Entries from May 2014 ↓

240 or 120

It will work better on 240. Maybe if you are using a current starved 120 outlet that may also be one of several on the same breaker. Then , yes, it will run better on 240.

It will also run better on a dedicated 20 amp 120 circuit. 240 x 10 = 2400 watts. 120 x 20 = 2400 watts. As long as your 2kw linear is getting all the input power it needs, it does not care if that power is coming from a 120 or a 240 volt circuit. You do need 20 amp breakers on the 120 circuit where the 240 circuit can do with 10 amp breakers.

QSK Second thoughts

The ability to copy code in between the dits and dots I send has never appealed to me. I guess maybe I do have trouble walking and chewing gum at the same time. So QSK operation has never been a high priority with me. Actually it did not even make the list.

I did set up a Kenwood TL-922 with QSK mainly because I wanted to ensure proper T/R sequencing.

Right now I have the TL-922 working along side an SB-200. The SB-200 does not have QSK and it lets me know when it goes to transmit by very loudly clanking the T/R relay.

I much prefer the strong silent performance of the TL-922 and as soon as I buy another vacuum relay the SB-200 is going to be QSKed too.


I recently acquired an old Heathkit SB-200. Did it work? I don’t know. I did not power it up until I replaced old parts and added new ones.

The first old parts that were replaced were all the diodes, resistors, and capacitors in the HV power supply/filter section. The old parts were carefully unsoldered from the circuit board and six new 220mf/450v caps, four 1N5408 diodes, and six 100k 3 watt resistors were installed. New holes were drilled to mount the new electrolytics. The mounting holes for the diodes had to be enlarged.

The HV lead to the plate choke was disconnected in preparation for testing and tube conditioning.

Before power was finally applied a soft start circuit was installed in the primary power wiring. I used a very simple 110vac DPDT relay and two 24 ohm 7 watt resistors, one in each leg of the primary power leads. The soft start relay coil was wired across (in parallel with) the 110vac blower motor coil. This ensures proper operation of the soft start for both 110 volt and 220 volt operation.

Granted that this amp worked for a long time without soft start so we might want to question the need for it. On power up with the soft start the soft start relay begins to rattle for a few seconds before the voltage across its coil becomes high enough to close the contacts and bypass the current limiting resistors. The start up current is limited to about 2 amps. The relay rattling and duration indicates that much more than 2 amps would be drawn on start up without the soft start circuit. Yes, the amp would still work without soft start but I believe the soft start will make the components last longer. My primary concern is the power transformer.

Now a word of caution regarding operation with an amp using a soft start circuit. In particular the soft start circuit implemented as described in this amp. In general it is good practice to ensure that the amp is plugged into a primary power service that can provide the required current. It does not have to be a 220 volt circuit. A 110 volt circuit will do just fine as long as it can deliver the 15 to 20 amps required. After all we are only talking about running a KW amp. Your hair dryer pulls more than a KW and it runs fine off 110 volts. If the primary outlet cannot deliver the required current, say because of a long extension cord or some other stupid reason, the 110 voltage to the amp will begin to drop. When that happens the soft start relay will start chattering as it begins to drop out.

If that relay drops out, even intermittently, the current limiting resistors in the primary power legs will be forced into the circuit and they will get very hot. If this condition is allowed to continue those resistors will smoke and stink up your operating position.

So, at the first sign of relay chatter, pull the plug. Turn off the amp!

After the soft start was installed, the amp was powered up (with the HV lead disconnected) and the HV was checked on the meter and noted to read 2250 volts. That is within a few volts of the calculated output voltage for a doubler power supply ( 800 X 1.4 X 2).

A single strand of hookup wire was soldered between two wire posts . The strand is in series with the HV supply. This is a homemade fuse wire designed to protect the HV supply in the event of tube failure.

The T/R relay in this amp requires that the exciter be able to switch -135 volts to ground. Exciters that provide relay contact amp switching can handle this requirement. Newer exciters are using solid state amp switching that may not be able to handle -135 volts. There are lots of solutions to this problem. Some of the more sophisticated solutions provide high speed QSK operation for serious CW operation. Others provide interface boxes that require external power supplies and additional cables.

I do have a TS-922 amp that is equipped with QSK. Operating with the TS-922 is very quiet. No T/R relay noise like with the SB-200 but a good QSK circuit requires at least one vacuum relay and costs around $100 for parts alone. Unless you are an avid CW contester, such an investment is hard to justify.

My operations do not require QSK even when working CW and I do not like interface boxes and external power supplies cluttering up my operating position. An amplifier needs to be able to work with all current and future exciters to be useful.

To make the SB-200 ultimately useful we need to add low voltage T/R switching. I used a 5 vdc relay but there is no 5 vdc power in the amp so I had to build a dc power supply too.

The easiest way to get the dc is to use the 6.3 volt filament transformer winding. A convenient connection to 6.3 volts can be had by using the 6.3 volt power to the pilot light used to illuminate the meter.

In my case I also wanted to replace the incandescent bulb with bright white LEDs. This required removing the front panel to get to the meter and wiring. If you are not going to use LEDs, you might be able to get to the connections without taking off the front panel. Bring the 6.3 vac out to the underside of the amp. Then install the components for the DC relay supply.

Note that the 6.3 volt filament winding is center tapped and the center tap is grounded to provide a return for the plate current. This configuration is perfect to implement a full wave, two diode, rectifier and filter. A couple of 1N5718 diodes and a 200mf/16v capacitor resulted in 4.5 volts. Enough to pull in a 5 volt relay coil. A miniature 5 volt relay coil. You don’t need a big relay here. We are only switching -135 volts at low current.

To replace the incandescent lamp I used two 12 volt LED strips containing three LEDs each and epoxied them to a small piece of circuit board. The LED strips were wired with the plus wire to one side of the double sided circuit board and the negative lead to the other side of the circuit board. The LED assembly was then glued with shoe goo to the back of the meter across the hole that was used by the incandescent bulb.

The meter needs to be removed from the front panel to do this. Be careful with the LED installation and make certain you can get the meter back through its mounting hole with the LEDs installed.

While the meter is out, use a q-tip to paint the hidden parts of the meter body with flat black paint. The color of the paint is not important but it should be able to block any light exiting from the clear meter body. This will prevent the annoying glare emitted from around the meter body. Now the only illumination will come from behind the meter scale.

The LEDs are powered from an 8.8 volt doubler supply that is built from two half wave supplies. A +4.4volt and a -4.4 volt half wave supply working off each leg of the 6.3 volt filament transformer and grounded center tap.

The -4.4 volt lead goes to the minus input to the LED bank. The +4.4 volt lead goes to the plus input of the LED bank. Thus we have 8.8 volts across the LEDs and plenty of bright white illumination for the meter scale.

Note that neither of the power leads to the LED bank uses the ground as a return. Doing so would effectively place the rectifier diode across the filament winding without any current limiting ensuring that the magic smoke leaves the diode and renders it useless.

The LEDs and relay need their own supplies otherwise the LEDs might have a tendency to blink when the relay is keyed.

The old keying line and its components were routed to the new relay. Now the external keying line has 4.5 volts on it and can be keyed by drawing low current that even the wimpiest solid state switch can handle.

Loading capacitor problem. The problem manifests itself by not allowing ‘loading’ of the amp on 80 meters. The load capacitor needs to be fully meshed at all times under all conditions on 80 meters. Turns out that this is caused by poor contact of the grounding wipers against the capacitor shaft. The poor contact is caused by time related oxidation of the metal surfaces of the wipers resulting from disuse over time. Movement of the load control when it is adjusted helps prevent oxidation buildup. A couple of shots of deOxit on the wiper surfaces corrects this problem and the load control works as it should.

It is working now. 650 watts output on 75 meters with an 80 watt input. Unloaded plate voltage at 2250, fully loaded plate voltage at 1950. Typical SB-200 performance.

A long time missing feature of the SB-200 was the lack of a standby switch. It is often desirable to disable the amp without turning it off completely. For instance, you might want to check the SWR of the antenna system at low power before enabling the amp to run high power.

I just finished replacing the relative power sensitivity pot with one that has an on/off switch. The switch may be disengaged by turning the pot all the way counterclockwise, putting the amp into standby as the keying line is interrupted.

Such action will make it impossible to use the internal SWR feature of the amp while it is in standby. It is assumed that the exciter would have its own SWR capability.

The addition of this switch to the pot was considered preferable to drilling an extra hole in the case to mount an additional switch.

Anyone need a good 1kw input amp? This one is for sale. I don’t need it. Too busy playing with the TL-922.

Linear Amp Design Quirks

Why did Heathkit (and others) use over 100 volts to bias tubes to cut-off and also activate the T/R relay?

The voltage cuts off the tubes ensuring there is no noise on receive. Using that voltage to actuate the T/R relay accomplishes removal of the bias while actuating the relay. Notice that the keying line is connected to the grids through a resistor as well as to the ground return for the relay coil. This way only a more common DPDT relay is required.

Why are the grids taken to ground through capacitors rather than tied directly to ground?

If the grids were directly grounded it would not be possible to use a negative bias on them to cut off the tubes. Yes, a cathode generated bias could be used instead to provide cut off bias but there is a more pressing reason for grounding the grids through capacitors. The grid structure of the tubes are inductive. If the grids are directly taken to ground their inherent inductive values are very likely to provide resonance at high frequencies and cause destructive parasitic oscillations. The capacitors from grids to ground counteract the inductance to help prevent parasitic oscillations. Note that these capacitor values are chosen to work with the specific tubes used in the finals. More capacitance or changing the value of these capacitors is probably not a smart thing to do. Common sense does not always result in good outcomes particularly when it is actually common non-sense.

One more unfortunate consequence of taking the grids to DC ground occurs when we experience a filament to grid short. When you short out the secondary of the filament transformer it tends to catch fire. Now why would the filament short to the grid? It could be caused by high currents due to parasitic oscillations. High currents create magnetic fields that can physically distort internal tube electrodes. Could it be that those parasitic oscillations were enabled by taking the grids to DC ground?

Could it be that by keeping an original well designed grounded grid configuration we could avoid two potentially serious problems?