component. Use the transceiver volume control to match the levels for comparison. Again, you cant escape atmospheric noise, focus on the specific noise source.

For different bands you may get better results with a different antenna (I don’t understand why but try it). You should run coax from the back of the QRM-180 to the noise antenna, but pretty much anything will work. RG-59, RG-174, or whatever, it doesn’t need or want to be 50ohm. Try a short dipole (I have one in the attic), or rain gutter, or a fence,... you never know what will resonate with your particular noise problem. Going vertical is a good option. Remember you are trying to receive just the offending noise, not the HF signal. If you can set up an antenna with some gain, great, point it at the source. When you have established a decent reduction, note the knob settings. It will be pretty repeatable for the same conditions on the same band. A word of caution, whatever you choose, don’t go with a big noise antenna that is too near to the one you are transmitting on. The noise antenna could easily pick up enough RF to blow the fuse on the noise antenna line.

 PCB assembly


You will need some basic tools: soldering iron with a smallish tip, your favorite solder, sharp flush cutters to trim leads, screwdrivers, and if you are not 13 years old, a good magnifying lens.  A volt-ohmmeter will likely come in handy.  Begin by sorting out the parts so as not to loose anything, get some good light, and a place to work comfortably for a couple of hours.  Complete assembly is probably a two evening job; there are a lot of parts.

It is generally best and easiest to install the smallest parts first, and gradually work up in size.  Put in about half a dozen parts then solder them.  Then trim the leads and put in the next group.  Checking them off of the BOM as you go will save time in the long run.  Don’t worry if you have parts left over, there are a few extra parts in the kit (J310, fuses, some extra diodes, the occasional extra resistor, I don’t count so good).

Caution:  there are THREE significant errors on the PCB… 
1- the collector and emitter of Q4 ore switched, follow the instructions to correctly cross the legs. 
2- the silkscreen for diode D19 is incorrect, it must be installed opposite from what is shown. 
3- C2 is a polarized capacitor but there is no plus sign on the silkscreen. 

4- I forgot that you will want to add R15 to the backside of the board... 

These errors will be corrected on the next layout, but other, similar errors will be inserted at that time.

1) Install the Zener Diode D7 first, this will keep it from being confused with the switching diodes.  As with all the diodes (other than D19 which is marked wrong), watch the polarity and match the band with the silkscreen image.

2) Install the 1N4005 black rectifier diodes.  Put D19 in backward, and install the rest of the diodes.

3) The resistors are next, but don’t install R1!, it is not needed.  And you might want to save R9 for later since it mounts vertically and makes the board a little difficult to work with for now.  Save the potentiometers until later for the same reason.

4) Install the inductors, there are two different values and they can be difficult to tell apart.

5) The capacitors are next, just save C3 and C4 until later.  Only C2,C3 and C4 are polarized, the rest can go in either way.  C2 is inserted with the negative end nearest D6.

6) Now is actually a good time to install the connectors.  You can use some judgment here if you like. There are screw type connectors in the kit that work well but some people would prefer to solder the antenna and transceiver wires directly to the board.  Your choice.  Don’t put a connector in the S1 position, you will solder wires here later for the power switch.  Using the supplied connectors on J2 and J3 is a good idea.

7) Take a break from soldering and wind the transformer.  Wind 10 turns of the enameled wire, and then three turns of the white insulated wire.  The wires can exit the ends or sides of the core, it doesn’t much matter.  Just make sure that when it goes on the board the enameled wires terminate on the C7 side.  After, you will want to trim the wires to about the right length to fit the board (not too short).  The white wire can be stripped with a good pair of strippers.  The enameled wire must be prepped by scraping the ends with a razor blade or sandpaper, then tin the ends to make sure that all of the enamel has been removed.  Double stick tape will hold the core in place on the board.

8) No transistors yet, but all of the larger parts can go in now.   Don’t forget to install R9.  Solder in the controls and the relays.  Note that RV1 is a 1K pot.  The others are 5K.

9) The power jack should be wired with about 6 inches of wire.  Wind 3 or 4 turns of the power wire through the ferrite ring.  The power switch can be connected directly to the board with about 2 inches of wire.  The RCA jack also needs about 2 ½ inches of wire to get to its board connector.

10) OK, Q4, the large power tab transistor….  The board design evolved a bit and somewhere along the line the footprint for this transistor got messed up.  The base pin is correct but the collector and emitter got swapped (sorry about that).  With a little leg twisting, however, the transistor will still fit nicely.  See picture.

11) Install Q3, Q5, and Q6.  You can use in-line sockets if you like but I tend to think that they cause more trouble than they are worth.  Q5 runs warm so install the provided heat-sink.

12) Install Q1, and Q2.   D8, the two color LED can also go in, shortest leg goes into the square pad.  Don’t cut the legs, instead install the LED as high as possible.  It will be bent to fit the front panel later. 

13) Assemble the small PCB by installing two diodes, the fuse, and the connector as shown.  Then attach about 5 inches of coax.  See pix.

14) The 1megohm resistor was originally omitted, but should probably be installed to establish a DC reference for the gate of Q5.  It can be easily tack soldered to the back of the board from one side of RV2 to ground.



You could test the board at this point but it is actually pretty easy to access and remove from the case so I recommend testing after the case is assembled and the board installed.

Case Assembly


1) It is easier to label the back panel before the connectors are installed. You should trim the provided label strip to a length that doesn’t interfere with the mounting screws and put it across the top back.

2) Install the connectors on the back panel nice and tight.  The anodizing on the panel is pretty tough so with a bit a bare copper wire, tie all the ground lugs together as insurance that they are all really grounded.

3) Add the bit of fiberglass shaft to the shaft of RV1.  Use the adhesive lined heat shrink provided.  This will allow this control to be adjusted from the back panel.  If you are lucky, I will have done this for you.

4) Insure that the board controls fit nicely in the front panel but don’t put it on just yet.  With another bit of copper wire, connect the cases of the GAIN, PHASE A, and PHASE B, controls to ground.  Tack the wire to the top of each control and to the small ground plane pads to the extreme left and right.  It doesn't hurt to also ground RV1 in the same manner but you will have to sand the case a bit to get it to take solder.

IMPORTANT-- make sure that all the leads are all trimmed close on the bottom of the board..  long leads will short to the case and cause lots of grief.

5) Slide the board into the one of the housing halves and after fussing with the LED to align it with it’s mounting ring, go ahead and put nuts on the controls and a couple of screws to hold the front panel to the  housing.  You may notice that the front panel is labeled on both sides, DON’T PANIC.  I considered for a short while that it might make sense for the unit to sit vertically.  It didn't.  Just put the horizontal layout to the front.

6) Install the rear panel and hook up everything other than the Noise antenna connections.  With the knobs facing front, the connectors on the rear of the board always have the ground pin to the right, active pin to the left.  Put the rubber feet on the bottom.

7) The transceiver cable goes through a tight grommet, but even so it would be a good idea to put a bit of your favorite glue on the cable to hold it firmly in place.

8) The LNA board simply sits above the main board held in place by a 90 degree SMA coupler to it’s input.  Power is provided by tack soldering two wires on the main board near poor little Q4.  (ground and the collector of Q4)

9) The input to the LNA comes from the rear panel (the BNC connector) through the small red PCB. 

Trouble shooting.
CAUTION: If the transceiver is connected, there is a possibility of high voltage and RF exposure. 
BE CAREFUL!  Turn RF power down on your radio and take care not to transmit.

It is fairly robust circuit so you should not experience any problems if you managed to get all the parts in the right holes and in the proper orientation.  But don’t panic if you do have an issue.  Should something not function, check the schematic for that area and recheck the parts related to that function.  It is surprisingly easy to mis-read a component’s value.  There are a few extra parts included with the kit in case you need them (extra fuses, a J310, etc.)   Start with getting the relays to click when the PTT is grounded.  When that works, you know that the power is right. 

Make sure that the LNA in not in backwards, same for all of the transistors, and the transformer.

If you still have signal problems, try connecting to the receiver but without the antennas.  Use a small probe or bit of wire to touch various points along the signal path.  Just your body as a signal source will introduce plenty of static to locate a fault.   If you can mess with the knobs and get a null in background noise, it is working properly.  You just need to get the best noise antenna and get good at fussing with the knobs.

All else fails, email me and I will try to help.      rob@ag6yj.com