Pilot TV-37 Restoration

What the internet needs is another......


There's no shortage of Pilot TV-37 "The Candid" restoration webpages on the net... it's a popular TV with a small footprint, a unique "wow factor" of a 3" screen, and common enough that even the budget-minded collector can afford one if they hold out long enough.

I acquired this set at a radio auction a few years back in unknown condition, and was willing to pay market value since I could cash-and-carry. No need to worry about shipping (it never ends well). The upside was the glorified cardboard (masonite) case was in pretty good shape, the downside was the original knobs had been replaced with generics (more on this later).

Other Excellent TV-37 Webpages
Dave McClellan Resto - Very Clean Example
Jon Stanley - Basketcase Rescue w/ Plexiglas Cabinet
Joe Sousa's Restoration - Excellent Tech Info
Phil Nelson - One of the First TV37 Pages Published
Javier Albinarrate's Restoration - Example of Green CRT


  • The first TV to break the $100 price barrier, coming in at $99.50 in 1948; pitched toward the (well-off) college crowd.
  • The only production TV with a 3" CRT (and a penchant for destroying them).
  • A legitimately lightweight set. The perforated hardboard cabinet and plywood base weigh in at 5 lbs, 6 oz. And the all-aluminum chassis with full tube complement is 8 lbs 9 oz.
  • For real portability, a case with handle was offered in the style of what you might stash your accordion in. As one does.

"As Found" Chassis is clean, but dusty. Good filament!


Let's just say some "unique" cost cutting measures were taken to bring this set in on budget. Here are the highlights....

  • To cut weight, the entire chassis, HV cage and heat shield are aluminum.
  • Tube shields are typically used to avoid RF & IF oscillation and interference. This set has none of that.
  • And of course no power transformer, making this a hot-chassis design with one side of the 120V line tied to it.
  • This also means the controls must be isolated: this is done by mounting them on phenolic boards in oversize chassis holes.
  • There are 21 tubes in this TV and their filaments were wired by an accountant (see below).
  • The tuner is comprised of a pair of ganged variable capacitors similar to those found in a table radio; one handles channels 2-6, the other 7-13. The three dual-triode 12AT7 tubes used as RF amp, mixer and oscillator are "split down the middle" with one half being low side and the other half high side tuning. A triple-gang switch swaps B+ between the two halves as well as redirects the twin lead input.
  • There are 4(!) IF stages, stagger-tuned to eliminate the need for intermediate shielding. But no AGC; the contrast knob serves to control the gain.
  • The 3KP4 CRT is electrostatic; it is affected by the earth's magnetism and centering will change if the chassis is operated on its side.
  • There's 2,500 volts on three of the back panel control shafts, and they do indeed bite!
  • Two versions of this set were made; the earlier has 5 rear controls and a fixed R for linearity. The later (like this one), have 6.


  • A "hot chassis" requires an isolation transformer; consider interaction with line-operated test equipment as well.
  • If you have a good 3KP4, protect its filament with zener clamping (numbers below) or a separate 6V transformer before you do anything else.
  • Apply a few wraps of insulating tape to the Focus and Centering "discs" on those three rear controls. 2500V is present on the shafts as well as disc mounting tabs. You -will- manage to graze one at some point.
  • Do not operate the set with any tubes missing, even if they're not in the primary series chain. Doing so will over-stress the others.


  • Knobs for these sets seem to go missing. New ones are being reproduced by Renovated Radios, and used are available on eBay, but if you need a full set, you're looking at dropping a C-note with shipping.
  • This design seems especially sensitive to tuner and video amp tubes. Even if they test good, even if they work somewhere else in the circuit, I can wipe out the picture entirely by juggling NOS 12AT7's around.
  • The oddball series string wiring means that you can kill the heater in one half of a dual-section tube if the string is over-stressed. So just because a tube has a glow, look closer to see that both halves are lighting. 
  • You can see from the below pics that there are A LOT of ground points relying on screws and rivets to the chassis. All of these are a potential source for intermittent conditions.
  • There is a screw at the front center top of the chassis, directly under the CRT, responsible for grounding the RF coupling to the chassis. Even if tight, you can have no signal or worse, intermittent coupling which can cause sync issues, picture noise, random dropout, and generally cause things you can spend hours on getting nowhere with. As it's mechanical in nature, continuity also depends on temp and vibe. It also requires CRT removal for access. Lots of fun. See below.

Ballast Tube with Open Section "Before" "After"


Two series strings are wired in parallel (with one string having two 35v tubes in parallel). Then these two parallel strings are series wired to two more parallel-series strings, with v-dropping resistors sprinkled throughout.  But unlike an AA5, or series string engineered sets of the 60's, the typical current draw and cold-resistance of of these traditional tubes are all over the map. 12AT7's mixed with 6AU6's and 35W5's. Fail one filament and the entire system tips out of balance. Worse yet, the CRT takes a major surge at startup, leading to early failure. And being the only TV to ever use this 3KP4 CRT, replacements are rare and expensive. You'll find many of these sets running war surplus scope tubes with green phosphor (3KP1). Being a rather simple design, this was supposed to be a quick and easy restoration..... Whoops.

Early Series String Wiring - 3KP4 is in primary string. Later Series String Wiring - 3KP4 in secondary parallel.

SUMMARY of Issues Found / Work Performed

  • Replaced all electrolytic caps, out of tolerance resistors, and HV coupling caps for baseline.
  • No Sound: A random rectifier tube was found in place of 35B5 audio output.
  • No Sound: Audio output transformer had an open winding.
  • Lack of horizontal sweep: Open section of ballast tube, no B- voltage.
  • Distorted Sound after warmup with cabinet in place (~1 hr): Speaker cone distorts when warm; patched up tears.
  • Erratic Picture (vibe related): Cold solder joint on contrast pot lead from an old repair.
  • Erratic Picture (vibe related): Intermittent contact of IF and Tuner tube pins.
  • Erratic Picture (sig strength related): Recap tuned C's in Video Detector/Coil Can and tweak.
  • Erratic Picture (sig strength/thermal related): Replace 1N34.
  • Erratic Picture (vibe related): Bypass RF input grounding screw at tuner with lead to ground. See pic.
  • Erratic Picture (vibe related): Clean/tighten under-chassis screws serving as ground points.
  • Erratic Picture (vibe related): Clean/adjust contacts on high/low band switch.
  • Filament voltage spot checks out of tolerance (affects gain): Install/tweak series resistors for modern line voltage.
  • Pic Foldover with strong signal: Subbed 12AV7 for 12AT7 RF Amp to reduce gain.
  • Preventative: CL-90 Thermistor on mains to limit inrush.
  • Preventative: Pair of 1.5KE8.2CA zeners across CRT filament.
  • Other: Replaced Sound IF tube socket (terminal barely hanging on)
  • Cosmetic: Polished nameplate and clear coated it. Touched-up outer cabinet dings and scratches.
  • Cosmetic: Sanded plywood base edges, fix damaged corner and fill, glue de-laminating plies, repaint with low-gloss black rustoleum.
  • Cosmetic: Novus treatment for clear plastic over CRT. Clean and treat bakelite with Tung Oil.
  • Cosmetic: Solid wood front lacquer cracked, chipped, missing. Sand and repaint.
  • Cosmetic: Ordered a reproduction tuning knob, polished up the incorrect, but matched, set and reinstalled for now.

Warning: The Following May Not be of Interest unless you're restoring one of these...      Completed Pics & Vid at Bottom

Note: Two production versions of this set were made, with minor changes. However, the series string wiring is very different between the two. Schematics can be found at Phil's Old Radios. Correct HV-rated caps (including the one hiding in the HV cage) can be sourced from JustRadios.com.

After pulling the cabinet, I was happy to see all tubes were present, and even happier to see the CRT was a white-phosphor 3KP4. But all would be for nought if it had an open filament, a not-unlikely possibility. After isolating the tube and slowly bringing it up on my bench supply, I saw a nice warm glow in the neck. No guarantee that it wasn't worn out, but a good sign to press onward.

The under-chassis was clean with nothing out of the ordinary, though clearly someone had been in there before lifting and re-soldering leads. I also noticed the two cap sections of the B+ pi-filter were bent over and soldered together, bypassing the speaker's field coil. Typical when a coil goes open, but this one had continuity; odd.

Spot-checking tubes top-side, numbers matched up to the Sam's listing. All 4 IF tubes are 6AU6's. The 3 tuner tubes are 12AT7. The 4 V&H Osc. and Drive tubes are all 12SN7. Nothing too exotic here, but the 12AT7's are growing in value thanks to the audiophiles. However....I did find a 35W5 rectifier tube where a 35B5 should have been for the audio amp. Likely accidental, the two are physically identical and their sockets are right next to each other. Hope that didn't drive the last guy nuts. A quick search here failed to turn up a proper spare, so in went an order.

While waiting on that 35B5, I performed a recap (all film, electrolytic, and HV coupling caps), control cleaning and replaced any out of tolerance resistors. I also installed a pair of back-to-back zener diodes (1.5KE8.2CA) across the CRT filament to clamp voltage. A CL-90 thermistor on the AC line limits inrush and only drops a few volts, which is fine given today's higher line voltage. It adds a little to the warmup time, but better than the alternative.

This time the CRT sprang to life with a vertical line. But no horizontal sweep and no audio. This ended up being a lack of B- voltage. This particular model uses a 3-section ballast tube, and one 10-ohm section was open. I pulled the "tube" and went to bend the metal tabs up to open it for inspection, and surprise-surprise, somebody had already been there, too. This turned out to be an easy fix with a 2-watt resistor that fit neatly under-chassis in the filament string.

Now we've got a raster, but still no audio. Could it be something easy? Of course not. Turned out the output transformer had an open secondary winding. Rather than blow 50 bucks on a used Stancor or Merit equivalent, I cannibalized one from an old AA5 radio (with cracked case) with similar output stage plate impedance. Now we were getting somewhere, or so I thought.

If you really enjoy troubleshooting, get yourself one of these.
t's a 75-ohm transformer that's internally shorted. A good signal goes in, and no signal comes out. Connect it to the TV you're working on
for hours of entertainment!

No Video?

You've got a set with an oddball tuner and no idea if the IF stages have been tinkered with. You've also got an ancient 1N34 diode as the video detector. RF source seemed kosher, but nothing was getting through. All tubes in the signal chain test good, and subbing their order made no difference. But various other signal sources had no effect, either. It's the old repairman's dilemma.... Is it the TV or is it the test equipment?

Let's add additional uncertainty.... Out comes the B&K Analyst and a solid state portable TV for verification, both conveniently stored upstairs in another building.

B&K Analyst with the Hood Up Good Image with Video Injection Rebuilding the Video Detector Assm

Video injection resulted in a stable image, working up the IF chain was mostly OK. And after more time than I'd care to admit, I found the culprit. A defective, internally shorted, 75-ohm balun on the input. Made in China, of course. While that confirmed the tuner was functional, the pic was pretty rough, and adjusting contrast (gain) would cause the image to foldover at max signal. Even when gain was pulled down, the image was still jumpy and had sync issues. Wondering if this was a feature or a bug (a set of this vintage can do funny things with -too- strong a signal). I posted my question to a FB TV Group where there was, shall we say, no consensus.


Now we had good audio and decent video, but still sync issues and vibe-related pic breakup. For good measure I rebuilt the video detector assembly; some caps had drifted and while the 1N34 seemed to be ok, I wasn't confident that it was stable over the entire signal curve, especially at elevated temp. Remember, this set has no AGC. Also note, this is a germanium diode, silicon won't do. Jameco has replacements. With that rebuilt, I could continue up the IF chain. Pin cleaning and ground tightening helped a bit more. And eventually I discovered a cold solder joint on the contrast (gain) control that would only reveal itself with a sharp rap and careful examination.

Thermal expansion is also real... A number of socket terminals needed a slight tweak for a better grip on tube pins. This set also uses the chassis as its ground reference, and there are a ton of single lug ground terminals sprinkled throughout. If the lug used a screw, as found in the tuner section, I could usually cinch things up with half a turn. Rivets were checked for tightness, and an alligator clip to the chassis made quick work of hopping term to term to rule out the connection. If especially suspect, a wire was added, tied back to another known good ground. Remember, the chassis is aluminum and is more susceptible to expansion than traditional folded steel.

Headache Screw Intermittent A-Go-Go

Next was working my way through the tuning section. On at least one other webpage, mention was made of a screw on the tuner that completely prevented a signal from getting through. While its location wasn't specified, I'm pretty sure it's this one, circled in red. This screw grounds the visible lug pair to the chassis through an aluminum bushing under the phenolic board. It can not be accessed from the front due to the pressed-on tuning discs, and it would be madness to try to get the variable caps out of there for access from the rear. I suspect the plating on the screw (and decades of oxidation effect) also plays a part in the less than ideal continuity. I ended up melting some solder onto the top lug and running a wire directly to a good tuner ground through an existing chassis hole. Major improvement noted.

This tuner is unique in another way as well. There's a bandswitch to select between channels 2-6 and 7-13. Besides switching B+ in the front end, it also redirects the RF. Dirty contacts there cause problems, too.

Where I got lucky is with the IF alignment...I tweaked a can or two, but otherwise left things as they were. Some have reported difficulty aligning pic and sound, or heavy buzz in the audio, but this particular example is pretty well behaved in that regard. One change I did make, was to substitute a 12AV7 in place of the 12AT7 on the RF input to temper gain since this set will typically see a direct input rather than rabbit ears. For a particular interesting read on alignment, check out Joe Sousa's restoration, linked above.

Chassis and tubes cleaned; mounted to base. Prior to Sanding & Painting Brass strip mounted on finished base.
Ready for cabinet.

The final step was to spiff up the cabinet a bit. Rather than go to extremes on the masonite, I touched up the dings and edgewear and left the original patina alone. The lower portion was a different matter. The brown lacquer paint was chipping, and there was almost nothing left of the black paint used on the edges of the plywood base. Add to that a few missing chunks and I decided I better do it right. With wood glue, clamps and some carefully cut inserts I repaired the base, then after sanding, gave it a few coats of satin black rustoleum with a small brush. The control face was treated to a full sanding and I found a good color match locally (see finished pic above). The control strip is brass with brown paint infill. Rather than polish it to a golden shine which would look out of place, as well as risk paint damage, I hit it just enough to remove oxidation and get the text to pop, then sealed it with a few passes of clear coat.

Finished! 3" Screen - 6AU6 for Comparison Rolling Bars a result of Camera Shutter Speed

Return to Linearlook.com
Cory Heisterkamp 3/2024