What you should consider is a "magic tee" or "hybrid splitter".  This is not a job for a simple 3-coil transformer.

The good news: the "magic tee" is no more parts.

Use a search engine on the phrase "magic tee".

AM

Colorado has long been a model state, from which California learned how to "do things right".  Colorado still has more stringent smog requirements on vehicles than in CA.

One of my homes is in Colorado.  When my CO-registered vehicles first start to fail the smog test in Colorado, I bring them to California, pass them through the super-simple smog test, register them, and sell them.  And usually get back about what I paid for them five years earlier, in CO.

AM

Sorry to hear that license class is considered important.

I believe we are all of the same class - "fellow brother and sister hams".

AM

Most dealers will honor their "Dayton" prices if you mail order during the Hamvention weekend.  Bonus - no sales tax.

OTOH, the '857, according to the owner's group on Yahoo, is turning in some very unpleasant reliability experiences.  I'd avoid it until they work out the bugs.

AM

KE4MOB wrote:

"An amp may work fine *and* show no symptoms of trouble, but still need the caps replaced.

"I recently acquired an SB-200 which worked great. Just for kicks, I unsoldered one of the filter caps and looked underneath.

"Sure enough, the cap had leaked and left a black sticky mess on the board. I removed the rest to find half the caps on the PS board had leaked.

"But the amp worked fine.

I would humbly submit to you that leaky capacitors do not qualify in your first sentence "show no symptoms".  Leaking is most definitely a symptom!

How certain are you that the black sticky mess came from the capacitor, rather than being charred resin from the low-grade solder that Heathkit supplied their kits with?  If it was really capacitor guts that had come out, they would have eaten their way through the PC board.  If the PC board was intact, then the mess did not come from the capacitors.

AM

I use a general-coverage HF rig "opened up" for transmit.  Disable both the PA and the driver, and send the pre-driver output through an old step attenuator.

The signal from such a radio is a whole lot cleaner than most cheap sig gens, and as a bonus, you can modulate the signal with any flavor you wish.

AM

If you replace 811As with 572Bs, then you MUST push the amp hard.  Why?  The self-gettering of the 572Bs requires that you get them hot.  If you only run a quad of them at 800 watts out, then you may not be getting them hot enough, and they may begin arcing, although possibly not at 1.7kV.

Remember, these tubes are supposed to operate with the anodes glowing somewhat.

I've not experienced this, but the manufacturers warn about it.

AM

I've done field strength measurements on ground-mounted versus elevated radials.  Elevated radial systems seem to improve the vertical's performance by about 10dB.  Varies with band.

The main function of the radial on an elevated vertical is to provide the "second terminal" for the power to go to.  A resistor has two leads - and so does an antenna.  I did an experiment once to see what happened if I installed four radials per band, but just let them hang downward.  Result - about one dB lower radiation strength than "normal" and SWR rose from about 1.2 to 1.8.  So, if you don't want the "radial cobweb", you can just let them hang downward.  Antenna was a bit more broadband as well, as I believe the coupling to the lossy soil was better.

A vertical is primarily a DX antenna.  You get awesome zero-hop coverage (very local), but the first hop is a LONG distance away - 3000 miles on most HF bands.

Raising the vertical gets you one strong advantage - ground-mounted verticals hear all the RF hash and noise in your neighborhood.  Elevated verticals hear the hash generated above your neighbor's rooftops, which ain't much.

AM

First off, as others have said, the TS-850 is a great radio.  I'll try to get to the point, though.

Point One: numerical specs are only a start.  The TS-850 has great numerical specs, but the AGC on that receiver over-compresses, making it "sound" noisy, even though it's not.  This is also true of the FT-1000 series when receiving strong signals.

Point Two: This is all pointless if you are not a contester or very serious DXer, or are trying to DX on 160 meters.  Fact is, all those IMD/BDR/compression specs only come to the fore when you are trying to hear a VERY weak station on a band that's crowded with very strong signals.  If you're a casual ragchewer, then anything, even old Heathkit boatanchors, will work well.

The disclaimers are now out of the way.

The single most telling specification on any receiver is the close-spaced IMDDR (IMD-limited Dynamic Range).  Here's what happens:  You want to listen to a station at 7.004 MHz, who is extremely weak.  There's a station at 7.003, and a station at 7.002 MHz also, and they are both very strong.  Inside the receiver, something (usually the front-end mixer, but it could be any stage with diodes or transistors in it) creates a very tiny amount of the second harmonic of all of these signals.  One of those signals is 7.003 MHz, the second harmonic of which is 14.006 MHz.  This very same non-linear element in the receiver also works as a mixer, and a very tiny mixing product is created - the 14.006 MHz signal is mixed with the 7.002 MHz signal, and the result is two new signals, one at 21.008 MHz, and one at 7.004 MHz.  We don't care about the 21.008 MHz signal, but that signal at 7.004 MHz is on the very frequency where we're trying to hear the very weak station.  This is "intermodulation" - two strong signals combine in such a way to create a weak signal at a new frequency.  

Now, very few hams have actually heard this - you don't really hear a "spare" new signal right where you don't want it.  But, bear in mind that NOISE is also a signal, and there is noise at 7.002 and 7.003, and noise can self-create IMD in the same mixer/amp stage.  Therefore, a poor dynamic range receiver can seem far noisier, because lightning, atmospheric, and other noise continues to create all sorts of IMD products that make the bands sound worse than they really are.  A receiver with lower IMD products is more linear - IMD is a way of specifying linearity.

"Blocking" is a different way of measuring linearity.  As a general rule, if one receiver is 1dB better in BDR, then it will also be 1dB better in IMDDR, but that is not always true.  The ARRL has not always measured IMDDR, so you may have to go by IMDDR.

Wide-spaced BDR or IMDDR is less important - they can use filters to improve this spec, yet you'll be left with a poor receiver when all "suspect" signals are inside the front-end filter's bandwidth.

As I implied before, unless you are really pushing the envelope of needing a high-performance station that is outstanding when pushed hard, any receiver whose specs measure within a few dB of others just won't be much different.  In fact, it's quite difficult to measure the IMD levels within an accuracy of better than about 2-3 dB anyway.

I'm mainly a CW op and like many, I don't like Kenwoods.  Kenwood has really staked a claim on being the best SSB radio.  The receiver AGC is optimized for SSB and on CW, the excess compression of it can make you think there's more noise than there really is.  The best receiver I've listened to (for long periods) for CW is the K2 - they have done something magic with the receiver AGC.  The Omnis are right there with the K2.  I personally like the FT-1000D because the audio peaking filter can help my very old ears and brain handle a CW signal in noise a bit better than other radios.  Notice that all the radios I mention have actual IMD specs that are within a few dB of each other.  

What's in my shack?  Surprising, actually - a Drake B-line for nostalgia is the "ragchew radio".  Very poor IF filtering means they're poor in pile-ups, but very easy on the ears for ragchews.  Pretty good receiver IMDDR, though.  A Kenwood TS-950 - same front-end as the TS-850 - this is the main DXing and contesting radio.  I'd prefer an FT-1000D, but my budget won't support it.  An Icom IC-706MkIIG.  I acquired this as a 2nd radio for SO2R contesting, but its front end is so awful that it is functionally useless in contesting.  Look up the IMDDR specs on that one, pretty sad. Actually, I have several homes in several states, and the '950 and '706 go from one to another.  Were I wealthier, I'd have a K2 in each location, and drag an FT-1000D with me when I travelled from one to another.

AM

Hey, I have a 638 also.  Came with an old Heathkit.  I use it to drive an Icom IC-745.  Icoms have a DC bias voltage fed to the mic from the radio.  I used that DC bias voltage to power an op amp circuit.  Just a "voltage follower", you connect the output of the op amp to the negative input and get 0dB gain.  Then, a pair of  220k ohm resistors to bias the + input to half the supply voltage, and a capacitor to couple to the microphone input.  A capacitor goes from op-amp output to the radio's mic input, and a 470 ohm resistor goes from the radio's mic input to the op-amp + supply (and top of the two 220k ohm resistors for biasing the + input).  Ground the - supply pin of the op amp.  I forget what op amp I picked - it was one that was designed to operate with Vss down to 3V, and single-supply also.  Radio Shack had it in stock.

GL

AM

You have to be driving the grids pretty hard or you'll end up getting this result.  With the LOAD cap fully meshed, the tuning condition should be max gain, but lowest saturated output power.  If you're driving the tubes very gently, then max power will be at fully meshed LOAD cap - but it won't be much power.  Drive the tubes harder, and then max power should occur with the LOAD cap more opened up - and it'll be higher power.

AM

I worked them on 160, 80, 40 and 30 meters.  On 30 meters, RX/TX antenna was a random length wire dipole at 30 feet elevation on RTTY.  On 160,80,40 it was CW.  Almost all cases, the opening was, as k6yr said, a morning grayline opening.  The exception was 160 meters, which was quite late at night.  I did this operation from northern Oklahoma.  On 160, I use a top-loaded vertical, with radials elevated about 15 feet.  Vertical radiator is only 35 feet high, "natural" resonance with top hat is 2.7 MHz, so I tune to get power into it.  80/40 were using a commercial vertical, with a mix of elevated and ground-mounted radials of random length.  RX antenna in some cases was the dipole used for TX on 30 meters.  My power level is 100W on 160 and 30, 1.5kW on 80/40.  It took a lot of calls on 160, a few calls on 80 and one call on 40/30...but then I'm a good op, and know how to pick the right freq and timing.

I never heard them on 20 meters and higher.

In no case was the signal from FT5XO really above the noise - in other words, the S-meter did not rise above the noise reading when I heard FT5XO.  I could have never understood them on SSB.  CW and digital were the only modes that would have been useful from my QTH, with mediocre antennas.

One strong suggestion: if you're going to work DX from a less-than-Big-Pistol station, you should definitely have a sound card-based digital mod/demod setup running.  When the DXped goes to digital modes, you have a LOT less competition, and you can copy them in far worse conditions than on SSB or CW.

W6EL prop is only one indicator.  Go to a DX cluster, and do a search for FT5XO.  Look at what time the stations from your region are working the station.  That's a better clue than propagation predictions.

Likely at this point, they're going to focus on extreme cases only - 160 meters, and digital modes.  Only one day left.

AM

VE7ALQ...ripple in a class AB amplifier only causes modulation if you are driving the amplifier so hard that you're clipping it.  That's how plate modulated rigs work.  They drive the PA extremely hard, well past the point of saturation, then they vary the power supply voltage, via a modulation transformer, to get AM.  If you use less driving power, so that the amplifier isn't saturating, then the modulation doesn't work.

At the plate of the tube, a given RF input power "demands" a min and max instantaneous plate voltage.  This varies, of course, at the rate of the RF.  For the 572B type tube, the minimum you can get to is roughly 200 volts.  If you try to drive harder, so that the plate voltage (instantaneous - you have to look on a scope with an RF-capable high voltage probe to see it - your plate voltage meter won't read this) goes even lower, then the thing saturates, clips, causes IMD or keyclicks.

So, to keep the amp clean, you back off the drive a little.  Let's say, then that at your drive level, you're only expecting the 572B anodes to drop down to 500V at the RF minimum.  You'ge got 300 volts of "margin" available.  The HV power supply could drop by 300V, and you won't get any reduction in output power.

Wanna see this for yourself?

1) Load up your amp
2) Hang a bit of wire on the input to your oscilloscope
3) Key the amp, with the rig sending a steady carrier
4) Use the rig's drive control to increase/decrease the power delivered to the amplifier and watch the 'scope.

You'll notice that you can see ripple when the drive is cranked up hard, but not otherwise.

AM

Before putting in those new caps, make sure you need them.  The OEM capacitors in these amps had tolerances sometimes of +/- 50%.  If the capacitor is simply lower in value than when it was new, the amp may work fine.  I have had SB-220s and SB-200s where the capacitors were as much as 60% low in value, and the amps worked great.  The problem that low-value caps causes is ripple.   But in a Class AB linear amplifier the ripple does NOT modulate the output, unless you are driving the amplifier all the way to saturation.  If you're doing that, then you're causing splatter or key clicks anyway.

I've repaired dozens of these amplifiers, and have not encountered a single one that actually needed new capacitors.

AM

The pi-L output network in your amplifier can match the SWRs that you mention just fine.  As WB2WIK said, if the amp produces power, then you have little to worry about.

If the SWR is high, you may be operating the tank circuit with more circulating current and/or peak voltages.  On 10 and 15 meters, most amplifiers are operating with higher tank Q than on lower bands, therefore the circulating current is higher to begin with.  This CAN lead to overheated and burned bandswitch contacts for 10 and 15 meters, but the deterioration occurs over many years.  Fortunately, 10 and 15 meters are, first of all, bands where amplifiers are simply not needed most of the time.  Secondly, we tend to use tuned antennas on these bands (Yagis, quads, etc).  As WIK pointed out, on 80 meters, no antenna is low SWR across the band, but on 80 meters, the amps don't run excessive Q to begin with, therefore it's not an issue.

I have an old SB-200 by Heathkit, and I'm a flagrant violator of the wive's tales of SWR.  I did lose a 10 meter bandswitch contact once, but I was running in excess of 800 watts on RTTY on 10 meters into about an 8:1 VSWR - and had been doing so for over five years in contests.

Short answer - unless you're going to be contesting into an 8:1 VSWR, don't worry about it.

AM