# Voltmeter Autorange Schematic

Figure 3. Schematic of the auto-polarity voltage monitor, ponents Rl and R2, measure the actual value of the dual lOkO potentiometer, Most inexpensive varieties top out at less than lOkQ.In my case, 1 found H420Q the maximum value. I obtained several pots and chose ihe unit whose individual maximum values were most nearly equal. The closer the values, the belter the tracking of the two supplies.

The rest of the process is a matter of follow ingdafa book formulas, w ith a fewr adjustments. For both regulators, the output voltage (V'o) equals the sum of Ri and R2 divided by R2. with the result multiplied by the control voltage (Vc). For the 78GU1. Vc is 5V, and for the 79GUI it is 2.23V, The data books recommend a I mA control current, which gives 5kO and 2.2kO for the positive and negative regulator values of R2. With the less than nominal maximum pot value, however, (8420Q instead of 10 k), 4,2kQ yielded a positive output range from 5 to i5V with the 78GU1. Applying similar logic to the 79GLM negative regulator gave a value of 1875Q for R2 and a series resistor of 2345Q for the potentiometer leg. Breaking the fixed values into fixed resistors and circuit board trimmer pots yielded the values in Figure 2.

The procedure for calculating the negative regulator values begins w ith the formula for RL which equals R2 times the difference between the output vottagc (Vo) and the control voltage fVc), all divided by the control voltage. At minimum output (5V* when the potentiometer is also at minimum. Rl will be 1,25 R2, and at maximum voltage (15V> with the pot also at maximum, RI will be 5,73 R2, Since the difference between Rl at maximum and minimum is 8400 O ithe measured range of the pou, 5.73 R2 minus 1,25 R2 will also be 8400 Q. R2 thus equals 8400 divided by 4.48 (i.e., 5.73 - 1.25). or 1875 Q. A I500Q fixed resistor plus a 500 QPC board trimmer give room for adjustment. Since RI at minimum (when the pot is at 0 Q) is 1.25 R2, which we just set at 1875 then the series resistance for the Rl pot leg of the control circuit will be 2345 Q. A 2kO resistor plus a 500 fi trimmer again provide room for adjustment.

Calibrating the tracking is easy. Check the positive supply range and later the 4.2kfl resistor until the output extends from 5 to \5\\ When the positive supply voltage range is satisfactory, then set both negative PC board trimmers to mid range. With the dual pot set to minimum value, adjust the trimmer in R2 to let the negative voltage equal the positive voltage (about 5V).

Turning the dual pot to maximum, adjust the trimmer in the R I lee to let the maximum negative voltage equal the maximum positive voltage. Ret rimming both the circuit board pots once more should yield stable results. Now output voltage tracking will depend on how well the dual pot sections track Even with inexpensive dual pots, the variation should run less than a tenth of a volt between supplies, If the error is greater, but consistently high or low for one supply, then repeat the calibration, If the error varies across the voltage range, a different dual pot may be in order. Careful pot selection can thus save the cost of precision potentiometers and the complexity of electronic tracking circuitry.

As the photographs show, construction is not critical. Perfboard works very well when placing one set of stand-off pillars under the transformer mounting wings, rather than at the corners of the board. Use heat sink grease between the variable regulators and their finned sinks. Also note that the pinouts of corresponding positive and negative regulators differ. Do not let local QRM cause a reversal here! My experience says that at least one of the regulators will fry at first test.

My perfboard construction techniques make use of 1-46 pins for off-board connections. In this project, there are many board-to-hoard and board-to-panel connections, so have a good supply of pins and plan their placement carefully. Since, as the photos show, the project will fill the case, be sure that no pins interfere with cabinet screws,

### Voltage Monitor

The voltage monitor makes use of some automatic metering principles previously reported in 73, but updates them to eliminate the need for measuring the forward voltage drop of diodes and for taking the drop into account when designing the power supply section. By using op amps to sense a portion of the positive and negative voltages, it's possible to control the voltage and not exceed the op amp limits, w hile still making accurate measurements. The trade-off for this convenience is the need to calibrate the circuit carefully.

As in the earlier voltmeter circuit, it begins with a timing clock {7555 or 555), The clock circuit provides short square wave pulses, as the on-time is controlled by the !5kQ resistor. Fhe diode shunts the470kQ resistor during capacitor charge, but the discharge goes through the resistor, extending the off-time to about 3 seconds. The 4013 flip-flop provides alternate 3 second periods for reading positive and negative values. The builder can alter the 470k0 resistor to change the read periods, or insert a 500kf2 circuit board trimmer pot in series with a 330kf} resistor to provide for an adjustable period.

Figure 4. Schematic of (he uuio-poiariry. auto-ranging current monitor,

The 4013 D-type flip-flop is in a sunward divide-by-two circuit. Be sure to bypass pin 14 to ensure good action. Some combinations of clocks and flipflops will miss some or even ail beats without it. The extra portion of the two-section chip keys the indicator LEDs and provides clocking pulses to the current monitor. The Set and Reset pins {6 and 4} are grounded for clocked operation through 20kO resistors. Setting either high with the single polet double throw, center off toggle switch will override the clock, locking the readout either positive or negative, along with its indicator LED.

If complement-Q controls the positive readout, then a high Reset line locks the circuit for continuous positive readout, Likewise, a high Set pin locks the Q output high for continuous negative readout. Returning the switch to center allows the clock to take over, and the circuit will cycle with the neM clock pulse. Thus, There is a simple but effective manual override for the automatic circuit whenever we wish to closely monitor any one of our voltage or current readings.

The voltage sensors consist of unity gain buffers following a simple resistive voltage divider network Feeding between one-tenth to one-third of the voltage to the buffers ensures that the voltage will never rise to near the op amp supply voltage. Near that point, op amps cease to amplify linearl) , and accuracy deteriorates. For this circuit, the network provides 20% of the supply voltage to the buffers. The positive voltage buffer is non-inverting, while the negative buffer inverts, The result is ihat the rest of the circuit always gets a positive voltage. The circuit is Similar to one developed by Pepper {Radio-Electronics + March 1983, page 64).

Following the buffers is the 4066 bilateral switch, whose individual switches close when their associated control pins go High according to the output from the flip-Hop, : he 4066 is an improved version of an earlier switch chip and shows a resistance of only about 80Q per su itch section. This low resistance is insignificant for these circuits, A DC amplifier follows the sw itch to set the voltage fed to the meter circuit.

i:or some applications, the builder may wish to use separate DC amplifiers for each buffer and install them ahead of the 4066 switch. A gain of 3,3 provides a maximum of 10V for an analog meter, well within the op amp limits. For use with digital voltmeter circuits, adjust the gain of this amplifier according to need. For example, with an original one-tenth sample at the resistor divider and a unity gain amplifier at this point, the circuit w ill shou .5 to 1.5 V for power supply settings of 5 to 15V (either polarity). Digital measurement would thus require only a change in decimal point position.

Since the DC amplifier is non-inverting, the feedback resisior network of Rf and Rg is easily altered. Keeping the lOOkO trimmer and the lOOkQ input resistor, Rf then equals lOOkQ times the difference of the desired gain and I , with another JOOkfi subtracted for the trimmer pot. Unlike the unity gain buffers, which require no offset balancing, the DC amplifier shows a remnant offset voltage which detracts from accuracy. The external balance circuit at the inverting input terminal provides easy adjustment for no output when the lOOkQ input resistor to the non-inverting terminal is grounded rather than connected to the 4066.

The analog meter circuit uses a Radio Shack 0-io-lSV voltmeter w hieh comes with an external series resistor of about l5kG* This circuit replaces the series resistor with another combination to drive the I mA meter movement. Since the op amp output voltage corresponds to the sensed voltage, 10V from the DC amplifier equals 15V at the resistor divider terminal, Ten volts at I mA requires a lOkO series resistor, which is shown as 9.1kQ plus a 2kD trimmer pot for calibration.

Construction is not critical here either, since only DC and slowly timed pulses are involved. Use IC sockets for construction ease. As with the supply board, use pins liberally for off-board connections. I he 3x3 'A-inch pertboard squares shown in the photos easily hold the circuitry with room to spare. Be sure to place the trimmer pots in easy access areas for calibration after mounting the board in the cabinet. Since lead length makes no difference, mounting all trimmers along the top edge of the board will ease later calibration,

### Voltmeter Calibration

Calibration of the voltmeter is a cinch. Balance the DC amplifier by grounding the lOOkQ input resistor at the switch (4066) end and adjusting the trimmer marked Rb until the output is zero. Adjust the input voltage setting trimmers so that the voltage to the buffers is one-fifth (or the desired fraction) of the power supply voltage. Using a convenient voltage, adjust the DC amplifier gain trimmer, Rg, so that the output is 10V (or the desired amount) for 15V from the power supply. Finally, set the meter trimmer, Rm, so that the meter shows 15V for 15V from the power supply. I he only other cautions concern the flip-Hop. Identify the positive and negative control lines and be sure that the manual override switch and indicator LEDs correspond correctly to these lines. Identify the positive and negative control lines for the current measuring circuit so that the meters will read together.

### Current Monitor

Although voltage monitoring circuits are growing more common in bench supplies, there's still little useful current monitoring. A single meter for gross current measurement provides little help for monitoring low current circuits, while a sensitive meter pegs long before the supph nears its maxi mum rated output. Automatic monitoring of both positive and negative current drain appears only in expensive industrial and lab equipment in the S2.500 and-up class. A simple, reliable, and effective current monitoring circuit, however, has long had a place in the data books.

The current monitor in Auto-VIM owes much to National Semiconductor s Linear Databook circuit for routinely converting current drain to a voltage output without resorting to ultra-precise resistor matching.

1 he sensor circuits in Figure 4 use different op amps to sense positive and negative current flow, The TL081 lor 1 .F35I > Bi-FETop amp uses P-channel inputs which work with input voltages close to the positive supply value, but fail as the input voltage approaches the negative supply voltage. By contrast, the newer Tt N FET op amp, the TLQ91, with its N-channel inputs, shows precisely the opposite characteristics. Between the two, we ob tain separate but parallel sensors for positive and negative supply currents.

The transistors, whose base current is controlled by the op amp output, control the voltage seen at the 3.3kQ resistor. In fact, the circuits provide an output voltage per mA of line current equal to .001 times the product of Rl and R3 divided by R2. The circuit shown provides ,0033V per inA, or 1,65V at 500 mA, Sensor circuit output is positive for the TL081/2N3904 combination and negative for the TL09I 2N3906 duo. Although most data book circuits show FETs rather than transistors used with the FE1 input op amps, the bipolar transistors work better at the 5V end of the power supply range. Note the 10-turn trimmei pots marked Rb, which will receive attention during circuit calibration.

A DC amplifier follows each sensor to increase ihe voltage to a level desired for measurement. As with the voltage monitor, ihe negative amplifier inverts while the positive does not, thus yielding positive voltages for the bilateral switch. Each amplifier has a gain of 6.7 so that the metering circuit w ill see 12V at 500 mA, which is uithin the linear range of the op amps and within the switching range of the 4066. Each section ol the LF353 includes an offset balancing circuit to decrease errors introduced by remnant voltage outputs, (to be continued)^

73 Review t

Number 6 on your Feedback card fci William J. Mick K3RVH

Heathkit SB-1000 Linear Amplifier

Shake and bake test results of a low-priced amplifier from Heath

Heath Company Benton Harbor M 49022 Price Class: S700

Whew!

Heath had me worried there a while. For years, hams could count on them as a reliable source of HF linear amplifier kits for the amateur market. Then, to widespread dismay, ham amplifiers disappeared from Ihe product line. Many hams would think twice about building a micro-controlled transceiver, but an amplifier is something comprehensible, probably even repairable, the perfect kit-based station addition. Fortunately, Heath came up with something nice to fill the vacuum, The SB-1000 proves that hams stiEl have friends in Benton Harbor,

The SB-1000 is a classic tuned cathode fed, grounded-grid design using a single 3-500Z power triode, The basic circuit has been an ARRL handbook staple for years. Variations on the 3-50GZ theme, with one or two Jbot-lies/' have been marketed by numerous manufacturers since the mrd-seventies, In fact, under the skin, the design of the Heath SB-1000 is based on the popular AL-80A linear built by Ameritron.

The SB-1000 is not a kit for the first-time kit builder. It contains a 3000 VDC supply, and obviously demands a workmanlike approach in assembly and checkout, Don't rush through this kit. or skip steps, unless you iike to live dangerously.

### Packing

Everything comes in a fifty-pound box internally sectioned with separate boxes and bags of parts to support each phase of construction. The manual and accompanying foldout charts are of the usual high Heathkit standard. There are afso a few pages of errata, mainly typographical and pictoriaE corrections, that must be integrated with the instructions. None of them appear to be concerned wiih the kind of detail that endangers life or property.

### Subassembly

The amplifier is built on a heavy gauge steel chassis, with an internal partition that separates the RF deck from the power supply and control circuitry. In typical Heath fashion, the first few evenings of construction concentrate on various subassemblies. For rne, there were about twelve solid hours of piecework before the thing started to look iike the picture in the catalog, tn a moment of weakness. I felt victim to the old kit-builder s affliction canrwaititis. There J was. loosely bolting the chassis oan-els together for a preview of the final product on the operating bench. This of course aggravated the condition, as the SB-1000 looks pretty good. The compact size, simple control layout and gray-toned color scheme go very nicely with most modern ham equipment.

### PS Rectifier Board

The ftrst piece to build is the power suppy rectifier board which requires aboui an hour. This board needs double checks of rectifier diode polarity, as each is soldered in place. The rectifier board was not easy to solder, even though it was pre-tinned. The instructions stated about two or three seconds of heat per joint, but it took twice as long as that to yield sound joints. Fortunately, all the diodes survived despite the additional heat.

The power supply filter board takes another hour or so. and took solder much more easily than the rectifier board. It uses ten large electrolytic capacitors in series, so polarity is crucial. The instructions state that the polarity of each can should be rechecked when the board is complete, but (couldn't doit. Thefinte plus signs on the capacitor tops were out of sight under the board, and not even a dental mirror helped. Builders who don't have a proctoscope may want to mark the sides of the cans before bolting them to the board

Less than two hours were required to build the boards that handle ALC. power measurement, and meter switching. Construction is fairty simple, except for diode polarity and a coupie of muiti-colored wiring harnesses.

in contrast to the circuit boards, the input filter unit construction demands patience and dexterity. This unit is a small shielded box that surrounds part of the bandswitch, and it contains a number of slug luned coils and capacitors. These form individual pi-networks for each band, and there is ample opportunity to connect the wrong bandswitch contacts, or to short some of the longer leads Liberal use of spaghetti tubing and artful dressing of the numerous wire leads will avoid problems. The Heath assembly pictorials are very clear, and therefore invaluable at this stage The coil forms snap into holes in the sides of the filter box, and carefully controlled leverage is the only way to install Ihe coils without breaking them.

"Special" RG-58/U

At one stage of input fitter assembly the instructions call for a length of smafl coaxial cable After twenty-five years of hamming I tend to associate the term "small ' coax with something like IRG-58/U. In due course a piece of cabEe marked RG-58/U was found among the SB-1000 parts, but the length wasnt right A piece of subminiature coax was discovered, and ils length did correspond to the instructions, so obviously this was intended for the input circuit.

I've used subminiature 50Q coax with 100 watt transmitters before, so the use of this really small cable wasn t too much of a surprise, However it was now obvious that the 1 large'1 cable used in the amplifier's output circuit was the piece of RG-58/U. This caused me some concern, Consultations with several other long-time hams didn't otter any comfort The ARRL Handbook tables show 650 watts and 1900 VDC as the upper (albeit conserva lively rated) limits for RG-58;U. In fact, the SB-1000 operating instructions actually recommend that RG 58 and RG-59 feedlines be avoided in favor of heavier RG-8 or RG-11 coax.

I made ¡nquiries of both Heath and Amer-itron concerning the use of RG-58/U for the amplifier's output circuit, and received quite similar replies. Although the piece of cable in question is simply marked RG-58/U, it is actually a special Teflon1" insulated cable rated at 2500Vr This is not garden variety RG-58? and it is certainly easier to handle lhan RG-8 when Wiring up the amplifier. Cables of this type can be found in a number of modern commercial amplifier products. Its heat resistance is a useful property inside power tube enclosures.

The rear part of the bandswitch assembly handles switching of the plate tank circuit, a tapped pi-network design that incorporates a big tapped toroid mductor for 80 and 160 meters. Doorknob padding capacitors are switched into the circuit on the lower bands, which permits ihe use of reasonably sized variable capacitors. It's a compact and practical design, and looks a great deal like handbook amplifier designs of recent years, excepi for the bandswitch.

### Bandswitch

Almost every homebrew transmitter, amplifier, or high-power ATU Ive ever built has involved a careful search for a heavy duty wide-spaced RF switch for the tank circuits. The SB-1000 bandswitch, a CentraLab designer-type unit, is not typical of handbook amplifiers, which usually specify something tike the Miflen 51000 RF switch, oraheavy-du-ty surplus monster. Now. it's hard to believe that iT or the ARRL. have been overbuilding power amps all this lime. On the other hand, neither homebrewers nor the ARRL lab are much constrained by the realities of commercial competition.

In response to my queries, Heath stated that the ceramic bandswitch in the SB*1 GOO is conservatively rated at 9 amps AC and at 2500V. Furthermore, Heath said that the SB-1000 was run through a rigorous series of FCC tests involving all manner of electrical abuse without any switch problems Ameritron pointed out that similar switches have been used on kilowatt linear amplifiers of various manufacturers, including Drake, Swan, Dentron, and Heath itself, for some years.

### Final Assembly

At this point. I attempted to suspend my prejudices and do the appropriate thing: finish assembly of the amplifier and proceed to beat the hell out of it. The rear panel went together m about three hours, complete with heavy duty primary power relay and RF-fiitered AC cable. Phono plugs for external RF relay control, 12V accessory support, and ALC output voltage were also wired up. Also on the rear panel is a safety interlock switch that cuts the AC power when the SB-1000 lid is removed.

The center partition panel holds the two big transmitting capacitors and coohng fan, and is an easy job. Likewise, the front subpanel, with meters and accompanying meter lamps, went together smoothly, right down to the Jackson vernier reduction drives used for the tuning caps. The method used to mount the meters is not very rugged being a couple of solder lugs ai diagonal corners but it does hold once the panels are bolted together.

integration of the front subpanel and center partition with the chassis base was not easy. A fair amount of warping and twisting is necessary to make screw holes and capacitor shafts line up properly. Various hardware items must be loosened and aligned to permit smooth control rotation and squaring of all the corners. Another hour or two saw the installation of the power supply rectifier and filter assemblies, and connection of the rear panel Numerous flying leads and wiring harness ends must be interconnected. Except for minor glitches, like a couple of bad screws, and an out-of-reach solder junction, everything went together pretty much according to the detailed instructions. There was one resistor whose leads could not be trimmed to the specified length because they were already too short.

Fitting of the front fascia and the 3-5002 tube were almost anticlimactic, and after about fifteen hours of construction, the SB-1000 was at last ready for testing.

For obvious reasons, one does not plug in and go at this stage. 1 spent a good forty-hve minutes venfymg connections and checking for solder bridges and pinched wires. The rear panel barrier strip was wired for 240V AC input, so the 120V plug was removed and a suitable 240V plug was installed. The Jid was set in place in order to engage the interlock, and the unit was plugged into the AC mains.

The thing I hate about high voltage equipment is having to get near that front panel the first time the main switch is thrown. I pushed the S8-100Q power switch with a piece of broomstick and a resounding THUNK! shook the house as the transformer field sucked in the sides of the loose lid. The power supply hummed a bit, but there was no sparking no arcing, no smoke. The front pane! voltmeter showed 3300 VDCh and the 3-500Z glowed encouragingly. Home-brew or kit-built, you get a lot of satisfaction when you first put the juice to the product of your labors and nothing bad happens.

### Final Pre-op Tuning

A detailed set of alignment procedures takes the constructor through the tuning of the input matching networks. For this step, the lid must be slid back a fraction of an inch from the front panel while a nylon alignment too! is used to peak up the coil slugs- This is definitely a situation where one hand stays in the pocket, It is worth enlisting a second operator to dictate the instructions, key the exciter for youT and act as saiety man. Everything went by the book, and after heating up the Heath Cantenna tor awhile, the moment of truth had arrived—it was time for the S8-10Q0 to speak to the world,

SB-1000 On The Air!

it spoke very well for itself, reaping a lot of favorable comments and no criticisms from any of the stations worked. An FT-102 and an IC-751A were initially used to drive the amplifier, and Iboth had more than enough power to yield full output from the SS-1000 In fact, as the amplifier is rated at 85 watts maximum drive, the exciter carrier levels were reduced slightly to avoid overdriving it. The input tuned circuits are fairly broadbanded, but ihe drive sometimes has to be brought up a bit when the frequency is away near a band edge.

The SB-1000 has adjustable ALC output of up to twenty volts to help regulate the drive for a clean signal, although this required some fiddling to adjust- The first few months of testing took place in a club environment, and the ALC connection seemed to be unplugged as often as it was connected. It is probably just human nature to resist something which seems to be retarding those satisfying meter swings.

During three months of on-the-air use, there was no evidence of breakdown or other inadequacy in the special RG-58/U, the bandswitch, or even in the antenna relay, which is a plastic insulated AC power type. The antennas used were a well-worn tribander beam and a G5RV multiband dipole, without benefit of an ATU.

The testing environment was one of the worst possible: a club hamshack. A number of hams had the chance to stress it and abuse it, and that's just what they did, mostly unintentionally. Appliance operators accustomed to auto-tune rigs have treated the SB-1000 like a broadband device, changing the bandswitch, but neglecting the tuning controls, Some people take a long time to tune up or forget to watch the grid current meter. In spile of the rough treatment, the amp worked well until halfway through the ARRL DX contest.

What actually happened no one seems to know, except that serious internal arcing was heard on 15 and 20 meters. Examination showed that the safety RF choke at the output of the pi-network was open, but it was hard to tell if this component was the cause or a victim. It was clear that some pretty high RF voltages had jumped from a stator contact on the bandswitch to the metal shaft There was a great deal of carbon build-up on the rotary wafer, but the rotary contacts were clean. The stator contact, which connects the padding circuit for the plate variable capacitor, was eroded completely.. Replacement of the small RF choke and removal of the doorknob cap permitted operations to resume on 80 through

10 meters, but the bad wafer will have to be replaced in order to reenable 160 meter capability.

Despite my original misgivings about the switch, I don't feel that it was the source of the problem. It did work for three months, and it look a contest operation with a lot of different operators to bring about the failure An insulated shaft might have prevented the arc-over, but after seeing the amount of dust and fuzz collected on the air i ntake vent and on the floor of the RF section, I am inclined to believe thai it was dirt that started the arc.

### Top Dollar Value

The price of the SB-1000 is very attractive at about \$700. Compared to the big "dollar-a* watt amplifiers on the market, this amplifier is an economical way for a guy who isn't GRO-crazed to boost his signal when he needs to, It is not built to be indestructible, but it isn't priced like a continuous service ampt either. While the amplifier is advertised as covenng 160 to 15 meters, any technically competent ham will be able to figure out how to make it work on 10 meters as well. If you know what you're doing, you won't even have to buy any additional parts.

One thing Heath might do is to review the list of required tools, which seems to be the same list whether the kit is a simple noise bridge or a high power antenna tuner. I found several points during construction where a 25 watt soldering pencil was not hot enough for the job A 100 watt gun was needed for some of the heavy power supply and tank circuit connections, and it was necessary to use a 250 watt gun when soldering the safety RF choke in the output circuit to the bandswitch frame.

### Conclusion

Heath, traditionally the friend of the ham on a budget, has once again provided a cost-effective piece of gear that will do both the manufacturer and the constructor credit. Thanks are due to Denton Bramwellat Heath and Tom Rauch at Ameritron for their courteous and helpful responses to my questions. Although I personally would have selected heavier com ponents for the bandswitch and antenna re-Jay, I am satisfied that the supplied parts are adequate for the job, The club members are enthusiastic about the SB-1000. and while we have asked more of the amp than we should have, we look forward to a lot more heavy use in the future. KJ