ZM4G glow logo
ZL2iFB glow logo
Tokyo HyPwr
TS850 mods

The following rough notes concern my beloved Amp Supply Co LK550 HF amplifier.  I bought the amp second hand from a white stick op in the UK during the early 1990s and was my faithful companion for 20 years until Elecraft brought out the KPA500.  The amp has 3 x 3-500z valves which, with sufficient drive (more than 100W I believe), makes it capable of kicking out rather more than the license limits.  I’m not one to push the red line constantly though.  For DXing and contesting, I’m much more concerned about having a clean signal and reliable equipment, which means deliberately under-running stuff.  Normally the KPA500 is sufficient.

Over the years, I’ve made a few repairs and mods to the LK550, usually with help from other knowledgeable hams.  This page is my attempt to give something back.

Warning - lethal high voltages inside

The LK550 HT runs at about 3kV off-load, and dips to about 2.6kV on load.  3kV at up to an amp or more will fry and most likely kill you in an instant, with no second chances.  This rates alongside skydiving, motor racing, coal mining, electric chairs, urban search and rescue, looking down the barrel, and playing with gunpowder and matches in terms of personal danger.  Even with the main switch turned off but the mains lead still in the wall receptacle, the full mains voltage is exposed on the main switch, and that can certainly kill you too.  If you don’t know exactly what you’re doing with high voltages, don’t risk it - go find a competent amateur friend or professional who builds and repairs valve amps instead.  Even if you do know what you’re doing, be extra careful with extra high tension if you value your life.

If you ever need to take the lid off the amp and you have been using it lately, first check that the front panel plate voltage meter is working by turning the meter knob to VOLT.  It should read about 3kV with the power on.  Now turn off the mains supply and unplug it from the wall.  Watch the HT voltage using the front panel meter until it has bled away, ideally to an indicated zero: mine shows over 500 volts after the first minute, and still over 100 volts after two minutes so I recommend waiting AT LEAST three minutes, especially as your amp may behave differently to mine.  Double-check that the correct mains plug is out of the wall and remove the Jones-type HV connector to the HT transformer from the rear of the amp.  Only now is it safe to remove the screws and carefully lift off the lid, and even so there may be sufficient charge left in the PSU capacitors to give you a nasty surprise: once the lid is off, use a “chicken stick” (an insulated wand with a wire tip connected to the chassis) or a good quality insulated screwdriver to short between the chassis and the anode caps.  If you hear a distinct crack, you were a bit too keen to get the lid off but be thankful that you heard anything because at least it means you are probably still alive (so far).  You may have just damaged your smoothing capacitors with that discharge though, so don’t make a habit of it.

LK550 repairs

High voltage power supply capacitors

According to my trusty notebook, the HV PSU electrolytic capacitors in my LK550 let out the smoke with a pop one day in 1994, blowing the mains fuse.  The originals were 7 x 450V 160uF capacitors which I couldn’t source at the time, so my friend Ross GW3NWS advised me to replace them with 400V 220uF “computer grade” units from RS Components.  I added two extra capacitors to the earthy end of the string on a little daughter board to give a bit more headroom for the normal 3kV HT: the new set should be OK up to 3.6kV and in practice has worked just fine ever since.

Be careful about the capacitors you use in this application.  I invested in new “computer grade” electrolytics.  “New old stock” electrolytic capacitors should be carefully re-formed before being subject to high voltages, and they may still fail anyway.  Poor quality capacitors may not have sufficient ripple current ratings and so may overheat and fail prematurely.  Capacitors that burst open make a real mess and stink so, trust me, this is best avoided.

I checked the original equalizing resistors, adding another 100 kOhm 2W resistor across each of the additional electrolytics.  I also checked all 32x 1N5408 diodes in the bridge circuit and they were all OK, despite the blown capacitors.  Guess I was lucky.

Antenna change-over relay

The little Jennings 24V vacuum relay (used to switch the antenna between a bypass line to the amplifier input connector used on receive and the amplifier output on transmit) doesn’t last forever.  The obvious symptom of something amiss is that the rig goes deaf after switching from transmit back to receive.  A sharp thump on the amp box or stamp on the PTT footswitch usually brings the receiver instantly back to life, but within a month or so even those tricks no longer work and it’s definitely time to swap out the failing relay.

The relay is located right behind the SO239 connectors on the rear panel.  It is mounted on the QSK board, with three RF connections on one side and two relay coil connections on the other.

Replacement relays are available from various HV and amp parts suppliers but shop around as prices vary markedly.  Second-hand relays pulled from commercial equipment may be cheaper but probably won’t last as long.  Similar relays are made by other companies such as Siemens.

I made a rubber gasket to reduce the loud click when the relay fires.  It does helps make the changeover quieter but wrapping some sort of sound insulation material around the body of the relay, or buying a relay already encased in silicone rubber, would probably be worthwhile as well.

[See the mods section below for ideas to extend the life of vacuum relays.]

Coupling & decoupling capacitors

In 2008, something in the amp let out the smoke with a loud bang.  Turns out a decoupling capacitor in the PA compartment had blown a chunk of itself out.  The capacitor decouples any RF that gets past the anode choke to ground before it has a chance to get into the HT power supply.  This was an easy fix to make - the only hard part being to find a source of suitable HV capacitors.

A pair of coupling capacitors feed RF from the anodes to the tank compartment.  I changed those at the same time.

LK550 fixed PA caps

Padder capacitors

On the low bands, the air-spaced variable capacitors don’t have enough capacitance to resonate the tank circuit so the bandswitch adds in some additional fixed capacitance. 

  • On 160m, the padder consists of a ‘ladder’ of several paralleled HV ceramic capacitors right beside the band switch (soldered between the two vertical wires on the far left of the whole amp shot below - I had removed the old ones at this stage).  The dark colour and waxy deposit visible on the photo above shows that these were suffering due to age and heat. 
  • On 80m and 40m, the padders are tucked underneath the LF coil to the left of the band switch on a pair of solder tag strips bolted to the cross wall/chassis.

80 and 40m padders

I replaced the padders at the same time as I replaced the blown coupling capacitor, figuring that these were probably to blame for the amp needing more and more load and plate capacitance to load up properly on the low bands. 

On 160m, unfortunately, the replacement padder capacitors caused a strange drifting effect: on key down, the output would climb gradually to the normal level over a few second but would drop back down again between overs.  I guess that it was a thermal effect so replaced the 160m padder capacitors again with a larger HV capacitor from Russia, but that didn’t work properly either.  I’m still looking for the right type of padder capacitors for 160m ...

Output valves (tubes)

I replaced a faulty 3-500z once due to a cathode-grid short caused by mechanical damage while transporting the amp between houses.  I also replaced all three valves with a matched triplet of Chinese ones from Taylor as the output seemed to be declining, and I wanted to have some spare valves on hand.  Changing the valves is easy - just treat them gently, take care to seat them properly in the bases and replace the anode caps properly.  Personally I try to avoid touching the glass with sweaty fingers and leaving fingerprints that will probably go brown later, using pieces or clean rag or tissues to hold them, but that’s not strictly necessary.

LK550 innards without valves and 160m padders

Thoriated tungsten filaments (heaters) in valves such as the 3-500 need to run hot for a few hours to reach full emission.  The suggestion is to turn on the amp after fitting new valves and leave it on standby for a few hours before transmitting (though I can hardly wait!). 

It is also recommended to run the valves hard (so the anode fins glow orange) every so often to heat the zirconium getter compound on the anode and so destroy any gas that may have leaked past the seals.  In my case, this means running key down QRO at least once a year, into a QRO dummy load.  And by the way, it’s best to run any spare 3-500s through this procedure at the same time since leaving them sitting in storage for too long increases the chance of a catastrophic flash-over when they are eventually put into service.  Alternatively, I guess it’s possible to use a variac to reduce the HT, but only if you separate the HT from the heater and LV control wiring which should remain at the normal supply voltage.

Main supply switch

A distinct thud when turning on the amp suggests that there is a sizeable surge as the HT capacitors charge up and the cold tube filaments suck up the juice.  Turning off the amp doesn’t make the same thud but there’s a good chance there is a pulse of back-EMF from the transformers, and quite possibly an arc in the mains switch.  That probably explains why I have had to replace the main switch at least two or three times already.  For the past few years, I have been using a high current mains switch (32A at 250v AC) intended for electric oven supplies, which meant resizing the panel cutout with a file but was otherwise quite simple to fit.  That switch has lasted longer than previous ones.

[A more permanent solution would be a step-start circuit - see the mods below.]

Anode choke

LK550 burnt plate choke During the past year or so, the anode choke on my LK550 has gone dark about a third up from the lower HT supply end, and the turns have become even more uneven there (I’d noticed some loose turns before).  Lately, the amplifier has started taking off occasionally , the key symptoms being no output, off-the-scale negative grid current, about 400mA of plate current and a burning smell, all of which continues even if I stop transmitting and release the PTT.  I’m pretty sure this is due to progressive damage to the anode choke , itself probably caused by me operating the amp on 30m using the 20m position (the amp’s design pre -dates ham access to the WARC bands).  What I presume was a choke resonance near 30m made the amp perform quite inefficiently on that band, needing about 50W input to squeeze 2-300W out instead of the normal 25W in for 500W out on the other bands.  The bunching up of turns is probably a sign of overheating, making the thin enamelled wire sag.

Rather than just rewind the choke and test it for unwanted resonances using a GDO, in my junk box I found a suitably meaty HF choke on a larger ceramic former than the original.  Using a coupling loop on my MFJ259 with the choke shorted out, it appears not to resonate in any ham band but there is a resonance just above 15m.  It is a tight squeeze in the PA compartment but hopefully the air flow is still sufficient to cool both the rear 3-500’s.  It’s a good thing I’m not running the amp flat out.

LK550 new plate choke


In case the surplus choke doesn’t work out, I’ve also bought a choke from Ameritron (now part of MFJ) that is rated for 4kV at 1.5A (6kW input).  Not bad for US$20 + P&P.

LK550 modifications

Extending the useful life of the antenna change over relay

Someone (I forget who) suggested that the reason the main antenna change-over vacuum relay fails periodically is that, over time, the actuator coil becomes permanently magnetized, just enough to prevent the normally closed relay contacts closing completely when the relay is powered off.  They suggested changing the polarity of the coil supply, which means swapping over the connections to the coil.  To avoid having to open the box and resolder the connections every time, I’ve fitted a small DPDT switch to the rear panel of the amp, wiring the coil connections to the middle pins and the supply leads to the outer pins on one side, with a pair of patch wires connecting them to the outer pins on the other side but swapped over.  I’ve marked the switch positions on rear panel “EVEN” and “ODD”, intending to flip the switch according to whether it is an even or odd month of the year.  I should be able to tell you if this works in a year or so!

[I’ve also heard it suggested that running a bit of DC through the relay contacts on receive will help remove any corrosion on the contacts, but I’m not entirely convinced: apart from the need to rig up a DC supply and fit coupling capacitors in the RF path, the whole idea of using a VACUUM relay is that there should be no reactive gases in there to cause corrosion.  However, it might be a way of rejuvenating some of the old relays that I have removed and kept, since they are quite expensive to keep replacing.]

[I keep meaning to confirm that the relay is not hot-switching on QSK as that will definitely shorten its life.  All I need is some time to rig up the oscilloscope, find a way to sample some RF on one beam (perhaps there’s a suitable contact in my external wattmeter?) and connect the other scope beam to the keyer.  Meanwhile, I just avoid using QSK most of the time, and make sure the K3 allows a few milliseconds’ grace between PTT and transmit.]

Quietening the antenna change over relay

Like most amplifiers, the LK550’s stock antenna change over relay makes a noticeable and annoying click every time it operates due to the mechanical movement.  Inserting a foam rubber gasket (homebrewed to fit) between the relay body and the flange has quietened it down significantly, so much so that I sometimes find myself checking the standby/XMIT switch or power meter to confirm the amp is actually working!

Fan speed control

NB  I haven’t made this mod yet - it’s just an idea at this stage.

The LK550 has a two-speed fan, manually selected from a front panel switch.  I always try to remember to flick it to the high position when transmitting but the noise means I prefer to leave it on the low position when idling.  I’m a bit worried about transmitting with it on the low position, even though at around 500 watts out it will probably be just fine, so I’m planning to dispense with the separate fan switch but use the transmit-standby switch to turn the fan to high when set to transmit.

Other suggestions include:

  • Automating the fan speed switch using a temperature sensor in the PA compartment - perhaps a mechanical thermostat as an electronic one may not work well in the high RF environment.
  • Increasing the low speed fan setting until the noise remains barely acceptable but I’d be comfortable transmitting at the 500W level.
  • Using an external electronic thermometer to monitor the PA exhaust flow, ideally with an alarm if it climbs above, say, 50C.
  • Coupling the PTT control line to the fan speed switch with a relay, such that the fan set to high whenever the PTT line is earthed.
  • Using a temperature-sensing proportional fan speed controller (perhaps one from a PC) with sufficient decoupling and shielding to ensure that it works even with lots of RF floating around.
  • Replacing the 4 muffin fans with a set of temperature controlled computer fans (though I would have to match the CFM ratings of the originals, and ideally the temperature sensors would have to be placed on the far side of the valves).
  • Replacing the 4 muffin fans with a “proper” squirrel cage blower on the rear panel (probably too much mechanical work for me, and those blowers are noisy too).

Hawke’s Bay
North Island
New Zealand

39o 39’ South x 176o 37’ East

Locator RF80HL

260m ASL


CQ zone 32

A1 Ops

DX CoC logo new 125 on black
Clublog logo 125