Clifton Laboratories 7236 Clifton Road  Clifton VA 20124 tel: (703) 830 0368 fax: (703) 830 0711

E-mail: Jack.Smith@cliftonlaboratories.com
 

This page documents changes to the Z90's Assembly Instructions that may not appear in your printed manual, or in the printed Errata that accompanies your manual. It also serves a channel of communications to document hints and kinks that builders may run across.

You may read the the most recent Corrections and Additions document by clicking here or via the Documents page.

I will list these by date, most recent first.
 

Troubleshooting the Mixer, Crystal Filter and Log Amp Stages. Since the normal operating mode sweeps the Z90/91's local oscillator, signal tracing through the mixer, crystal filter and log amp stages can be difficult.

In working on a Z91 experiencing a problem whilst the 1 KHz filter was selected, I set the Z91 to Signal Generator Mode, commanding it to 10 MHz. This converts the Z91 into a fixed frequency receiver, with an input frequency of either 2 MHz or 18 MHz (The Z90's IF is 8 MHz, so the sum and difference input frequencies are 2 and 18 MHz.) When in signal generator mode, the rest of the Z90 continues to operate, except that the signal level is not decoded and emitted over the serial data line. The 1 KHz bandpass filter is selected in signal generator mode.

To signal trace,connect a signal generator to the Z90/91's input BNC connection and adjust the signal generator to the input frequency, 18 MHz in the example and observe the signal level throughout the Z90/91 with an oscilloscope or spectrum analyzer, should you have one. (Of course, if you use a spectrum analyzer for troubleshooting, it is necessary to use a series resistor of 1 Kohm or so at the probe end  to avoid loading down the Z90/91's stages.) Set the signal generator for about -40 dBm output (about 2 mV if your signal generator is not calibrated in dBm.)

When signal tracing into the crystal filter stage, it's also critical that the signal generator be adjusted to place it within the filter passband. To make this adjustment, place the oscilloscope probe at the junction of L303 and X305. Slowly adjust the signal generator frequency until the resulting signal peaks. Assuming you have calibrated the Z90's time base and that your signal generator is also calibrated, this frequency will be within 2 or 3 KHz of 18.000 MHz. Then, check the signal level at the junction point of each crystal, i.e., between X305 and X306, between X306 and X307, etc. Because the scope probe will slightly load each stage with stray capacitance, it may be necessary to touch up the signal generator frequency setting to provide for a peak output as you go through the filter from section to section.

If you need to check the 200 Hz bandpass filter, temporarily ground pin 6 of either K201 or K401. This will pull in both relays and switch the 200 Hz bandpass filter into operation. Of course, be careful and double check that you are grounding the correct pin.

There's nothing magic about my selection of 10 MHz as the fixed frequency command for the Z90. Use any value that will put the input frequency within range of your signal generator. If all you have is an audio signal generator that can, say, reach 50 KHz, set the fixed frequency to 8050 KHz.

I found normal signal levels in the malfunctioning Z91 until I reached the junction of X307 and X308, where the signal level went to nearly zero—C318's hot lead had contacted the ground foil and was shorting the signal to ground. (This particular Z91 was a prototype and was built on a printed circuit board without solder mask, so the ground plane is fully exposed.) Reforming C318's lead slightly solved the problem.

Modifying the Z90 Display to Remove Stray Character Glitches. Some Z90s have experienced stray character glitches. This usually manifests itself by odd characters flashing on the screen. In some cases, it results in the display getting out of synchronization and reverting to a white horizontal line. Two customers have experienced this problem, which was resolved by exchanging their display modules with replacements. The problem is associated with particular displays, as the trouble would follow the display even when installed on as many as three different Z90s.

The problem results from connecting the display's Frame Ground pin to the Z90's chassis ground. If your Z90 display is working normally, there is no reason to make the modification.

The fix is to clip pins 18, 19 and 20 from the 20-pin header soldered to the display panel, as illustrated in the photographs below. This change isolates (floats) the display's Frame Ground.
 

Mounting the Z10010 Bandpass Filter.  Mike, W2PY, has provided a detailed description and dimensioned sketch, of how he modified the Z90's back panel to mount the Z10010 bandpass filter. Here's how Mike did it:
 

Jack:

 I have enclosed a photo that illustrates how I made the holes in my Z90 rear panel to accommodate the filter.  The photo shows my rough drawing (including my notes) and the tools I used to drill the panel.

 I used a 2x6 board with several nails partially driven into it as a “holding device.”  I covered the rear panel with wide masking tape to avoid scratches.  I carefully measured and drew the design on the inside of the rear panel in pencil.  Then, I center punched each of the 4 holes, several times, to make deep starting guides.  Contrary to my pencil notes, I actually drilled two 1/2" initial holes using the Black & Decker "Bullet" drill shown.  I like these drills because they have a small starting point and they cut a nice clean hole in almost anything, including relatively thin materials.  Then, I enlarged the holes to 5/8" using the step drill shown in the photo - you can see my 'stop' marks on the step drill.  Finally, I enlarged the final holes slightly from the step drill size of 5/8" to a final size of 21/32" using the de-burring tool shown in the photo, although they were pretty close at 5/8” and really needed little additional size.  I drilled the two smaller holes using the correct sized drill bit.

 All of the drilling was done using a Makita battery-powered hand-held drill.

 I had to remove the orange anti-disassembly material from one of the corner screws so that I could get the filter to mount flush to the rear panel.  Although I could have drilled 4 mounting holes, I suppose, 2 seemed to be sufficient.  I used the original filter screws to mount the filter on the rear panel – they were just long enough to work properly. 

I do have a drill press, but it was behind several motorcycles and other junk at the time, so I just went ahead and used the hand tools as described.  When I drill the new replacement panels you sent (which are sooooooo nice) to accommodate the filter, I plan to use my drill press to ensure a "prefect" result, although as you have seen, from my earlier photos, hand tools will produce a nice job if used carefully. 

 -73-

-mike-

 W2PY

Poorly Etched PCB.  I received a report that a Z90 printed circuit board had poorly etched pads associated with the 1 watt, 100 ohm surface-mount resistors (the bias resistors for the Gali-74 amplifiers). The pads were not fully separated from the ground plane. Although this problem will normally be detected by the stage resistance and voltage checks, it's always better to find a problem before parts are installed than discover them afterward. Hence, if you have yet to assemble your Z90, it may be worth checking these pads for an etching problem. A magnifying glass may be useful, or you may make ohmmeter checks to verify that isolated pads are not tied to ground by poor etching.

This is the first PCB production problem report I've received, so I believe it is an isolated incidence. But, boards are normally etched in large panels and the sheared into their final size. This means that there may be another board or two with etching problems.
 

Installing the Z10010 Bandpass Filter Inside the Z90 Enclosure. Mike, W2PY, provides photographs showing how he mounted the bandpass filter inside his Z90 enclosure, following Doug, N6TQS' concept.
 

Z10000-K2 Buffer Installation. Glenn, VK4TZL, provides a pair of photographs showing his Z10000-K2 board installed in his K2.

Extra Signal on Display. Matt, KC0UKK, provides a screen capture image from his newly constructed Z90 and asks a question.
 

The good news is that when the antenna is connected, the background noise level obscures the BFO leakage to the point where it is difficult to see.
 

Crystal Grounding. Matt, KC0UKK, offers a suggestion for builders using the revised crystal grounding technique Bob, K7HBG, developed:

I followed Bob's (K7HBG) lead and used #26 buss wire to ground the crystals to the via (except X304) located under each crystal. Instead of soldering to the crystal after it was installed, though, I soldered the wires to the crystal on my bench where it was easy to work and then installed the '3 lead' crystals into the board. I checked the resistance of the cans to ground and I was done.

Z10010 Bandpass Filter. Doug, N6TQS, offers a suggestion on mounting the Z10010 4915 KHz filter inside your Z90:

Z90 Stray Characters on Display. Friday, I received a trouble report about stray characters being seen on the Z90 display. The characters appear at random and remain in place until the Z90 is rebooted. (By the way, it isn't necessary to cycle the power switch to reboot a Z90--you can also do it by holding down the top and bottom soft key switches for approximately one second.)

I missed the "install the heat sink" sentence, so after I put it in 
the box, I had to take the board back out.  In the process of that, I
broke off one of the mounting sockets for the HV inverter.  So, I
fixed that, and built up a good solder fillet around those sockets.
Doing that, while keeping them aligned, was a hassle.

Z10000-K2 & Z10000-U Buffer Amplifier Coaxial Cable Connection. Doug, N6TQS, reports on coaxial connections at the Z10000 buffer amplifier:

And I'll do the "wrap the wire on the coax shield" method.  I used that on one of the coax runs on the buffer board, too.  It's not quite as clean as a really good quality coax end treatment using either the "pull though the shield" or "comb out the shield", but it's MUCH easier, and almost as good.

Z90 & Z91-Bending the Tinplate Shield. Doug, N6TQS, has a useful recommendation:

I'm using a combination square to make the bends in the tin.  It's working very well, to make sure I get square and sharp bends.

Z90 & Z91; Measuring Current and Voltage During Checkout. I recently received a question about unusual voltage measurements during the Stage 4 current and voltage checkout. The builder is using a laboratory-grade power supply with a Fluke 77 digital multimeter on the 300mA range to monitor current consumption and a Fluke 179 digital multimeter to measure voltages. He reported that the DC voltage into the Z90 dropped from 13.5V to 12.3V when powered up after completing Stage 4. All the internal measurements are within tolerance.

The problem this builder is seeing is that the Fluke 77 meter used for current measuring has a non-zero effective resistance. Digging into the Fluke 77's instruction manual reveals that on the 300mA range, the meter has a "burden voltage" of 6mV/mA. This seems rather innocuous, but in fact it means that on the 300mA range the Fluke 77 appears to be a 6 ohm resistance, plus whatever resistance is in the meter's over-current fuse and connecting leads.

From Ohm's law, R=E/I. The "burden voltage" is actually a roundabout way of defining the meter's equivalent series resistance in current mode; R = 0.006V/0.001A = 6 ohms. Hence, during Stage 4 checkout, if the Z90 draws 200mA, there will be a 1.2V drop across the multimeter, as illustrated in the following figure. (0.2A * 6 ohms = 1.2V)

• Adjust the power supply to provide 13.5V to the Z90 while under load. This is not an optimum solution, as it requires you to adjust the voltage as the current changes.
• Use a power supply with remote voltage sensing and connect the sense leads after the multimeter's output. Good solution, but most power supplies do not have remote sensing.
• Use the power supply's internal metering, if available. This is my normal approach; we are not trying to measure the current down to the nearest milliampere. The analog meters in my various power supplies are adequate for the type of checks necessary whilst building a Z90. The drop across the internal current sense resistors are compensated by the power supply.
• Use a higher current scale on the multimeter. For example, the Fluke 77's 10A range has a burden voltage of 50mV/A, which is 0.05V/1A = 0.05 ohms. When the meter's internal fuse and test lead resistance is added, the net equivalent resistance is still below 0.5 ohms. During the last stages of Z90 construction, when the current draw is about 1A, the drop across the multimeter will be less than 0.5V. Although you will not have nearly the resolution and accuracy when on the 10A scale, it will still be more than adequate for the types of measurements required during the Z90 build.

Z90 & Z91 Heat Sinks. Bob, K7HBG, suggests that if you have it on hand, apply a small amount of thermal compound (a/k/a "heat sink goop") to the 7805 regulators before installing the heat sinks. Applying heat sink compound is not necessary, as the regulators have ample margin without it, but if you have silicon grease or other thermal compound on hand, by all means apply it to the 7805 regulators.
 

The buffer board went together very nicely, and I'm sending along a photo of it in my "full-house" K2.  I'd have made the output cable a little shorter, but wanted to give myself length for another try if I messed up.  But that small coax is pretty easy to work with.

Minor gotcha, mostly my fault-  There are a couple of places where there is a square of pads, and I caught myself about to install a component at right angles to the proper direction.

The K2 installation is very clean-  Note that if the hole in the buffer board were closer to the edge, it could have been mounted using the screw hole next to C220 on the edge of the K2.  I may do that on the other K2.

BTW, my K2 is fuller then most- I originally had the EA3BLQ digital interface installed (and I may leave that in the other K2), and now have the W3FPR fixed level audio output mod installed.  I'm primarily a digital mode op.

Doug notes that it's easy to install a surface mount part at 90 degrees to its proper orientation in a couple places. The top PCB is silk screened but even so, it's possible to make that error. I know that I've come close to doing it a couple times. You may wish to refer to the photograph in the Z10000 manual to avoid problems.

The photo below shows Doug's Z10000-K2 installation. He followed the mounting hole location suggested by Bob, K7HBG, in the upper corner of the K2's rear panel. A dimensioned drawing provided by Bob is in the Z10000 manual.
 

Z90 & Z91 Stage 6 Crystal Filter Redux. This morning, I ran a series of measurements on a late prototype Z90, with the same PCB layout as the production board, except that it does not have solder masking and silk screening, thereby simplifying access to the top and bottom ground planes.

I ran a series of tests with three crystal grounding configurations:

• Top grounding with a wire bus, as described in the current Z90 Assembly Manual
• Floating crystals--no ground connection at all
• Bottom grounding, similar to Bob's recommendations.

My data shows essentially zero difference in filter flank rejection under any of the three grounding configurations for either the 200 Hz or the 1000 Hz filter. A copy of my measurements is available by clicking here or via the Documents page. Since grounding and flank rejection is highly sensitive to the physical PCB layout, it is not safe to assume that the Z90 data can be applied to other equipment, such as the K2.

In a spectrum analyzer such as the Z90, leakage shows up as baseband lifting--the low level of signal that makes it by the filter is shown as increased signal levels on the filter flanks. The visual effect is much like the baseband is lifted above its normal resting place. With the Z90, baseband lifting is usually observed with signal levels around -40 dBm, as the Z90's typical dynamic range is 55-60 dB.  The measured data shows no perceptible change in baseband lifting under any of three grounding arrangements.

My conclusions are:

1. Bob's recommended filter grounding approach is safe to use, but does not alter the Z90's performance,  for the better or for the worse.
2. If you decide to follow Bob's grounding approach, I highly recommend installing the crystals first, before the other Stage 6 components. In addition, you may find it beneficial to identify where on each crystal you intend to make the ground connection, and tin the crystal with a small solder dot before installing it. Lightly fluxing the solder dot location with a liquid solder flux, such as contained in Kester's #951 Flux Pen is recommended.

The Z90 Control software has an automatic baseband clamping option that may be employed to keep the visual appearance of the scan constant even when signals strong enough to cause baseband lifting are present. However, this feature applies only to the computer software display, and does not alter the trace as seen on the Z90's LCD.

Z90 & Z91 Stage 6 Crystal Filter Revisited. Bob, K7HBG, has implemented a variant of the crystal grounding technique discussed on 02 November 2006.

Here's how he describes it:

I did my crystal grounding in a bit of a compromise between flat on board and slightly elevated. I soldered a grounding wire to the via under each crystal, bent it 90 degrees and installed the crystal over the wire. I reformed the wire up the base of the crystal to it's side. I then cut the grounding wire to leave about 1/4" (6mm) from the crystal base up it's side and soldered it. [JRS Note: Vias are the small unused pads that tie the upper and lower ground planes together. One is located at the center of each crystal outline.]

Bob has provided a photo showing his crystal installation.
 

You also can see that Bob has installed the crystals with the identifiers all facing in the same direction, as well as installing all resistors and inductors so the color codes run in the same direction, east-west or north-south These are small things, but add a definite touch of professionalism to his work.

I believe that it would be easier to implement Bob's method if the crystals were installed first, before the other Stage 6 components. My rationale in installing the crystals at the end of Stage 6 is that it is easier to install the smaller capacitors and inductors with the crystal pads unused. However, if Bob's alternative grounding method is used, access to the crystals for ground wire soldering will be much improved if the other Stage 6 components are not in the way.

Here's how Bob describes and illustrates the coax preparation.
 

Z90 & Z91 Stage 6 Crystal Filter. I have previously recommended mounting the eight 8 MHz crystals X301…X308 approximately 1/16” (1.5mm) above the PCB surface. After discussion with Bob, K7HBG, I have decided that the builder should feel free to mount X301…X308 flush with the PCB, as the solder mask provides adequate insulation of traces that might come in contact with the crystal holder.

The method of grounding the crystal holders illustrated in the manual involves running separate horizontal ground busses for the 200 Hz and 1000 Hz filters and connecting them via a vertical connection to ground pads. Bob suggests that builders might wish to consider following an Elecraft K2 modification to its crystal filters, involving soldering the base of the crystal holder to the PCB ground plane instead of a wire buss arrangement. This will require removing the solder masking near each crystal to expose the ground plane for a large enough area to permit a short solder connection to be applied. If you plan to use this grounding approach, you might also reverse the installation order and install the crystals first, before the other Stage 6 components, so as to provide better access to the PCB and the crystal bases.

Although I agree with the potential merit in Bob’s suggestion, I will leave it up to individual builders to decide whether they wish to make this modification as I have not tested it.

I am including pictures of the front bail assembly, in case there are any questions about how it goes together. These things can be a source of confusion for the un-initiated. The first one I ever installed on my Elecraft K2 had me somewhat confused.

One thing...these screws go in hard, so be sure you have them all the way in the bottom cabinet. A larger than normal Phillips screwdriver will aid in getting the screws seated all the way. They will protrude through the bottom of the cabinet about 3 or 4 threads or so.