I'm starting simple, with an old style amplifier of an earlier generation: the Drake L7. This amp is capable of over 1000 watts CW and several hundred watts more on SSB peaks (PEP). It is a grounded grid class AB (linear) amplifier utilizing a pair of 3-500Z tubes with a combined plate dissipation of 1000 watts. Initial checks show the power and electrode currents are within spec and power output is as it should be. With a stiffer power supply it could do more but it already reaches the legal limit in Canada.
An amplifier like this is not entirely straight-forward to install and use. There are many considerations to get the best from it during contests. I have not decided how far to pursue these objectives with this amplifier and may instead make this a temporary resident in my shack. For my present purposes it is acceptable. A more modern amplifier is in my future.
The limited use it has seen in the short time it's been here has been positive, both in operation and on air results. Because it has been many years since I've run QRO there are things to relearn. There is no rush and I am not easily seduced by transmit power as my ongoing use of QRP should impress. For me QRO is about entering contest categories where I will compete with the big guns to discover how much my antennas (and the operator!) can accomplish against the best. Successful QRO contesting requires a different style of operating to be competitive.
In this article I'll briefly cover what I believe are the most important aspects of integrating an amplifier into a contest station. There is more to it than finding room on the shack desk and plugging it in.
Electrical service
An amplifier that puts out 1000 watts of RF consumes substantially more than this. Going by amplifier efficiency alone the heat generated is ~800 watts. For amplifiers utilizing vacuum tubes count on ~150 watts for the filament(s). Since power supplies are not 100% efficient there is at least another 100 watts of heat generated. The rest of the circuitry consumes comparatively little power and can be ignored for this calculation.
 |
| My favourite electrician: Geoff VE3KID |
The total power consumption is ~2200 watts. That's approximately 9 amps for a 240 VAC mains branch circuit. Peak consumption is higher on SSB but the amperage impact depends on power supply design.
Let's assume that we need 12 amps at 240 VAC. The branch circuit should be capable of 20 amps to allow a safety margin and to avoid inadvertently tripping the circuit breaker. I put two branch circuits into the shack. Receptacles are on the floor behind the operating desk to avoid fussing with fishing wire and punching holes in the wall.
My friend Geoff VE3KID, a retired electrician, did the hookup and installation of a 50 amp sub-panel and branch circuits for the amplifiers. He previously installed the service into my garage and workshop (shown in the picture). Running cables and putting in boxes was my job, which he inspected before doing the panel work and hookup. Geoff is also the fellow who installed the circuit for my 30S1 way back in 1985 when we were both much younger.
Although many hams are comfortable working with lethal voltage and current, hiring an electrician is strongly recommended. Where required have your utility inspect and approve the work. Aside from the immediate risk there is potential for future electrical failure or fire and a declined insurance claim.
Noise
Here I am referring to audible noise, not RF, although that too can be a problem (see below). Audible noise sources include fans, open-frame relays and power supply hum. Some can be managed by positioning the equipment. In other cases the amplifier may need modification.
L7 fan noise is modest. The stock L7 antenna relay is very loud and the power supply (a separate unit) hums when in SSB mode (high setting of the plate voltage). Neither is unusual with "classic" equipment of this vintage. Both can be cured with some work. I have not decided what I will do or if I will do anything. Since the amplifier may not be here for the long term I may let it be.
Headphones that totally enclose the ears reduce the noise to a barely acceptable level. This was never a problem with QRP! It takes some getting used to, or in my case re-acclimatizing myself to what I experienced many years ago when I had a Collins 30S1. It, too, was quite loud.
RFI
Since urban and suburban properties are small most hams have their antennas close to the house. The antennas may be directly above the house and shack. Should you have RFI problems with low power it will only get worse with high power. A previously RFI free shack can easily sprout many problems when you install an amplifier. Be prepared for it.
Your computer and peripherals are especially susceptible, exhibiting faults that can be difficult to identify and isolate. Keep a supply of RFI suppressing toroids and use then on unshielded (and even shielded) Ethernet and USB cables, power cords, video cables and so forth. RFI in appliances outside the shack can cause greater grief since your family will exert pressure on you to fix it or QRT.
RFI problems in the house and with neighbours have happily diminished over the years due to changes in technology. The ones that do appear can be more difficult to deal with since they involve digital technology. Proceed with caution.
My major RFI fear is with respect to SO2R and multi-op contesting. Until I have band pass filters and automated band switching the impact of 1000 watts on an adjacent receiver is not to be contemplated. Until then my QRO operating will be strictly SO1R.
Heat
An idling tube amplifier dissipates in excess of 200 watts. Waste heat increases to well over 1000 watts during operation. Although solid state amplifiers have much lower idling power because they have no filaments their operating heat dissipation is approximately the same.
Heat fills the shack and can make it uncomfortable in warm weather. An open window or air conditioning can remove the heat. In winter the shack can be made cozier by the additional warmth, and can also make you drowsy. These problems are exacerbated during contests since high power stations tend to run all the time, resulting in more on time for the amplifier(s).
If you live in a rural area like I do the cost of the electricity can be a burden. Less so in winter when the heat removes load from the furnace, even if electricity as fuel is typically more expensive than what is used by the furnace: mostly fossil fuels in Canada. My house has a ground source heat pump that runs on electricity but with a 5:1 ratio of heat output versus input. Space heaters and amplifiers produce heat at a 1:1 ratio.
Placement
Amplifiers are large and heavy. This can make them difficult to place on the operating desk. As you can see in the above picture I put the L7 on top of the FTdx5000MP. I can do this because the transceiver is structurally robust and the power supply is in a separate enclosure under the desk. A thick cable interconnects them, including a heavily insulated lead for the 3000 VDC plate supply.
The less weighty RF deck of the L7 is an easy reach for tuning and monitoring and has lots of space around the cabinet for forced air ventilation. For smaller, more fragile rigs the amplifier is best placed on the desk outboard of the rig. There is more reaching involved but then the rig is typically narrower (e.g. K3) or entirely absent (e.g. FlexRadio).
Broadband amplifiers with automated switching can be hidden away from the operating position, taking up none of the valuable desktop. You might notice this in pictures of big gun stations where there is no more than LCD displays, keyboards, VFO knobs and paddles. That's very nice but also very expensive, at least at present.
Tubes
Have you checked the prices of high power RF tubes lately? They're high and getting higher. One reason I chose an amplifier with glass envelope tubes is that they are less dear than their metal-ceramic cousins. However they're larger, requiring a larger cabinet, and more fragile.
When run properly both styles of tube can last for many years, but one mistake can be costly. Tetrodes in particular are very sensitive to abuse since the screen grid is typically unable to dissipate any heat at all. That's unfortunate since tetrodes typically have lower distortion. Over-driving grounded grid triodes is also a recipe for early failure. Too many hams push their amplifiers hard, reducing lifetime and polluting the bands with distortion products. Many learn their lesson the hard way when they need to replace an 8877.
Suffice to say it is important to operate amplifiers with care, retuning as necessary when changing frequency, band or antenna. Diligence can ensure that you'll never have to replace those expensive tubes. Solid state amplifiers are no panacea since they have other concerns, such as intolerance of mismatch and, in many cases, greater distortion products. On the positive side these newer solid state amplifiers have safety features to protect you from faults and operator error.
Ceramic tubes such as 8877, 4CX1500B and many others require a 2 to 3 minutes warm up before they can used. This can be aggravating for DXers. Don't bypass the safety feature that enforces the delay or you'll face a drastically reduced tube lifetime.
Tuning
Speaking of tuning, it is worth mentioning a few of the challenges involved. Most hams are accustomed to broadband transceivers that handle frequency, band and antenna changes without operator intervention, including an ATU with memory. That is rarely the case with a kilowatt amplifier. The impact on contesters such as myself can be significant.
Tube amplifiers in particular almost always require manual tuning of the output tuned circuit. The output impedance of a tube or FET amplifier is higher than 50 Ω, while bipolar transistor amplifiers have a lower impedance. In all cases impedance transformation is required. Broadband transformers can be effective for most solid state amplifiers, but these can be large and expensive at the kilowatt level, and are usually impractical for tube amplifiers due to the large impedance ratio.
Broadband transformers also require that antennas be close to 50 Ω if they are to perform well. Manual tuning of tube amplifiers can match a wider impedance range at the price of operator effort. Having your antennas exhibit low SWR makes operating any amplifier much easier and safer.
Contesters and other active hams will happily deface their amplifiers with stick on labels for penning marks for various antennas and frequencies so that tuning can be accomplished quickly. Fewer contest QSOs are lost and DXers are less likely to miss out on rare DXpeditions. It can also help to avoid tuning errors. If you do use labels try to find those that are easy to remove and do not leave a permanent mark on the panel, otherwise you will have an eyesore with a lower resale value.
We can't forget the amplifier input which must also be tuned since the input port is rarely 50 Ω. Untuned amplifier inputs are found on lower end equipment and may require use of an ATU on the transceiver. Even with tuned inputs it may be necessary to use an ATU on the WARC bands, depending on amplifier vintage and quality. Since my amplifier has tuned inputs I have to turn off the rig ATU when I go via the amplifier to an antenna with a poor SWR. That's one more detail to worry about until all my antennas are good 50 Ω matches.
You may notice that if the rig and amplifier could be integrated that the transmitter output network and amplifier input network could be collapsed into one. Since that it rarely the case, and indeed the rig is often operated alone, an impedance of 50 Ω between the units is sensible.
Another thing: did you ever notice that many modern rigs don't have a button to transmit a carrier for adjusting an external tuner or amplifier? Something needs to be rigged up to implement this feature. On my FTdx5000 there is a jack on the rear panel (Tx Req) that you supposedly ground to do this, but mine doesn't doing anything. I'm working on it.
Sequencing
Although not often thought about in the typical ham's station, once you add an amplifier there may be a need to deal with sequencing. For our purposes the only signal of significance is when you go key down. Problems can arise depending on where the signal comes from: keyer, computer, transceiver or other peripheral controller, and the speed of the switching circuitry in the rig and amplifier.
Most high power amplifiers use relays at the input and output to switch the amp in and out of the circuit; during receive the amp is bypassed. The output relay tends to be of the large and slow type to handle the power. If RF appears at the input port before the relays fully engage hot switching can occur. This can damage the relays, cause arcing in the amp and truncate the first transmitted symbol. This is mostly a problem on CW, on that first dit or dah, and on digital modes, but typically not on SSB unless VOX is used.
When using a "smart" keyer such as a Winkeyer or contest software the sequencing problem can be mitigated with some experimentation. When playing memories -- these are the majority of contest transmissions -- delay can be added between assertion of PTT (Tx enable) and the first message symbol. On SSB it is even easier since it is unlikely that the message begins at full volume. The trouble is with VOX on SSB but especially with CW sent with paddles. In the latter case there is no way to advance PTT assertion, and delaying symbol transmission will adversely affect the operator since audible feedback will be out of sync with the fingers.
In initial testing of the L7 the serial assertion of PTT from Winkeyer to rig to amplifier does not affect the initial CW symbol. Monitor the bands during a contest and you will find instances where the first symbol is short or entirely lost. The problem gets worse as CW speed increases. If you have any instances of arcing in the amp (relays, capacitors, etc.) stop immediately and deal with it.
There are two alternatives to serial chaining of PTT. One is to key the rig and amp in parallel from the computer or Winkeyer. Before trying this ensure that the voltage and current are within the capability of the Winkeyer or computer circuit you are using. Those big relays in older amplifiers can draw more current than you expect.
The second alternative is to modify amplifiers with electronic switches and fast relays. These are far quieter as well (see 'Noise' above). Vacuum relays at the output port of the amplifier may be required. The best systems are QSK compatible.
What all of this means to me
From the foregoing discussion you may appreciate why I have been slow to include an amplifier in my station. Aside from trying it out on the bands during ordinary DXing I have yet to put it to the test in a contest. I am in no rush since the contest season is winding down with the arrival of spring. Good thing, too, since my antennas are in rough shape.
Soon enough I will try it out in a minor contest. That will give the amp a good workout and will force me to exercise my QRO contesting skills and tactics. It is very different in comparison to low power and QRP contesting. I will also have to practice rapid band changes.
With the arrival of spring and sporadic E season it is a shame that this amplifier does not work on 6 meters. At some point I will need more power on that band for optimum DXing performance. That will shortly be the subject of another article for the blog.
That I now have an amplifier does not mean I will no longer operate low power or QRP. I enjoy all power levels and will continue to be flexible in my operating choices. I prefer to delay QRO contesting until I have made more progress on low band receive antennas. Otherwise I will not be able to copy many of the smaller stations who will call me. I do not want to be an alligator.
My second amplifier, when I get one, will likely be more modern than the L7. I would like one with better operating features, in respect to many of the considerations raised in this article. Broadband tuning would be especially welcome. When a suitable amplifier becomes available I will make my move. I can afford to wait.
No comments:
Post a Comment
All comments are moderated, and should appear within one day of submission.