Heathkit SB-221 brought back to life

Here’s an 11-minute overview of my SB-221 repair project. I offer some background about the amplifier, and what I did to bring it back to life.

The project

If you have an SB-220 or SB-221 HF amplifier, you may one day need to repair it. These Heathkit amps are built like tanks and take a beating only to come back for more. But they are only as good as the skills of those who build them from the original kits, and any modifications added later.

In my case, I found myself undoing some modifications and fixing some glaring problems, but working with an essentially well-built unit first put together in 1978.

This amp was purchased for $275 from Gene VE7KB (Kelowna, BC) in Nov. 2010, who was handling the estate sale of a Silent Key (a ham radio operator who had passed away). It had been unused since at least Sept. 2005.

The rebuilding project began in mid-April 2016. At 1:55 p.m. on May 7, 2016, the old SB-221 came back to life after 11 years (at least) of inactivity. It took me about three weeks, each day between 2 and 8 hours learning and fixing. Finally, the first contact was made — with W1GQ on 20M in the New England QSO Party at 0243z on May 8, 2016.

[Note: the refurbished SB-221 was used full-time in the May 29-30, 2016, CQ WPX CW contest — it performed flawlessly for the entire 29 hours on the air, making 1,175 contacts in 54 countries using 80M through 15M. Read the post-contest report].

This is not a how-to article. It is merely a radio operator’s notebook for this project. Please don’t ask me to diagnose equipment trouble, as I am not expert enough to answer, and there are people online who can actually help with good info.

Google is your friend. Nothing in this project notebook is my idea. I searched the web and studied photos and documents for many, many dozens of hours to find the information required to repair my amplifier. Your project will certainly be different. That learning process is more valuable than just following a recipe to make a quick repair. I encourage anyone working on these high-power amplifiers to take the time to understand each circuit, each part, and make your own decisions about what to do.

For great information starting points, check out:

• The SB-220 Compendium, by my friend John White VA7JW
• Circuit Improvements for the Heath SB-220 Amplifier, by Rich Measures AG6K

And now, on to the repair notes….

Initial condition

At some point in its history, the amp was worked on by Steve VE7DQQ, who kindly left his card in the amp for future users like me. I communicated with Steve by email in 2010 about the modifications he made. While he couldn’t remember, he did recall the amp after seeing photos and it was good to establish some provenance for this piece of equipment after many years.

Steve’s modifications included:

• 10M mod: Installation of the 10M input tuning network and band-switch contacts (unlike the SB220, the SB221 did not originally come with 10M components – they had to be ordered separately back in the 1970s).
• More convenient input tuning section: Steve rotated the entire input tuning section 90 degrees so that instead of facing the inside of the front panel, the input coil slugs (“hex holes”) can be accessed from the top of the amplifier (via an access cutout in the mesh screen) for simpler “live” adjustment. Still need to be very cautious around the high-voltage feedthrough there.
• Low-voltage peripheral power transformer: A 120V-to-14.5V AC transformer and a full-wave bridge rectifier powers a red LED on the front panel (standby switch “ON” indicator) via a DPST toggle switch on the front panel. The after-market transformer is Toshiba but unbranded (#22213906) and provides 14.5V AC to the silicon bridge rectifier — marked “2W06 / 8034” — similar to 7805 commonly available).
• 12VDC T-R/BIAS relay: The original 120VDC transmit-receive/bias relay has been replaced by a 12VDC relay with pinouts identical to the most recent Harbach replacement relay (except the coil is 12VDC not 120VDC) – this means in my amplifier the KEY (RELAY) JACK IS 12V NOT 120V – so it’s a soft-key equivalent.
• Standby switch: front panel toggle switch is standby — switches the antenna relay coil power in and out. Second pole switches 14.5VDC for the red LED indicator through a resistor.
• Zener replacement string: A high-voltage diode string built on perf board had been installed to replace the meter rectifier board’s original Zener diode.
• Meter diode protection: Both meters had diode protection (back-to-back diodes across the terminals). Both meters showed deflection when tested.
• Meter lamp RF blocking: A pair of small RF chokes at the non-RF end of the big RFC-3 filament choke near the tube sockets (I believe to reduce RF-induced lamp brightening on lower bands).
• Adjustable ALC: A front-panel 100K potentiometer adjusts ALC level (fed from orange wire to terminal strip under the input tuning section).
• Relocated the relative power sensing circuit from the output SO-239 at the back to a location near the relay. The rectifying circuit (germanium diode) on its own 3-lug terminal strip (Strip B in Heathkit schematics) is the sensitivity circuit for relative power. This is normally located at the rear antenna OUT jack. Technically, the alternate location is fine as it ties directly to the Antenna Out center conductor at the relay.

The amp was wired for 240V mains power. As I don’t have a 240V line to the radio room, one of the first things I did was rewire the amplifier’s line voltage selection for 120V.

The 1978-stamped Eimac 3-500Z tubes were removed and installed in another SB-221 to see whether they were still good. The tubes were in fine shape, at least as strong has the 1995 Eimac tubes that were in the other amplifier.

Other notes from initial inspection

• The resistor-biasing mod (removing 120V from the relay’s top pin, and moving the 100K resistor’s top from the top pin to the second pin) had not been made.
• The sockets looked very good — clips were all tight, jaws touching when no tube is inserted.
• The fan motor is Alliance Co. Model JSU-6 (115v) – this may be the original part.
• The fan blade was yellowed with age and heat, but identical to the Harbach replacement. It had burn hole on one blade, very likely from a grid RF choke catching fire during some cataclysmic event. I replaced with identical blade ordered from Harbach in 2010. (Harbach blade is Thorgren #5C175and moves 400 cfm @ 3000 rpm – no longer avail at Grainger).
• Filter capacitors were the 1978 kit originals – P/N 25-224 — 200uf 450V 35mm diameter. They have gray plastic skins, not bare aluminum. No visible defects.

Parts ordered

Ordered from Digikey:

• 8 x $8.11 381LX331M450A042 (330uf 450V 35mm dia.) caps
• 10 x 100K 3W resistors (P100KW-3BK-ND) for filter capacitor bank (2 spares)
• 20 x 27-ohm 1/2W resistors for spares for grid pin resistor-capacitor networks
• 2 x 1000pF (.001uF) 6KV ceramic caps to replace missing part, and a spare
• 4 x 0.82-ohm 3W resistors for metering board (need 1, and 3 spares)

Ordered from Harbach:

• RM-220 metering rectifier board (2016 design)
• I already had on hand a Harbach SS-221 Step Start board, but only the 240VAC version (determined by the pair of power resistors used). I ordered 2 x 20-ohm 10W wire wound resistors for step-start board as these can be added in parallel to make the board work for 120V mains supply voltage
• 2 x 10-ohm 10W resistors (for direct use on the step-start board on 120V mains configuration).
• A quantity of five spare 0.82-ohm 2W resistors (on the metering rectifier board, these usually fail during glitches)

Ordered from Rich Measures AG6K:

• A parasitic suppressor kit for 3-500Z tubes — includes a 15-ohm 10W wire-wound glitch resistor (replaces RFC-2) to protect filament transformer in glitch
• Spare 0.82-ohm 3W resistors (yes, even more)

SB-220/221 relay modification

Heathkit originally configured the transmit-receive and bias switching relay to connect 110V across the coil (to key the amplifier) and in receive mode to bias the tubes to cutoff (so they stop amplifying power).

In Circuit Improvements for the Heath SB-220 Amplifier, Rich Measures AG6K recommends a simple and very beneficial modification to the relay wiring to protect the filament transformer from melting down if there is a filament-to-grid short in one of the tubes.

Note: my relay is 12VDC, but most SB220/221 units use a 120V version of this relay.

LEFT COLUMN

PIN 1   LARGE BARE WIRE TO PIN 2
PIN 4   SMALL COAX CENTER CONDUCTOR TO TUNED INPUT
PIN 7   SMALL COAX CENTER CONDUCTOR TO RF INPUT JACK
PIN A   SMALL WIRE TO ANTENNA RELAY JACK OR TO SK-220 UNIT
        (via front panel standby switch)

MIDDLE COLUMN

PIN 2   LARGE BARE WIRE TO PIN 1
PIN 5   LARGE BARE WIRE TO TANK CIRCUIT
PIN 8   LARGE COAX CENTER CONDUCTOR TO RF OUT JACK

RIGHT COLUMN

PIN 3   EMPTY (originally had ONE END OF 100KΩ RESISTOR & SMALL RED WIRE 
        FROM 120VDC SUPPLY & SMALL RED JUMPER WIRE TO PIN B)
PIN 6   (new) ONE END OF 100KΩ RESISTOR and LARGE BLACK WIRE TO 
        RECTIFIER/METERING BOARD
PIN 9   OTHER END OF 100KΩ RESISTOR AND GREEN/YELLOW WIRE TO 
        CENTER TAP OF FILAMENT TRANSFORMER
PIN B   (coil)  12VDC (stock config. had SMALL RED WIRE FROM 120V SUPPLY)

Repairs and upgrades

• Installed upgraded Harbach RM-220 metering/rectifier board (2016 design, have a 2010 built for the other amp if ever needed, but it has a 2006 design working already).

The RM-220 rectifier metering board’s 2010 design by Harbach (left) is updated for 2016 with much larger components and a larger board (right).

Note: the 2016 design provides shorter standoffs (spacers) to allow wiring and diode string to stay clear of the filament transformer shell – be sure to replace these BEFORE wiring the board into place, as doing it at the end is much more difficult, despite Harbach putting this step near the end of the instructions.

• Installed Harbach soft-start SS221 board.
• Replaced RFC-2 with 10-ohm 10W wirewound glass-encased “glitch” resistor (at feedthrough).
• ALC adjust pot on panel had lugs at bottom – touching or almost touching the chassis. Turned to the side and tightened.
• Standby LED was lose – used RTV adhesive on back to secure.
• HV meter’s + lug was loose. Fixed.
• Cleaned RF compartment. Tightened HV passthrough and the shorting strap and bar (pitted copper from arcing (at some point, the amp must have been powered up while shorted, and the bar had a tar-like lumps which were sanded off.
• Checked keying voltage at RCA back apron. Measured 14.5VDC – keying line from FT-2000 handles that very nicely.
• Installed a .001uF (1000pF) 6KV ceramic cap from HV passthrough to ground (currently missing part) (Had to install a lug in the empty screw hole, used a 6-32 screw from the RM-220 Harbach kit and substituted a nylon screw when installing the metering board.
• Carefully checked metering board wiring – strange configuration from previous work (B- wire not present at top of capacitor bank, and Hole B black wire comes from Plate meter instead – very non-standard). All fixed.
• Replaced 1uH RF chokes on one grid pin at each socket with 27-ohm 1/2W resistors.
• Added additional 470pF ceramic caps in parallel with 200pF mica caps on all grid pins.
• Installed Q-damping 10-ohm resistors (from Rich Measures kit) between the .01uF caps (the capacitors are soldered to the coax center on the terminal strip on the RF side of the RFC-1 filament choke and now go through a resistor to each filament terminal.
• Built and installed anode (plate) parasitic suppressors – each anode has 2 x 100K+coils and pair of 1-ohm fuse resistors.
• Removed old Zener perf board (was connected from HV passthrough to Hole J on the metering board)
• Installed upgraded capacitors and bleeder resistors (saved old set as spares)

The original filter bank of 200uF electrolytic capacitors with original green 30-ohm resistors.

The new capacitor bank — eight 330uF 450V 105C electrolytic caps with new 100K-ohm resistors — which dissipate abut 35W less heat. That’s stiffer voltage and less heat. Good stuff for $8.40 per cap and resistor.

• Relocated relay resistor to second pin down, and removed the red 120V wire entirely as the after-market 12VDC relay uses a separate power source.
• Replaced powr cord’s 240V plug with a 120V plug.
• Replaced the old fan blade with Harbach (Thorgren) new model – identical swap in
• The four feet had been put on with machine screws and nuts, making them damaging to remove the case – gouging the sides of the front panel as the case slid off. I discovered tapped holes in the chassis that fit 1-1/4″ 6/32 screws.

Wiring issues addressed

• HV/RelPwr meter was wired backwards – wires didn’t go to the correct lugs or board holes. Lug 1 of pot should be black wire to right(minus) of HV meter. Red wire from lug 4 of HV switch moved to left (+) of HV meter (it was on the negative post side, which is wrong).
• Multimeter switch had no HV connection, and the lug was broken off. Orange wire was corrected — goes to Lug 1 of meter switch(previously was soldered to lug 1 of relative power sensitivity pot which is wrong).

• HV feedthrough should have heavy blue wire to Hole J on metering board.
• Black wire from Hole C on the metering board went nowhere (cut off and tucked away) – Hole C should go to Plate meter negative (right) terminal
• Instead, for some reason the Plate meter negative terminal’s black wire was incorrectly wired to board Hole B (Hole B should have a black wire from top of filter capacitor bank (B-)
• Heath instructions for “two brown and one green wire” through the grommet for meter lamp power and relative power had been replaced with black shielded two-conductor cables. That’s fine – except the shield of one meter power cable had broken away from the 3-lug terminal strip. Resoldered and fixed.

Fan shroud

A metal shroud around the fan improves airflow around the two tubes, and across the filament pins on the underside. The idea was posted by David K5DBX and it seemed a simple addition with good results.

I purchased 10 feet of 12″-wide aluminum flashing  from Home Depot (the roll cost me $11,  and is about 22 ga. metal stiff enough to be solid and thin enough to cut with scissors. There’s enough on a roll to make man y, many fan shrouds — I have two SB-221 amplifiers, and will have the shrouds in both.

I cut a strip of aluminum 23″ long and 2.75″ wide. To install it, I had to remove the tubes and the fan blade, allowing me to slip the slightly compressed (concentrically “wound up” a bit) ring of metal into the round opening. It pretty much formed a perfect circle when I released the metal.

I used clear RTV silicone adhesive to secure the two ends of the sheet together at a convenient location on the underside, then gently applied more RTV caulk-like around the perimeter of the shroud, smoothing it with a damp finger tip as I went.

Tuning index markings on panel

Last summer at a garage sale I picked up a Brother label maker (deluxe model) for $1. It prints black text on white adhesive tape. I decided to make custom tuning marks for the SB221. It worked very nicely.

Next, I will find the Brother cartridge to print white text on clear tape and re-do the markers so they look closer to th e original screen-printed markings. And the size of the label pieces won’t matter as much (I messed up on the 20M and 15M Load marker sizes).

Another idea I would lie to try is cutting a large piece of iPad screen protector film and attaching it to the faceplate with the knobs off, Then I can affix any kind or markers I want, or use a permanent marker, without marring the faceplate’s original finish.

Final layout

To-Do list

20M input tuning: The 20M input tuning showed a 2:1 SWR — but it is 1:1 on all other bands.

Rather than mess with the 20M tuning coil (nobody has touched it), I decided to take a close look at the mica capacitors at the 20M tuning coil. Sure enough, there were two capacitors soldered to the coil — a 180pF on the output side, and a 220pF (not 400pF as indicated in the SB221 Heathkit manual) on the input side, both to chassis ground lugs.

The 180pF cap was burned up — a big chunk  was missing. The cap fell apart in two pieces as I unsoldered it.

Now, the chances of finding a 500V or better 180pF mica capacitor in the junk box are slim to none around here. I just don’t have much of a junk box. But I do have a pile of old circuit boards from various equipment people have handed over for parts (knowing I might find a use for them).

I searched every board I could find. Looked at every mica cap — nothing even close to 180pF.

Then I looked under a DTMF touchtone keypad on an old PBX telephone circuit board. There were several 500V mica caps — 560pF, 270pF, 100pf and 20pF. I unsoldered them and confirmed their values with the multimeter and they were each within 5pF of correct — pretty good tolerance. On a whim, I checked the 560pF in series with some other smaller values, just to see if I could get close to 180pF.

Providence was at work. The series value of the 560pF and 270pF measured 181pF! — as close as I would ever get, even with a brand-new high-quality 180pF cap with 1% tolerance.

Soldered them together, then soldered them into the amplifier (which took some doing in tight quarters). Put the amp back together and turned it on.

20M SWR is now 1.4:1, which is just fine by me — the 220pF cap may have been damaged by whatever blew out the 180pF cap. My educated guess (seeing as it was me) is it was a hot-switch incident (changing bands while transmitting) while testing far too late at night — like 3 a.m. — recently.

On my next Digikey order, I ‘ll get 180pF (338-1082-ND) and 220pF (338-3101-ND) mica caps and replace both on the 20M tuning coil.

Power-up grid and plate current: After completing the rebuild, I noticed that upon power-up the grid current meter was showing negative grid current for about one second, before returning to normal (no grid current). Tried starting the amplifier again, this time watching plate current, and it, too, went to about .180mA for about a second — seemed the soft-start board was not wired in quite right.

The amp had been running just fine in normal operation for a few days, but the odd power-up behaviour had me bothered. So I opened things up for a closer look. I can’t find the cause of the momentary rise in grid current and plate current upon power-on. Double-checked the Harbach soft-start wiring — in fact, I took the board out and retraced all the AC mains wiring through the circuit breakers to the front switch, then re-installed the Harbach soft-start board per the detailed instructions.

The start-up rise is still there. And it seems to be only while the soft-start resistors are in circuit — once the relays short out those 10-ohm resistors, all is entirely normal (i.e. no plate current or grid current during receive).