Ham To Ham #20 - May 1997

73's Ham To Ham column

c/o Dave Miller, NZ9E

7462 Lawler Avenue

Niles, IL 60714-3108

e-mail: dmiller14@juno.com

Remember, I'm always looking for good ideas, tips, suggestions and shortcuts to keep the Ham To Ham column lively...please send anything that you might think would be of interest to other readers to the address above. I'll let you know as soon as possible whether I feel it can be used or not. I try to never leave folks "hanging", wondering if their contribution was accepted or not (as you might have experienced with other publications in the past). Anything in the area of ham radio, common electronic test gear or computers is fair game. This month's edition contains an interesting variety of such examples.

Chill out!

Most of us are now using computers in our day-to-day ham radio operations, be it on packet, RTTY, pactor, amtor, SSTV, FAX or even CW! Computers have become inextricably linked to our ham shacks, as Uncle Wayne predicted they would years and years ago when 73 pioneered the amateur radio magazines in force-feeding us computer related articles (I'm glad he did)! But the speed and processing power of today's computers also places an extra burden on the CPU chip(s), and those little microprocessor blocks can generate quite a bit of heat while they're operating. Temperature problems are becoming more and more prevalent with the gaining popularity of high-speed motherboard designs and faster CPU chips. Generally, the faster a system runs, and the more work that the computer is expected to do, the more internal heat it will generate. If the power supply fan stops or isn't big enough, you can almost be assured that the heat will not be dissipated fast enough.

Here are some factors to keep in mind...a typical computer system today runs between 110 and 130 degrees Fahrenheit internally. While the CPU and some the system chips can be rated up to 140 to 160 degrees Fahrenheit, the overall system generally isn't. Chips will often behave erratically above 130 degrees Fahrenheit, and the hard disk itself (inside of the disk drive) will expand more and more as the computer's internal case temperature rises. Eventually the data can become unreadable, or data written to the drive under hot conditions, may not be readable later on...when the system is cooler. Some drives may even fail-safe to OFF at somewhere around 140 degrees Fahrenheit.

In general, no computer system should ever be expected run reliably over 140 degrees

Fahrenheit, and to be on the safe side, the internal case temperature should always be kept below 130 degrees Fahrenheit. Typical symptoms of system overheating may include erratic operation, random lock-ups or the appearance of error messages without apparent rhyme or reason. Failure to restart properly until the system has been turned off for a few minutes (or longer) can be a pretty decisive evidence that internal overheating is the culprit.

Quick fixes include running your computer at a slower speed ("normal" instead of "turbo" for instance), making sure that the power supply's fan is in good working order, clean and is not blocked by any obstructions (internal or external). Trying to keep an air-flow buffer zone around the main desktop case (or tower), and avoiding "burying" a tower system in an out-of-the-way spot that may not provide for sufficient air circulation.

You might also consider installing an extra fan(s) to move more air, but be careful that you don't worsen the situation by blowing heat-filled air over an area that's already overheated. In addition to moving the system to a cooler (or at least more "open") environment, look into installing a heatsink or heatsink-plus-fan on the CPU itself. These add-ons are available from most computer parts dealers.

Here's a true-life case-in-point. I recently upgraded my earlier 486SX-33 machines to a new high-speed processor (a 586-133) and was surprised to see how hot the new processor actually did get...160 degrees (F) when it was just loafing along! The upgrade hardware didn't really lend itself to installing one of the ready-to-go processor fans available (the ones that simply clip over the microprocessor chip), so I took a slightly different route. I had a very small 12 volt DC fan (about 2"x2"x1" deep) in my spare parts cache (with the cost of spare parts, it's not a "junk box" anymore, HI), so I decided to assign it to the job of cooling my new processor upgrade. The standard computer internal power supply will easily accommodate the small amount of current drawn by the little fan and 12 volts is available by tapping into the yellow (+) and black (-) leads that go to the usual 4-pin connectors of the average computer power supply.

There was an open area just to the side of the processor's ZIF socket, so the careful application of a dab of silicone bathtub sealant seemed just right to quickly and easily "mount" the tiny fan...so that a steady stream of air would flow across the upgrade's top heatsink. That easily accomplished mod lowered the heatsink's temperature from 160 degrees down to 90 degrees Fahrenheit! A 70 degree difference...more than I had hoped for! I used a digital contact thermometer (calibrated to 1/10 of a degree Fahrenheit) to measure the before and after temps right on the surface of the processor's heatsink. That alone convinced me of the value in moving air across the top of these chips!

I also made sure that I didn't end up blowing the processor's heat over some other sensitive component. In my case, the heated air ended up against the side wall of the computer's outer metal case itself, so that it could be dissipated to the outside for the most part. If the inside surface of that outer case is painted with some flat black metal paint, all the better. Flat black paint will help to absorb the heat and distribute it to the steel case for dissipation (a silver or light colored interior surface only serves to reflect some of the radiant heat back to the inside of the computer).

To help you keep track of processor and/or internal case temperatures, a relatively inexpensive indoor/outdoor LCD thermometer (such as the 63-1009 or 63-1019 sold by Radio Shack (reg. trade mark)) can be mounted on the outside of the computer's case, and the remote sensor placed strategically inside the case for continuous monitoring of the processor, power supply or other heat prone area. If the system begins to act strangely or you notice that the internal case temperature is running above what it normally does, you can take appropriate action before disaster strikes. Sixteen bucks for a thermometer is certainly a small fraction of the cost of a new processor or hard drive! If anyone's come up with an automatic temperature alarm circuit that they've found to be effective for this purpose, why not send it in and I'll share it with the rest of the readership? Something that would display the temperature, sound an alarm at a preset maximum and/or safely shut-down the system (for unattended packet BBS or node systems) would be nice to have.

de Dave, NZ9E

Short stop

Here's a great test bench tool from Jim Wood at Inovonics of Santa Cruz, CA: "Figure 1 shows a relatively easily built multivibrator 'beeper' circuit that's capable of tracking down shorts on PC boards, in cables or in other circuits that might otherwise defy location by normal means. It's possible to resolve resistance right down to the thousandths-of-an-ohm level by using the small changes in the pitch of the tone emitted from the 'Short Stop's' built-in speaker (the human ear is very sensitive to even slight changes in pitch at the 1-kHz operating frequency of the unit).

The 'Short Stop' is basically a classical free-running multivibrator, built around a pair of the commonly available 741 op-amps...or a single dual op-amp...but with a few important additions! Transistors Q1 and Q2 deliver a plus and minus 10 volt square wave to resistor R4 and are capable of supplying 100 mA when the probe tips are shorted. Resistor R5 assures that the open circuit voltage never exceeds 0.1 volt. When the probes are open circuited, the gain of U2 is equal to the R4/R5 divider loss, making the two op-amp outputs identical. The tone drops to a low 'growl', effectively silencing the speaker when the probes are in the 'open' state. When the probe tips are a perfect short however, U2's inputs are effectively shorted and the speaker will produce a 1-kHz tone. Anything between these two extremes, even a very tiny resistance, will produce a different sounding tone, telling you immediately if you're 'headed in the right direction' toward the short. Deceptively simple, but very effective.

Construction can be done on a home-brewed PC board or by using the time-honored perf-board and wire jumper technique; circuit layout is basically non-critical as long as reasonably good close-wiring procedures are followed. The separate plus and minus 12 volts DC can be obtained from a dual +/- test bench power supply, or two 12 volt "electronic cigarette lighter" batteries (Radio Shack # 23-154) can be used to provide portability. There is one other precaution that must be observed. Notice that the schematic in Figure 1 shows two (2) wires going to each probe...it's important that you follow this recommendation. U2's differential input must have its own separate path to the probe tips in order to eliminate test-lead-resistance from the measurement. The miniature 'zip-cord' style of speaker-wire that's readily available makes a good 2-conductor test lead fulfilling this requirement. Along the same line of reasoning, the test probe tips themselves must make very low-resistance contact with the short-circuit being traced; I've found that a pair of H.H. Smith #317 probe-tips are idea...their tips are needle-sharp for piercing a wire's insulation or for digging below the flux or oxide barrier on a PB board's trace...and they're easily replaceable as well. You can experiment with different probe tip ideas if you like, just make sure that you end up with ones having the very least contact-resistance possible.

I hope you'll give this circuit a try, it can save you untold amounts of troubleshooting frustration the next time you need to chase-down an elusive short circuit."

Moderator's notes: I'm going to build this one! Jim's circuit understandably was voted the "Best of Issue" in Electronic Design Magazine for April 3, 1995. Finding unintended solder bridges, shorted parts and other near-zero-ohm faults (without having to disconnect every soldered-in part) should now be possible using Jim's circuit. Thanks for a super idea, Jim.

Note: If you would like a copy of the circuit diagram for Jim's innovative "Short Stop" circuit, send a SASE to Dave Miller NZ9E, 7462 Lawler Avenue, Niles, IL 60714-3108 and indicate the circuit name on your request. No requests will be honored without a self-addressed and stamped envelope (SASE).

Driving home the point

From Phil Salas AD5X of Richardson, TX: "For outdoor ground connections, I've been using 5/8-inch ground rods, instead of the more common 3/8-inch or 1/2inch variety of rods. The 5/8-inch ground rods are much sturdier and therefore much easier to drive into the ground and an 8-foot ground rod is under $10 at Home Depot (reg. trade name). To make it even easier to drive the rod into the ground, first dig a hole about a foot or more deep and fill it to the top with water. Allow the water to seep into the ground and soften it up, then drive in the ground rod. The deeper the original hole, of course, the less effort will be required to drive in the rod. I also use #4 stranded copper house wire as the ground wire itself. Strip off the wire's insulation and then attach the copper wire to the ground rods with bronze or brass ground clamps. Less expensive aluminum clamps will cause galvanic-action problems when used with copper wire, especially outdoors. The #4 wire and bronze ground clamps are also available at Home Depot.

On a little different topic, I've changed over from using hose clamps for attaching telescoping antenna tubing sections together. I now drill and tap a #6 screw hole in the larger diameter tubing element and then just 'snug-in' a standard #6 machine screw to hold the inner tubing section in place until I find the final dimensions. Then I remove the #6 screw, drill through the inside tube as well, and attach both tubes together with a #6 stainless steel sheet metal screw. It's a much better (more reliable) connection in the long-run and you can still change the over-all dimensions at a latter date if the need arises.

Moderators notes: As an addendum to Phil's first suggestion, I've also heard of using a husky electric "hammer-drill" chucked onto the ground rod, as an aid in driving it into stubborn soil. If you don't have an electric drill with a chuck large enough to accommodate the larger diameter rods, you might be able to find one at your favorite local tool rental center...and of course Phil's hole-and-water idea still applies to make the job that much easier.

If you utilize Phil's sheet metal screw (instead of a hose clamp), for holding telescoping antenna parts together, don't be tempted to "improve" upon it by drilling another through-hole for a second screw (on the opposite side) at the same level as the first...you could weaken the tubing to the point of seeing it bend over during a high wind condition. If you feel that a second screw is needed, place it at a different level on the tubing's face.

Correct drive for modern linear amplifiers

From Richard Measures, AG6K, of Somis, California: "Back when amplifier-tubes were less costly, two-tube grounded grid amplifiers were more common, but today, with tube prices much higher, single-ended amateur linears have made greater in-roads. There's nothing wrong with single-tube linear amps, in fact, with the elimination of the possibility of a push-push parasitic oscillation taking place between two tubes, and the resultant "big-bang" taking out expensive tubes and other parts, inside of your linear, is actually a bit less! The biggest single drawback to modern one-tube amplifier design seems to be that none of the currently available Hi-Mu tubes are particularly compatible with the 100+ watt output drive from the typical amateur HF transceivers in use in most of our shacks. The result can be grossly overly-driven and splatter-prone "linear" amps...inhabiting our ever more crowded HF bands.

If you look at the specs for today's popular amplifier-tube crop, you'll find that the bulk of the new ceramic triodes require anywhere from 22 watts to 75 watts input RF for full output, anything beyond that results in saturation drive. Yet how many popular transceivers are rated in the 22 to 75 watt range? Some have effective output drive reduction circuitry, but many do not.

Here's what happens: most of us who use linear amplifiers simply depend upon the ALC (Automatic Level Control) within the transceiver, and/or the ALC fed-back from the linear, to take up the slack, but does that really answer the question effectively? If you examine most ALC circuits, you'll find that the ALC time-constants are such that nothing is done until the amplifier has already been overdriven...a little like closing the barn door when the horse is already most-of-the-way out! This allows the first syllable (or so) of each word - in the case of SSB speech - to overdrive the amplifier to varying degrees, depending upon the actual ALC time-delays in your set-up and the power in the operator's voice (as well as in the strength of the vowels and consonants of the word being spoken). You can hear this effect if you own two fairly equal receivers. Try tuning one to the center of a distorted sounding SSB signal, and the other to a point 4 kHz or so away, and then listening to the two receivers with one side of a stereo headset connected to each. Every word -or part of a word - that's overdriving the sending station's "linear" amplifier will be heard simultaneously in each receiver, whereas proper drive will be heard only in the "on-frequency" headphone. Listen around and try it...it's an interesting experiment. It's neither good for our relations with our neighbors on the bands, nor is it healthy for the expensive tube in our amplifiers, to permit such a condition to exist over the longer run.

There's a simple and inexpensive way to correct the problem, however, if you're willing to do just a little bit of investigation. Grounded-grid amplifier tubes are driven via a coupling capacitor, transferring RF energy directly into the tube's cathode - since the grid appears as ground to RF. The proper value of resistance, inserted between that coupling capacitor and the cathode, will absorb any excess driving power, keeping it from overdriving the tube...it's a form of negative feedback. The harder the cathode is driven, the more drop across this new resistor, exactly the effect that we want, and it happens automatically and instantaneously. A secondary benefit from this scheme is that the IMD (InterModulation Distortion) level attainable using this method, is actually better than IMD levels attainable with driving the tube at its rated power! Here's the rub, it's impossible to give you an exact value for the resistor you'll need in your particular amplifier - that's where the "investigation" part comes in - but I can give you a practical example to help you get started.

The commercially available unmodified single 3-500Z HF linear amplifier that I was using for this example, reached maximum output power with only 55 watts of input drive. The actual anode current observed at maximum power output will vary somewhat with the emission characteristics of the particular tube installed, and in this particular example it was 400 mA. Driving this amplifier then with a 100 watt input (45 watts of over-drive), only increased the anode current by 100 mA (or up to 500 mA total anode current). That non-linear increase represents the distortion product in this otherwise "linear" amplifier. After installing a 20 ohm resistance in series with the tube's cathode drive coax cable, however, the result was an observed anode current of 420 MA (at the full 100 watts of drive), or within 5% of the tube manufacturer's power specs, and the amp showed excellent IMD performance (22 ohms would probably have been perfect). To determine the wattage rating for the resistance needed, you'll have to take into account the service that the amplifier will normally be subjected to. Full-power, long-winded RTTY or Data transmissions will require about 40 or 45 watts of dissipation capability from the resistor, but ordinary SSB and average-duty CW will bring that figure down considerably...perhaps to only 8 or 10 watts. The resistance will probably end up having to be custom made, from several flame-proof, metal-oxide-film non-inductive resistors, in the 2 to 3 watt class...to achieve the final value and wattage rating desired. Digi-Key Corporation of Thief River Falls, MN (Tel: 1-800-344-4539) sells metal-oxide-film resistors if you can't obtain them locally, but we still have to figure out the actual value we'll need.

Unfortunately, there's no hard and fast formula that will give us this exact value, the process instead combines an educated guess along with some cut-and-try. The resistors needed are inexpensive enough, so if you fail to come as close as you'd like on the first guess for the final combination, a new set of resistors is still a good deal less costly than a new tube!

Start by verifying the accuracy of your amplifier's Ip (anode current) meter, then using the example given above as a guide, install the combination that you think might be closest, resistance-wise and wattage-wise, to what your particular tube(s)might need. Tune up the amplifier with maximum CW driving power applied from your particular transceiver. If the anode current is still beyond what the tube manufacturer recommends (from the manufacturer's spec sheets), then the input cathode (pad) resistance needs to be increased; if the anode current is too low, then you've added too much resistance - that's the cut-and-try part. By the way, metal-oxide-film resistors won't "flame-out", but they will decrease in value if their wattage rating is consistently overloaded; they'll also open-up like a fuse if a catastrophic overload occurs. That last point actually offers some degree of protection for the tube in the event of a true catastrophic overload. If you allow some "cooling space" between resistors in a series/parallel "bunch", a 2 watt MOF resistor is often capable of dissipating 20 watts for 5 seconds without incurring permanent damage. That kind of overload would more than likely permanently change the value of a carbon composition resistor, but metal-oxide-film devices are considerably more tolerable of the presence of the occasional overload. Once you've determined how much resistance is needed and what wattage dissipation will give you the margin of safety you'll feel comfortable with, you can install the permanent resistive "pad".

Here's a final thought, perhaps in the past you've been tempted to reduce an obviously high anode current reading by underloading your amplifier with its tuning controls. This has been widely done over the years by many in our hobby, but it's a poor technique from a linearity standpoint. When an amplifier is underloaded, grid current increases, also increasing the non-linearity and splatter. All amplifiers should be tuned so that most of the electrons leaving the cathode end up at the anode, not diverted to the grid (as underloading does), by too much load capacitance inserted during tune-up. The best way to tune an amplifier, linear or non-linear, is to peak it for maximum power output, with full drive applied. If the anode current is beyond the manufacturer's specifications, then reduce the current by the method described using a series cathode resistance, not by underloading. Also, even though your amplifier's manual may tell you differently, it's only possible to correctly tune a high-power linear for maximum output and maximum linearity by applying full power to it, preferably into a well-shielded 50 ohm dummy load. You can also pulse the amp on and off at a rapid rate and mentally add-to the difference in meter readings, or you can use a two-tone oscillator and a monitor scope for tune-up, but most hams find these too clumsy and opt for the maximum CW key-down tuning technique instead. Once the amp is correctly tuned into a 50 ohm load, it can be switched into a correctly matched antenna system with no re-tuning needed. Making your transmission line and antenna appear to be a 50 ohm load to your amplifier has been widely covered elsewhere, and it's usually quite easily accomplished with a properly rated and pre-adjusted matching network, or what's more commonly called an "antenna tuner."

Moderator's note: I've incorporated Rich's suggestion in my own Heath SB-1000 HF linear amp with very good results. I chose to bring out the RF input circuit to the rear panel of the amp via an RCA phono jack and then the single 3-500Z's cathode RF feed back into the amp via another RCA jack (using small diameter Teflon (reg. trade mark) coax cables. I then mounted the resistive pad Rich described in an external, shielded mini-box attached to the rear panel of the amplifier externally. This allowed me to experiment with the resistive pad values outside of the amplifier's case. I ended up using four 8-ohm, 20 watt, non-inductive resistors in series (Radio Shack #271-120) to achieve the results described by Rich. Another "well done" goes to AG6K for helping to clear up an area that's often misunderstood by many of us in ham radio. Rich has done a great deal of research into the causes and cures of linear amplifier maladies, and his suggestions, as always, are well worth seriously considering.

I hope that you've found the ideas in this month's issue to be useful, and again I invite you to submit your own favorites to share with the rest of 73's readership.

Murphy's Corollary: No matter how noble the result of an experiment may be, there are always those who will ignore or misinterpret it and continue along with their own "pet" theory.

A very special thanks to this noble experimenters:

Jim Wood

c/o Inovonics, Inc.

1305 Fair Avenue

Sanata Cruz, CA 95060

Phil Salas AD5X

1517 Creekside Drive

Richardson, TX 75081

Richard L. Measures, AG6K

6455 La Cumbre Road

Somis, CA 93066

Note: The ideas and suggestions contributed to this column by its readers have not necessarily been tested by the column's moderator nor by the staff of 73 Magazine, and thus no guarantee of operational success is implied. Always use your own best judgment before modifying any electronic item from the original equipment manufacturer's specifications. No responsibility is implied by the moderator or 73 Magazine for any equipment damage or malfunction resulting from information supplied in this column.

Please send all correspondence relating to this column to 73 Magazine's Ham To Ham column, c/o Dave Miller, NZ9E, 7462 Lawler Avenue, Niles, IL 60714-3108, USA. All contributions used in this column will be reimbursed by a contributor's fee of $10, which includes its exclusive use by 73 Magazine. We will attempt to respond to all legitimate contributor's ideas in a timely manner, but be sure to send all specific questions on any particular tip to the originator of the idea, not to this column's moderator nor to 73 Magazine.

Note: If you would like a copy of any of the circuit diagrams or figures referred to in

this column, simply send a SASE to Ham To Ham Column, c/o Dave Miller NZ9E,

7462 Lawler Avenue, Niles, IL 60714-3108 and indicate the month and circuit or

figure name on your request. No requests will be honored without a self-addressed and

adaquately stamped envelope (SASE).