In the fall of 2009 I acquired an Ameritron AL-80B linear amplifier for my ham radio station. My amp was purchased used. It employs a single 3-500Z tube and is designed to produce up to a kilowatt of signal in SSB service on the ham bands. I bought this amp because I did not have one capable of operation on the 160m band, and I needed an amp for that band. I like the 3-500Z. It is a tough tube which typically enjoys a long lifetime, but is relatively inexpensive to replace if necessary. Plus, it is a glass tube, and nothing beats having a glowing filament and plate at your side when operating on a dark winter's night.
The AL-80B has proven itself a good performer. I especially like the dual crossed-needle meters which quickly provide you with all the information you need to know to tell how the amp is operating. The amp is reasonably quiet in operation and not difficult to tune. Of course, nothing is perfect, and I have noticed some anomalies with the amp in the course of using it. What follows is an attempt to describe what I've found and capture what I have done to address these concerns.
The first thing any ham typically does after purchasing a new piece of equipment is to put it through its paces. With my amp, I noticed that the B+ voltage, which should nominally run about 3100V on the AL-80B, was instead running only about 2800V, as indicated on the panel HV meter. The Power Out (PO) also seemed a bit low; the amp was only putting out about 840W on the 40m and 20m bands (normally the "hottest" bands for an amp) and about 100W less on 160m.
I was curious if the other voltages were also off normal, so I measured the filament voltage and the voltage on the 12V bus. Oddly, these were both high; the filament voltage measured about 10% above the 5.0V it should ideally be and the 12V supply was running in excess of 15V. The 12V supply didn't concern me much, as the internal components it powers (mostly op-amps) are all rated to something like 32V, but I found the filament voltage a bit disconcerting.
Initially, I was powering this amp from a fairly soft 15A 120V circuit. I realized this was not ideal, and used the Thanksgiving weekend to pull a 20A 240V circuit into my shack. I elected to use 12 gauge wire (20A capacity) because I had some spare 12-2-with-ground ROMEX on hand and the total run was very short; less than 25 feet from the breaker to the amp (including the power cord).
Rewiring the amp for 240V is well documented in the Ameritron manual and was a breeze to do. The hardest part was forcing myself to cut the nicely molded 120V plug off the line cord in order to replace it with a 240V equivalent.
After the conversion, the amplifier did indeed put out a kilowatt on 40m and 20m, but just barely. It is still perhaps 100W shy on 160m, not that the difference would be noticeable on the air.
One comment on the 240V conversion: The power switch on the front panel of the AL-80B is a SPST design, which is fine for 120V-to-neutral, or even the 240V-to-neutral service found in other areas of the world. However, with the typical American two-phase 240V service, with the power switch in the off position one side of the transformer remains live with respect to ground. This only presents a problem if there is a fault, but still, I'd rather have seen Ameritron use a DPST switch here.
As indicated earlier, the filament voltage my amp presented was higher than I would have liked to see: After the 240V conversion, I was measuring 5.5-5.6V at the base of the tube. What constitutes an excessive filament voltage has turned into an almost religious argument in the ham community. The purists will point to the Eimac spec of 5.0±.25V, and recite the empirical data that every 3% rise in filament voltage over spec will cut the tube's useful life in half. Those in the opposite camp will argue that in ICAS service tube life is never what it is in commercial service, and other things, like repeated power cycling, will come back to bite you well before the filament wears out. Still, the higher-than-necessary voltage bothered me.
A couple points are worth making here. First, to accurately measure the filament voltage, it must be done at the pins of the tube. With a draw of 15 amps, the voltage drop in even a short run of wire may be significant. Second, the voltmeter must be accurate. Sometimes, voltmeters drift out of spec, especially on the low voltage A/C scales. Make sure you use a meter you trust. Finally, give some thought to how labile your line voltage is when determining the set-point for your filament voltage. Mine can vary quite a lot, and on hot summer afternoons can be 10-12% below what it is on cold winter nights. You do not want your filament voltage to ever drop to the point where it becomes the impediment to producing full power out, as this can cause more damage to the tube than running it at an excessive voltage.
A minor question arose at this point: The tube's cathode is fed via a center-tap in the filament transformer, so that the RF current's path to the tube's filament is electrically balanced. Was it necessary to distribute any added resistance evenly between the two sides of the filament to maintain this balance? Although an unbalanced feed can introduce hum on the amplified signal, I concluded that for the small value of resistance required, the effect would be inconsequential, and thus I ignored it, choosing to add all the resistance to one side for simplicity.
In the end, I decided to drop my filament voltage by about .5V. At 15 amps, this amounts to adding about .0333 ohms of resistance to the circuit in a form that is capable of dissipating 7.5 watts. Some folks achieve this by running fine Teflon wire from the transformer to the tube socket. For convenience, I chose to use resistors. I bought three .1 ohm 5W "sand" resistors, which I wired together in parallel and placed in the circuit at the point where the filament supply attaches to the choke on the tuned input board. This was easy to get to and did not require any disassembly. You can see the results in the photo above.
With the mod in place, the filament voltage fell to almost precisely the 5.0V that the specs call for with this tube and I have not noticed any detrimental effects from this mod, including hum, to date.
Addendum: The October 2011 issue of QST carried an article by Charles Rankin, WA2HMM, entitled The Care and Feeding of a 3-500ZG Amplifier, in which he uses a coil of standard THHN house wiring instead of resistors to make a similar modification to his AL-80B. The publication of this article lead Tom Rauch, Jr, W8JI, Ameritron's principal engineer for this amp, to respond with his own web article presenting his side of the story. Tom makes a strong case for not attempting to minimize or alter the 3-500z filament voltage and raises several points I had not considered, although in my case the intent was not to minimize the filament voltage, but to merely bring it down into the range the Eimac specs recommend. It appears the selection of the appropriate 3-500z filament voltage in an amateur radio linear amp is no simple task, and one Tom and Ameritron spent more than a little time engineering in the AL-80B. I'm still happy with my decision to lower the filament voltage on my amp, but I encourage everyone considering such a mod to read and digest Tom's assessment before making a final decision.
The AL-80B comes with a built-in softstart circuit. This circuit uses a series resistor to limit the inrush current to the tube's filament and to the HV capacitor bank during the first few AC cycles after the amp is powered on. Once the voltages come up to about 1/2 their design values, a relay shorts out the resistor, permitting the power-up sequence to complete and the amp to operate normally.
Ameritron chose to use a 10 ohm, 10 watt, resistor in the AL-80B's softstart circuit. While this may be barely adequate for 120V operation, I judge it inadequate for 240V service. Depending on where in the AC power cycle the switch contact closure occurred, the amp sometimes turned on with a loud "thump" and the shorting relay engaged almost instantaneously.
I opted to increase the resistance to 30 ohms, which I think allows for smoother, more consistent, power-ups than the design default. I did this by lifting one end of the existing 10 ohm resistor from the power supply board and adding another two 10 ohm resistors in series as shown to the right (click to enlarge). Just be careful to dress the leads on the added resistors well away from the edge of the board so that they won't inadvertently short against the cabinet when it is reinstalled.
Another area where there seems to be as many opinions as there are hams is whether or not to use ALC with an amp. ALC is a neat concept. It theoretically allows the amp to feed back to the rig the fact that it is being overdriven so that splatter and even damage may be averted. In practice, though, it suffers from hysteresis effects and sometimes induces other problems, like audio pumping. It is by no means a panacea and many hams believe that, as a cure, it is worse than the disease. However, I've chosen to use ALC, at least in some circumstances, and I'd like to take a moment to explain why.
By default, the AL-80B has none of the protection circuits normally found in higher end amps that knock them offline in the event of a fault. The 3-500Z is a robust tube, capable of withstanding a lot of abuse, and it is a relatively cheap tube as well. Thus, automatic fault protection was no doubt deemed not worth the cost, although grid fault protection can be purchased via an optional add-on board. The amp derives its ALC voltage from a measure of its tube's grid current. The acceptable level of grid current is selectable via a potentiometer on the front of the amp. When the set level is reached, the amp begins to assert an ALC voltage, and if ALC is used, this should cause the exciter to begin to throttle back its power.
Now suppose a fault develops in the antenna, feedline, or the amp itself. The net result is that the resonant pi circuit is going to see a huge shift in its loading. It will become either over- or under-loaded, and most likely severely so. If it becomes over-loaded, the net effect will be to reduce the tube's efficiency. The detrimental effect on the tube in this mode will be high plate dissipation, and while this can certainly be hard on a tube, the plate of the 3-500Z is quite tough and it is unlikely to be damaged if the fault is caught within a few seconds. On the other hand, if the tube becomes severely under-loaded, the voltages in the pi circuit will become quite high, possibly leading to internal arcing (if arcing is not the root cause of the fault to begin with). This will be combined with very high grid current. The grid of the 3-500Z is a much more delicate structure than the plate and can easily be damaged by such a fault before the operator has the time to react. However, with ALC feedback, the amp can make the exciter fold back almost instantaneously, possibly saving the tube. All this is by no means a given, but I figure why not give the amp a fighting chance to protect itself.
I've found another advantage of using ALC with one of my rigs, an old Kenwood TS-430S: This rig does not have a PO control that works for voice modes. The only way to control power on SSB is via the Mic Gain control -- an exceedingly poor substitute for the real thing -- and there is no PO control at all for FM. However, with ALC connected, I can use the "ALC Set" control on the amp's front panel as a surrogate external PO control. This works splendidly on FM, less so on SSB, but it's certainly better than nothing.
I have several HF rigs in my hamshack, and I enjoy using each of them from time to time. The rigs are connected through a coax switch which selects which rig is active. However, with the addition of the AL-80B to the hamshack, I found myself constantly fumbling behind the amp to identify and swap the proper RCA connectors which carry the keying and ALC signals to and from the different rigs. It soon became obvious that I needed another switch to control which rig controlled the amp. But, what sort of switch would be best suited for this task? I purchased a cheap passive 4-port stereo audio switch from Radio Shack with RCA jacks for the audio-in and audio-out connections. I dedicate the Red (right channel) side of the switch to the keying lines and the White (left channel) side to ALC. Now, selecting the active transceiver consists of flipping the coax switch to the desired position and pressing the corresponding button on the "audio" panel to connect the correct keying and ALC lines to the amp. If you try this in your environment, make sure the audio switch is passive, not amplified — the selection must be made mechanically, not electronically. Also verify the unselected ports are not shorted by the switch if you use ALC.
One final comment before I close: There are three closely spaced RCA jacks on the back of the AL-80B: The amp's keying relay, the ALC jack, and a jack marked "12V DC" which makes the amp's 12V supply available to power low power external devices. I've never used this jack, except to measure the voltage on the amp's "12V" bus, but to me it seems like an accident waiting to happen. It is altogether too easy to envision a scenario where one is fumbling to change cords and accidentally plugs the rig's keying or ALC line into the 12V jack by mistake, probably with disastrous results. For this reason, I have fabricated a special "dead" RCA plug fob which I keep plugged into this jack, so that nothing else might find itself connected to it by mistake.
The AL-80B is a real workhorse. It's a relatively inexpensive and tough amp which will get you well within half an S-unit of maximum legal power, at least in the US. With the proper care and feeding, it should provide years of faithful service. I've enjoyed using my amp on a regular basis since I acquired it in 2009, and have come to appreciate it even more after delving into and tinkering with some of its idiosyncrasies.