My work knife made an odd sound the other day and only would open a short amount without help. It is an “assisted open” knife with a torsion bar that helps snap the blade open with one hand. This is an important feature for safety if only one hand is available, and comes in handy in tight spaces.

I didn’t know if I would have to buy a new knife or if it could be repaired. I watched this video and it looked really simple. But first I had to go to the Kershaw (Kai) website and request a new torsion bar for the specific model knife.

I received the new torsion bar in about a week (free of charge). I opened the knife on my work bench and took out the torx screws (T6 or T7 and T8).

The new torsion bar just sits into the two holes shown below. No tension is applied at this point.

This part was simple. But during reassembly, I found it difficult to get the torx holders (nuts) on the other side of the knife to seat properly. On closer inspection I found that the torx holders have a flat edge so it can only fit one way into the hole (and this provides a way to hold the nut so it doesn’t rotate when adding the screw).

Once the screws were back in place, the knife closed and opened like new.

Looks aren’t everything (wiring panels)

Today I had to replace a card access panel that was installed about 20 years ago with a newer (but still obsolete) panel with larger dimensions. The previous technician chose to put 4 panels in a square formation with PVC and metal conduit between them. He chose to route cabling for power and data up through one panel and through another until he reached the other panel. It looked very clean and tidy. But it made replacing one of the panels an absolute chore. There is one building where this approach was done with 14 panels!

Here is a picture of a smaller cluster of panels using PVC interconnects. It looks tidy, but when the first blue panel fails, all of the other panels have to be taken down in order to replace it. Bad design. Cabling could still be tidy and individual to the panels. Now both the cabling and the interconnects have to be changed. Likely I will remove all the interconnects after clearly labeling each cable, and perhaps color coding them per panel for quick routing.

On the one I recently changed out I also had to label the cables (since he never bothered), and cut off about a dozen zip-ties inside a single panel. Again, it looked tidy, but tracing the unlabeled cables was near impossible until they were all cut free. Some cables had to be retracted through 3 of the other 4 panels while not disturbing or shorting any of the other panels.

Then I had to remove the conduit interconnects so I could remove the old panel. I got the new one situated and ran the cabling directly to the new panel instead of through the other panels. This makes any future troubleshooting and wiring logic a lot easier since I can tug on a cable and not confuse it with the cabling of another panel.

External cabling is fine so long as it is still tidy, and tidiness makes troubleshooting easier as well. Label your cables so another tech can understand what panel they belong to as well as the purpose of the cable. I’ve started using shrink wrap labels where possible because they are easy to read and don’t smudge. Don’t use these where the cable may experience high heat or the text may disappear. Document any unusual wiring and keep that information handy. I have hundreds of pictures and documents for this facility stored on their server that gets backed up nightly, because I won’t always be around to help them with the next project.

Gasoline can hell

Since the new gasoline/petrol can rules came out, it has been nearly impossible to find a decent plastic gas can. The nozzles are so complicated now that dispensing gasoline typically gets fuel all over the place, thus increasing the hazard dramatically. I’ve heard that the new anti-explosion measures were put in place to help protect Darwin Award candidates who pour gasoline on fires.

I bought some replacement nozzles on Amazon to try and bypass these new gizmos, and they looked very promising. Fast-forward two months and the thin indented gaskets used on the nozzles soften in the presence of fuel, and tear when the nozzle is screwed too tightly onto the gas can, meaning the new nozzle will also dribble rather a lot of gasoline when used. The other problem is the gasket they use is indented to fit over the lip of the nozzle and the white anti-flame gizmo inside the nozzle. This weakens the gasket and lets it tear or twist when the collar is attached, making the seal break and lets gasoline leak.

I chose to remove the white part and replace the single indented gasket with two gaskets, one on either side of the black nozzle lip. This seals both areas that could leak gasoline.

I’ve read that nitrile is the best material for gasoline resistance, but I also bought a sheet of Fel-Pro rubberized cork gasket sheet and cut a couple of gaskets to try. (Cut to the chase: the Fel-Pro cut in 44mm OD 33mm ID works perfectly, but only tighten collar until it stops. Any further will deform the gasket.)

COLLAR THREAD DIAMETER IS 1 5/8 INCHES OR 41mm (40mm OD gasket is slightly undersized for the can-side gasket, it falls out when taking off the yellow collar. 41mm OD should stay in place.) Trough of thread is 44mm. This is important for sizing a gasket that will fit inside the threads and be held in place.

COLLAR LIP HOLE IS 1 5/16 INCHES OR 33.5mm (33mm ID gasket hole actually lets nozzle neck fit through the gasket easily while being held in place by the collar lip)

Some brands of gas can like Blitz use a pressure fit rounded plastic part instead of a flexible gasket. That seems to work better than the thin flexible gasket, but I’m hoping a thicker gasket of nitrile will fix the issue of leaks on the replacement nozzles. If not, I may be investing in an expensive metal “Jerry can”.

Here was the original can and spout, Scepter brand:

This is what I had after replacing the nozzle:

The Fel-Pro rubberized cork gaskets worked perfectly. I cut out two (very rough) gaskets and placed them on each side of the nozzle lip. One seals between the red tank and the nozzle, and the other between the nozzle and the yellow collar. The lip of the nozzle is sandwiched between the gaskets. At first I tightened down as hard as I could, thinking it would seal better, but that leaked a lot because the gasket became deformed. I then re-seated the gasket and tightened just until it stopped. No leaking at all! W00t! This may be true for the indented gasket that comes with the kit, but I haven’t tried this yet.

I tried to find a better and cleaner way to cut the material than scissors and a drill, but this worked. Just don’t leave any pieces that could come off into your gas tank. Metal dies in 41mm and 34mm are too spendy for the amount I would use them. Cheapest I found were Mayhew though they can cut both holes at the same time.

I did just try 41mm OD x 33mm ID x 4mm nitrile rings. Failed. So far the solution seems to be the Fel-Pro material, perhaps because it is flat it seals better. I can’t seem to find flat o-rings.

Example of poor vibration dampening

At work I found an example of how NOT to use vibration dampening. This air conditioning unit is resting on 4 rubber/cork pads, but then bolts were run through the pads into the rigid wood base, which is also bolted to the rigid metal base. This allows the vibrations to travel freely through the bolts into the frame and into the metal building. A pad really isn’t the best approach here, if the amount of vibration would allow the unit to move off the base. Spring dampeners on each foot would be a better way. Spring dampeners also need to be matched to work with the weight of the unit.

Ferrules for stranded wires

One of the byproducts of the virus quarantine is spending a lot of time watching YouTube videos. One of them had to do with using ferrules on stranded wire when assembling projects, in order to prevent frayed ends from possibly shorting to other connections, and to provide a more solid electrical connection. Ferrules are generally not required by code (though may be by project specs), and most block screw terminals are rated to accept stranded wires directly, but ferrules can often make a better quality connection.

Ferrules (in this context) are thin copper tubes coated with tin, and can either have an insulated collar or not. Ferrules keep all the strands of a stranded copper wire together and the crimping tool compresses the copper tube around the strands so that it is resilient and mechanically the same as a solid core wire for the length of the ferrule. Ferrules are superior to tinning with solder, and more consistent and reliable. Temperature in a circuit can also melt solder under certain conditions, making it less reliable than ferrules.

The more I use them, the more I prefer the uninsulated ferrules. On some projects I use insulated ferrules to make certain wires stand out visually. The collar is designed to fit over the wire’s insulated coating.

I’ve had to deal with frayed low-voltage wires for years and decided to get both square and hexagonal ratcheting crimping tools, and an assortment of ferrules. The square shape works best with terminals that compress into a square shaped hole, and the round ones for round shaped holes. However, if the wire is small enough compared to the terminal opening, the shape of the crimp may not matter.

outside diameter comparison

terminal opening difference
barrier terminal strip
crimp tools

One of the first things I discovered is that one should check first whether a ferrule will even fit into the terminal or not, prior to attaching it to the wire. The best way to test this is with a piece of spare wire with a crimped ferrule. Some terminals (such as the blue half-moon opening kind in the picture above) that can take an 18AWG stranded wire may not be wide enough to accept an 18AWG rated ferrule, or may require a particular shape of crimping. I have some terminals on older panels that don’t have very large holes, even with the screw retracted fully. Also, some ferrule brands have thicker metal, so the same gauge rating can vary in outside diameter. The tolerances are tiny, but significant.

When stripping the wire for attaching a ferrule, the wire must reach the end of the ferrule or the ferrule metal can break off in the terminal.
strip wire to length of ferrule
broken ferrule

There is also a trapezoid crimper that produces a sort of D or B shape; a DuPont or Molex pin crimper for small wires and JST connectors; and the more common lug crimpers for attaching spade, fork, or ring connectors. I won’t be covering those or the many other specific job crimpers, though for each kind a ratcheting tool is preferable to ensure that the right amount of pressure is used on the connector.

There are also differences in the kinds of screw terminals, even though they may be marketed under the same common name, such as “rising clamp terminal block” (see the green terminal block in the pic above). Some indeed rise as the screw turns, and the square shape is maintained. Others sold with the same name have a circular hole with a spring metal tab that screws down onto the wire (like the blue terminal block in the pic above). That clearly should have a different name. This latter kind also tend to have the smallest available space for the wire and may not accept a ferrule easily, unless it is under-sized.

Dual ferrules are designed to take two stranded wires and create one crimped connection out of both. That is great if the terminal is large enough to receive the ferrule. Do not twist the wires together first, just strip them, line them up and slide them in prior to crimping.

In some low-voltage situations with multiple wires going to a single terminal point, I have been forced to use a wire nut or Wago connector to join the wires together, and then use a ferrule on a single pigtail wire to the terminal. The point is to still have a clean connection at the terminal with less chance of shorting strands, but it includes a wire nut or Wago connector also. Wago connectors have the advantage of being able to disconnect a single wire from a group without breaking contact for the other wires. Wire nuts may not always reliably connect multiple stranded wires. Wires can twist around and appear to be connected electrically, but not be. A terminal strip is another possible way to connect multiple wires together, and makes for easy troubleshooting of circuits.

Also, there should generally only be one ferrule per enclosed screw terminal. Two may fit into an open screw terminal (barrier terminal). Multiple stranded wires may fit into an enclosed screw terminal, but multiple ferrules generally will not, especially if they have insulated collars. It is better to use an intermediate terminal (Wago or wire nut) to join the multiple wires and then use one that goes to the final terminal.

Always check your national and local codes regarding wiring. There may be specifics that must be met, particularly in commercial installations. If you might possibly be sued, make sure you cover yourself by following wiring code.

Smoke alarms going off for a few seconds

My wife had been complaining about the smoke alarms going off for about 10 seconds every other day. These are fairly new units (2 years old) with new batteries. I replaced all the batteries again and we got 1 day without the alarms sounding.

Today I began pulling all of them down, trying to find the culprit. When I walked into the master bedroom, I smelled hairspray. Hairspray is an aerosol of sticky micro-particles that gum up exhaust fans, white noise generators, and smoke detectors. I have already replaced the exhaust fan in the bathroom once, and cleaned the fan in the white noise machine.

If I’m right, I can replace the entire smoke detector unit and we should be good for another couple of years.

UPDATE: March 28,2021  Yes, hairspray was apparently the problem. No odd alarms since then. Although I can’t recommend covering the alarm, that is the only way to keep it from getting clogged again outside of using the hairspray in a different location.

PIR (REX) Shutter Adjustment

I was installing my first Passive InfraRed (PIR) Request To Exit (REX) device recently and it said that I could adjust the shutters to narrow the field of where the device looks. It said that there are 8 indents that can be selected.

At first I was looking at the plastic Fresnel lens and the 8 sections it has, trying to understand what I was supposed to do with it. (The directions were text only, no diagram).

While testing the device, I noticed the two shutters that are under the front lens cover. Suddenly it all made sense, so I post the pictures here to help anyone else that didn’t immediately see the shutters either.

PIR1PIR2PIR Shutters

Yanmis YF-7 two-relay timer

I bought a Yanmis YF-7 from Amazon about a month ago and am finally getting around to playing with it. It can be powered by a 9V battery and has a programmable timer for controlling two on-board relays.

[EDIT: Unit failed. I had it in a drawer for a few months, took it out to try using it and it won’t turn on. It just heats up like it has a short (none visible) and nothing is connected to it besides a 9V battery. Sadness is mine.]

The main problem I’ve found so far is that there is no documentation for it. They sent no manual with it, and an internet search yielded just a tiny amount of information from the German Amazon website, and then from competitor’s models that have similar programming. They did say that I could email them for a manual, but I don’t like giving a live email address out. I’d much prefer that they had a link for downloading a PDF manual.
Click here for a PDF of this document

So, I will post what I find here, and try to keep it updated with the latest information.

YANMIS YF-7 USER GUIDE by W. Jason Woodrow 2019

Yanmis YF-7
Part Number: Yanmisepv0txaurd
UPC: 746771544984
Description: Multifunctional Delay Relay, DC 7-30V

9V model buttons

The unit can be powered by a single 9V battery. The unit has two relays each with a red LED on the board that indicates which one is currently powered. They operate independently, and can operate simultaneously depending on programming.

When the unit is powered-on, by default it will display dashes on the LED display “- – – -” meaning all timers have run out of time.

The buttons at the bottom are K1, K2, K3, K4.

K1 held for 1 full second will enter program mode.
In Program mode, P will blink in the leftmost column. Choose which program you want, based on the number. P-##. Lowest ## is 11, highest is 48.
Settings are saved as they are made.

P: choose program from 11-18, 21-28, 31-38, 41-48 (excluding 19, 20, 29, 30, 39, 40), 32 programs total (although there are 3 that appear to be duplicates of others).

K2 increments the left digit
K3 increments the right digit
K4 decrements the right digit
Holding any of the these buttons cycles through the numbers

A timer (controls Relay 1)
B timer (often controls Relay 1)
C timer (controls Relay 2)
D timer (often controls Relay 2)

In program mode, press K1 briefly to display the A timer. See the chart below describing how the decimal affects the time. Press K1 briefly again to show B timer, etc. D timer is the last setting.

When display shows four dashes “- – – -” you have exited program mode and saved the settings. To enter Program mode again, press and hold K1 for one full second.

NOTE: In timer mode (A, b, C, d) K4 moves the location of the decimal or removes it. K2 increments all three digits, K3 decrements all three digits. Holding K2 or K3 cycles through the digits.

This part is a little confusing, and it can affect the way the other timers are interpreted by the device. If you stick to one method for all the timers, you’ll get the most consistent results.

A004. (decimal showing after last digit) is 4 minutes (Range is 1 to 999 minutes)
A004 (no decimal showing) is 4 seconds (Range is 1 to 999 seconds [16min 39sec])
A04.0 is 4 seconds (Range is 0.1 to 99.9 seconds [1min 39.9sec])
A4.00 is 4 seconds (Range is 0.01 to 9.99 seconds)

Through testing each mode, I was able to figure out if it started the timer based on power-on, trigger pulse high, trigger pulse low, or trigger steady power. A couple of modes were too odd to completely figure out.

For Triggers, I used a wire to feed from the negative power input to the two triggers (called Signals in the picture).

P-11: Trigger 1 keeps Relay 1 activated as long as trigger power remains. Same for Trigger 2 and Relay 2. Timers are not used.

P-12: Each trigger toggles the relay closed or open as many times as triggered. Timers have no function

P-13: Trigger 1 closes relay 1 for time set on timer A. Trigger 2 closes relay 2 for time set on timer C. B and D timers have no function.

P-14: appears to be the same as P-13

P-15: appears to be the same as P-13

P-16: appears to be the same as P-13

P-17: appears to be the same as P-13, except that the timer only shows “—-” (4 dashes)

P-18: Triggers inactive, Timers A & C start at power-on.

P-21: Trigger 1 runs Timer A then closes Relay 1 until power-off. (Trigger 1 can run timer again, but relay remains open until power-off).
Trigger 2 runs Timer C then closes Relay 2 until power-off.

P-22: Similar to P-21, except power to triggers have to remain on (cannot be a pulse). Timer will reset if trigger loses power.

P-23: Pulse to Trigger 1 runs Timer A then activates relay 1 until next trigger pulse which powers off relay 1 and starts timer A again. (Relay is off while timer runs, relay activates when timer reaches 0).
Pulse to Trigger 2 runs Timer C then activates Relay 2 until next trigger pulse which powers off Relay 2 and starts Timer C again.

P-24: Trigger 1 power must stay on, runs Timer A then activates relay. When Trigger 1 goes low, Timer A runs for time set on B and then closes relay. HOWEVER, see note below.

Trigger 2 power must stay on, runs Timer C then activates relay. When Trigger 2 goes low, Timer C runs for time set on D and then closes relay. HOWEVER, see note below.

HOWEVER, it also depends on how decimal is set on timers A and C.
A04.0 B07.0 C02.0 D03.0
A runs for 4 seconds, then runs for 7 seconds (Timer shows A during 2nd time)
C runs for 2 seconds, then runs for 3 seconds (Timer shows C during 2nd time)

A4.00 B07.0 C02.0 D03.0
A runs for 4 seconds, then runs for 0.7 sec, so it is like decimal was moved on B to B0.70
C runs for 2 seconds, then runs for 3 seconds (Timer shows C during 2nd time)

P-25: Toggle. Trigger 1 power must stay on, runs Timer A then activates relay. Stays activated when Trigger 1 goes low. Trigger 1 power must stay on, runs Timer A then deactivates relay.
(Same for Trigger 2 and Timer C)

P-26: Any change of state to Trigger 1 starts Timer A. If power to Trigger 1 is left on, Timer A will trigger when power is removed from Trigger 1. Or trigger can be pulsed with same effect.
Same for Trigger 2 and Timer C.

P-27: Same as P-23

P-28: Power on starts Timer A, then Timer C, Relay 2 activated, Timer A again, then Relay 1 activated. Relays stay activated until power-off. Triggers no effect. (I don’t understand this one’s purpose)

P-31: At power-on both relays activate, runs timers A, C, deactivates Relay 2, A, D, deactivates Relay 1, B, D, B, D, activates Relay 2, B, C, deactivates relay 2, B, activates Relay 1, D, A, C, A, D…. (I don’t understand the purpose of this one)

P-32: Loop. Trigger 1 power must stay on, activates Relay 1 runs Timer A, then deactivates Relay 1 while Timer B runs. Loops as long as power is on trigger.
Same for Trigger 2 and Timer C and D.

P-33: Loop. Same as P-32 BUT only needs a pulse for trigger. Next pulse to trigger cancels timers. Trigger 1 pulse activates Relay 1 runs Timer A, then deactivates Relay 1 while Timer B runs.
Same for Trigger 2 and Timer C and D.

P-34: Similar to P-31, except after activating and deactivating both timers, it stops. Triggers inactive.

P-35: Pulse to Trigger 1 starts timer A, then Relay 1 activates with Timer B, then stops until next trigger.
Same for Trigger 2 and Timer C and D.

P-36: Same as P-35, but Trigger power must remain on.

P-37: Pulse to Trigger 1 starts timer A and activates Relay 1, Timer B starts and Relay 1 deactivates, then stops until next trigger.
Same for Trigger 2 and Timer C and D.

P-38: Pulse to Trigger 1 starts timer A and activates Relay 1, Timer B starts and Relay 1 deactivates, Timer A repeats once more and activates Relay 1, then stops until next trigger.
Same for Trigger 2 and Timer C and D.

P-41: Trigger 1 goes LOW, Timer A starts and Relay 1 activates. Stops when Timer A finishes.
Same for Trigger 2, Timer C and Relay 2.

P-42: Trigger 1 goes LOW, Timer A starts. Timer B starts and Relay 1 activates. Stops when Timer B finishes.
Trigger 2 goes LOW, Timer C starts. Timer D starts and Relay 2 activates. Stops when Timer D finishes.

P-43: Same as P-42

P-44: Nothing seems to happen at all.

P-45: Power-on Relay 1 on for Timer A, then Relay 2 on for Timer B. Stops after Timer B.
Triggers not effective.

P-46: Trigger 1 HIGH, Relay 1 on for Timer A, then Relay 2 on for Timer B. Stops after Timer B.
Trigger 2 not effective.

P-47: Same as P-46.

P-48: Loop. Power-on Timer A on and Relay 1 activates, Relay 1 deactivates for Timer B. Timer C on and Relay 2 activates, Relay 2 deactivates for Timer D. Keeps doing this.

Work knives

I usually carry a knife at work as one of many tools. My last knife was a CRKT M16-12Z tanto style with serration near the handle. It served me well for many years until I used it to clean embedded moss from underneath my roof shingles. Before I realized what was happening, I had a nice fold-out butter knife. The abrasive shingles had ground away so much metal that the knife was just about useless.
knife damage

I talked to coworkers about knives they use, and two names stood out: Benchmade and Kershaw. I opted first to get a Benchmade Mini-Griptillian (blade alloy is 154CM). The lock on this knife was different than the CRKT, and the clip was oriented in the opposite direction (anchored on the butt of the handle). I carried it for about a month and it took some getting used-to having the knife in a different orientation than I had carried for years. But the blade was very sharp, and kept its edge well. It was smaller than I really liked, but had a good feeling handle. Another difference that I didn’t like was that it opened with just a thumb post, and had no index finger post.

Eventually, I found myself wanting a larger knife, and one that was oriented like my old CRKT. I looked at the Kershaw blades. I bought two, as they are far cheaper than the Benchmade. I bought a “Blur” and a “Clash” (blade alloy is Sandvik 14C28N).


I really like the shape of the Blur. It is pretty, a good size for my hand, and opens easily. The part I don’t like is that it only has a thumb post for opening. I also liked that the clip is anchored near the blade. This make drawing and opening the knife quicker and more natural for me than the other orientation.

The Clash is a similar knife, not as pretty to my eyes, but has an index finger post for opening. That makes it more natural for me, since my CRKT had both kinds of openers, and I used the index opener. The plastic handle is too slippery for me and that is a bit dangerous when closing the knife one-handed. A more rubberized handle would be welcome.

I find that the thumb posts invariably etch a hole in my pants. On my old knife, I took it to a grinder and removed about half of the thumb post on each side. I did something similar to one side of the Blur. I also feel like it would be easier to slash my thumb if I slip.

My ideal knife would be the Kershaw Blur with an index finger opener instead of a thumb post, and a more rubbery grip. I like the S30V version for the steel, but prefer a black blade.

The glass break on the butt of the Blur is a nice addition, which I wish were on the Clash. I’ve not seen anything online about retro-fitting a glass breaker.

I’ve chosen to carry the Clash, due to the opener and the orientation of the clip. The blade holds an edge well, perhaps not as well as the Benchmade which is made of a better alloy. I wish it had a more rubbery grip and spine, especially when I am closing the knife. It slips too easily in the awkward phase of moving the locking lever and closing the blade with one hand.

The assisted opening on the Kershaws is a nice feature when I’m in tight spaces, often up in a ceiling doing cabling work.

I use a small EZE LAP diamond stone to keep the edge and tip nice and sharp.

NOTE: One potentially dangerous part of the Clash is the way the blade locks with a “locking liner”. The vast majority of the time, this works very well. I was recently using the knife to clean out some fireproof foam from a pipe, and occasionally putting pressure on the BACK of the blade. Due to my grip touching the locking liner, the blade became free to close. I wasn’t injured, but it did get my attention. This kind of locking mechanism demands that the blade only be used with pressure on the sharp side until you intend to close the blade.

Garage Door Opener Radio Interference (and fix)

Our garage openers are both Linear LDO50 which uses 318MHz frequency for the receiver, and are about 12 years old. The security type is “MegaCode”.

We live directly across from a major airport, and not far from a smaller airport, so we typically get a LOT of radio interference with our garage openers. I can often see the receive indicator on the openers glowing solid red, or blinking rapidly for several seconds at a time. This causes the door to open slightly and then stop, and this repeats several times. Some days I just had to get out of the car and use a key to get in through a side door. The wall buttons for the openers always work, so I know there is no physical issue with the openers.

So I decided to try adding a new receiver that uses “Security+ 2.0” and 3 simultaneous frequencies (310/315/390MHz) since it is unlikely that all 3 will be getting interference at the same moment.

I bought the Liftmaster 850LM receiver that can control up to 3 openers. I also bought the compatible Liftmaster 892LT two-button remote controls.
NOTE: The remotes are able to use three different modes, and you must choose which one you will use.

1. I first set each remote control to use Security+ 2.0 instead of just 315Mhz or 390MHz.
Remote control

2. I plugged-in the receiver and pressed the Learn button for Channel 1 until the yellow LED turned on. I then pressed the left button of the first remote control to associate it with Channel 1 of the receiver. I repeated the process for the second remote. (Channel 1 can support up to 50 different remote controls! Channel 2 & 3 can each support up to 20 remote controls.) I then did the same for Channel 2, this time using the 2nd button on the remote.

3. I attached the receiver to the angle iron that holds the first garage door opener motor. I plugged in the power adapter and tied up the slack in the wires.
1st unit wiring

4. I used common telephone cable to do the wiring. I only needed 2 wires to each opener, so I cut off the remaining wires. I chose at random to use Blue and White. I chose to use Blue as the “common” and white for the other. No electricity is passing through these wires, it is simply providing “continuity”. That is, when the receiver gets an authorized signal, it essentially closes a switch that connects the two wires. The garage opener senses that the two wires have touched and opens or closes the door.
Receiver wiring
Closeup of receiver wiring

5. I stripped the ends of the wires and attached a short piece of cable from CH1 of the receiver to the Pushbutton and Common terminals on the first opener. I stapled the cable to the ceiling using curved staples for wires.
Closeup of opener wiring

6. I did the same steps to connect CH2 of the receiver to the common and pushbutton terminals of the second opener, this time using a 20ft piece of cable.
2nd unit wiring

That’s all there is to it. I tested each button on each remote and the doors opened and closed as expected.

I then erased the memory of the built-in receiver in both Linear units, and gave one of the new remotes to my wife, and took the older remotes, removed the batteries, and stored them in a box. Time will tell if this solution works well.

August 2020 UPDATE: It’s been a year and this new system has worked flawlessly. Previously there were days where the old opener was so jammed by interference that no amount of clicking would open the door. Now that is not a problem.