The Collins R-390A was built for AM, CW and MCW, and it shows the moment you tune in a sideband signal. This guide covers the practical side: how to copy SSB on an unmodified receiver, how to fit the Lankford/Cornelius 2-diode AGC modification, how to test it, and how to add an outboard detector through the rear-panel IF output. For a side-by-side look at the adapter options, see the R-390A SSB adapter comparison guide. For how owners have tackled the problem since the 1950s, see SSB and the Collins R-390A: From Military Converters to SDRs.
Why SSB Is Hard on a Stock R-390A
In 1963 Captain Paul H. Lee, W3JHR, set out the problem. The R-390A’s AM diode detector adds considerable distortion to SSB, and the levels of incoming signal and BFO injection aren’t right for proper SSB detection. Turning down the RF gain fixes the levels, but it throws away the AGC. The receiver was designed to hand SSB to an outboard converter through its 455 kHz IF output, and Chuck Rippel’s owner’s manual calls SSB “a kluge on the R390A at best.”
The answer is to keep the signal small compared with the BFO. On a stock set you do that by hand with the RF gain. The Lankford mod lets the AGC do it for you.
Part 1: Copying SSB on an Unmodified Receiver
Step 1: Zero the BFO
The BFO PITCH knob’s zero mark needs to sit exactly on zero beat, or every offset below will be wrong. Don Stepka’s method from the R-390 mailing list:
- With the BFO off and BANDWIDTH at 0.1 kc, tune an AM carrier to the centre of the passband.
- Turn the BFO on and bring it to zero beat with BFO PITCH. Don’t touch KILOCYCLE CHANGE.
- Loosen the BFO PITCH knob and refit it pointing exactly at 0. Leave the BFO on.
Step 2: Set the controls
The Y2K-R3 manual’s SSB procedure starts from the CW settings and changes a few of them. Use this table as a starting point:
| Control | Setting | Notes |
|---|---|---|
| FUNCTION | MGC | Don Stepka found AGC or MGC makes little difference on a stock set, because the low IF level generates little AGC. |
| AGC | SLOW | Stepka preferred MED or FAST. Lee found FAST pumps and MED is a little slow for voice break-in. |
| LIMITER | OFF | |
| BANDWIDTH | 4 kc (or 2 kc) | The 2 kc filter is tighter; the 4 kc filter sounds more natural. |
| AUDIO RESPONSE | WIDE | |
| RF GAIN | Low, then raise | The manual starts from RF GAIN 10. Stepka and others turn it well down, then bring it up only until normal band noise is heard, because too much signal swamps the BFO. |
| LOCAL GAIN | High | Stepka set it all the way up and controlled the level with RF GAIN. |
| BFO | ON | Zeroed as in Step 1. |
Step 3: Offset the BFO for the sideband
Set BFO PITCH to the opposite side of the carrier: minus for USB, plus for LSB, as both Lee and the Y2K-R3 manual give it. A good offset is a little under half the bandwidth: about 1 kc with the 2 kc filter and about 2 kc with the 4 kc filter (Stepka). Others on the list used 1.5 to 2 kc on the 2 kc filter, or up to 2 to 3 kc, so try a few settings and keep what sounds best.
Step 4: Tune and ride the gain
Tune the signal in with KILOCYCLE CHANGE until the voice sounds natural, then trim BFO PITCH for best intelligibility. If voices turn harsh or distorted, the signal is overpowering the BFO, so back off RF GAIN rather than LOCAL GAIN. On a stock receiver you’ll ride the RF gain as signals fade and change. That’s the chore the Lankford mod removes.
Part 2: The Lankford/Cornelius 2-Diode AGC Modification
This is the most widely used SSB improvement for the R-390A. It grew out of the earlier Cornelius AGC mod in Hollow State News 1 and 10, which replaced R546 with a diode and changed two other resistors. Dr. Dallas Lankford published his simpler version in HSN 23 (Fall 1989): add two diodes and remove nothing. Later issues, 27 and 36, revisited it. Two small-signal diodes go across two resistors in the AGC circuit. They let the AGC time-constant capacitor charge quickly without passing through the resistors, while it still discharges through them, giving fast attack and slow decay. List members explained that it works on SSB because the R-390A derives its AGC ahead of the point where the BFO is injected.
Parts
- Two 1N4148 (or 1N914) silicon small-signal diodes. Lankford first suggested the 1N270 in HSN 23, then the 1N4148 or 1N914 in HSN 27. Later list reports found the 1N270’s back leakage limits the AGC voltage. The diodes need very high back resistance, so avoid germanium types.
- Optional: one 47 pF mica capacitor, for extra BFO injection (see below).
Where the diodes go
| Diode | Across | Orientation (from Lankford’s HSN 23 schematic) |
|---|---|---|
| D1 | R547 | Anode toward pin 2 of V506A, so the diode points away from that pin |
| D2 | R546 | Cathode toward pin 1 of V509A, so the diode points toward that pin |
V509A is the AGC rectifier and V506A the AGC time-constant stage. These orientations match the schematic Lankford published on page 4 of HSN 23, and later descriptions on the R-390 mailing list. Both diodes point from the AGC line toward the rectifier, so the negative AGC voltage builds up quickly. It’s still worth checking them against your receiver’s schematic before you solder.
Fitting it
- Unplug the receiver and discharge the filter capacitors.
- Find R546 and R547 using the schematic and parts-location diagrams in the TM or the Y2K-R3 manual.
- Solder each diode directly across its resistor, observing the orientation above. Nothing is removed and no holes are drilled, so the mod can be undone with a soldering iron.
- Optional BFO injection boost: solder the 47 pF capacitor in parallel with C535, the existing 12 pF BFO injection capacitor. Lankford recommended it to give enough BFO signal to prevent audio distortion on SSB. One list member used 33 pF instead.
- If you added the capacitor, readjust C525, the BFO neutralizing capacitor, for a null (minimum reading) on the carrier meter. This keeps the extra BFO energy out of the AGC circuit. On the receivers described, C525 is reached through a hole in the left side panel of the main frame.
Using the modified receiver
Set FUNCTION to AGC and leave the RF gain up. Lankford found the MED position gave acceptable SSB with his version. He measured attack times of 0.001, 0.01 and 0.2 seconds and release times of 0.01, 0.12 and 2.5 seconds for FAST, MED and SLOW. Later list reports mostly preferred SLOW, so try both. Then follow Steps 1, 3 and 4 above: zero the BFO, offset it for the sideband and tune.
Testing the Result
- RF gain test: on a steady SSB signal with AGC on, slowly reduce the RF gain. The audio level should stay roughly constant until the carrier meter falls to 0.
- Full-gain test: with RF GAIN at maximum, SSB should be clean and undistorted.
- If it still distorts: try a new BFO tube (V505). A weak BFO tube can look like a failed mod.
- Bench checks: Larry Haney’s notes on r-390a.net (linked below) give signal-generator readings for the detector and AGC voltages with and without the mod.
Side Effects and Later Refinements
- Several list members reported no change in AM performance, and one found CW somewhat better.
- Overall audio gain drops slightly.
- Some heard a slight pop as the AGC first engaged on strong, heavily processed signals. SLOW AGC removed it, as did replacing a 1N270 with a 1N4148, and setting the limiter to 2 or 3 helped.
- Larry Haney’s later analysis lists further trade-offs: the AGC delay is removed, weak-signal signal-to-noise is lower, audio output is reduced, the noise limiter distorts more and the carrier meter reads low. His follow-up modifications (2019 to 2024) address these and are worth reading before you start.
Part 3: Adding an Outboard Detector Through IF OUT
If you’d rather not touch the receiver’s internals, use the IF output the R-390A was designed with. It’s the BNC jack marked IF OUT at the upper left of the rear panel (J116). It supplies 455 kHz at about 180 to 220 mV into 50 ohms, taken between the third and fourth IF stages through cathode follower V509B, which also isolates the receiver from the load. It was intended to feed the military CV-591/URR SSB converter.
Over the years that jack has fed:
- military converters such as the TMC CV-591A/URR
- commercial outboard detectors such as the Vintage Radio Products PD-1 and the Sherwood SE-3, which connect with a cable to IF OUT
- home-built detectors, such as OH2BRY’s adapter, which needs no receiver modification but wants the receiver on AGC rather than MGC
- SDRs and panadapters, as in Hermes Lite 2 as an R-390A companion panadapter
An outboard detector supplies its own carrier, so you’d normally turn the R-390A’s BFO off and use the detector’s sideband selection. Take audio from the adapter to your amplifier or speaker. One list design, WB2LHI’s, fed its audio back into the receiver at the diode-load terminals instead.
Or Leave It Stock
Not everyone agrees the R-390A should be modified at all. Chuck Rippel told the R-390 mailing list in 1997 that he neither modifies the receiver for SSB nor recommends it, because it was built for AM and CW. Both approaches in this guide are reversible: the 2-diode mod comes out with a soldering iron, and an outboard detector simply unplugs.
Frequently Asked Questions
How do I copy SSB on an unmodified Collins R-390A?
First zero the R-390A BFO: with BANDWIDTH at 0.1 kc, centre an AM carrier, switch the BFO on, zero-beat it with BFO PITCH and refit the knob pointing at 0. Use the 4 kc or 2 kc bandwidth, LIMITER off, LOCAL GAIN high and RF GAIN low, raised only until band noise is heard. Offset BFO PITCH minus for USB or plus for LSB, about 1 kc on the 2 kc filter or 2 kc on the 4 kc filter, then ride the RF gain.
How do I fit the Lankford 2-diode AGC mod to an R-390A?
Work with the R-390A unplugged and its filter capacitors discharged, because its B+ voltages can kill. Solder one 1N4148 or 1N914 diode across R547 with the anode toward pin 2 of V506A, and another across R546 with the cathode toward pin 1 of V509A, checking orientation against your schematic first; nothing is removed. Optionally add 47 pF across C535 for more BFO injection, then readjust C525 for a null on the carrier meter.
How do I test the Lankford AGC mod on an R-390A, and what are its side effects?
With AGC on and a steady SSB signal, slowly reduce the R-390A RF gain: audio should stay roughly constant until the carrier meter falls to 0, and SSB should be clean at full RF gain. If it still distorts, try a new V505 BFO tube. Overall audio gain drops slightly, and Larry Haney lists removed AGC delay, lower weak-signal signal-to-noise, more noise limiter distortion and a low-reading carrier meter; his 2019 to 2024 follow-up mods address these.
What signal does the R-390A IF OUT jack provide for an outboard SSB detector?
The R-390A IF OUT jack is BNC connector J116 at the upper left of the rear panel, supplying 455 kHz at about 180 to 220 mV into 50 ohms. It is taken between the third and fourth IF stages through cathode follower V509B, which isolates the receiver from the load. It was intended for the CV-591/URR SSB converter and suits outboard detectors such as the Sherwood SE-3 or an SDR; normally turn the R-390A BFO off when using one.
Sources & Further Reading
- SSB conversion compilation, 1963–2000 (r-390a.net), including Capt. Paul H. Lee’s 1963 article and R-390 mailing list posts on the 2-diode mod and BFO technique
- Hollow State News #23 (Fall 1989), page 4: Dallas Lankford’s 2-diode AGC mod and schematic (VK6ADA HSN archive, issues 13–25)
- Larry Haney: fixes for the Lankford 2-diode SSB AGC mod (r-390a.net)
- Y2K-R3 R-390A manual, Chapter 1, Chapter 2 (operation) and Chapter 3 (theory) (r-390a.net)
- R-390A IF deck notes (r-390a.net)
- Chuck Rippel: R-390A owner’s manual (amwindow.org)
- Dr. Dallas Lankford: Boatanchor Legends tribute (VK6ADA)
- Hollow State Newsletter archive (VK6ADA)
- Dallas Lankford Collection of Modifications and Projects (kongsfjord.no)
- R-390A Aftermarket SSB Adapters: A Comparison Guide (VK6ADA)
- SSB and the Collins R-390A: From Military Converters to SDRs (VK6ADA)
Mike Peace VK6ADA / r-390a.net and foxtango.org Administrator