31
\$\begingroup\$

In old movies (or new movies playing in these times), I often see people smacking the top of electron tube TVs or screens. Somehow it seems to help to stabilize/sharpen the picture. But why?

Are there any reasons an electronic circuit could go from not working as expected to working as expected from a sharp mechanical jolt? What are the conditions that cause such an issue in the electronics? It would appear that the Atari ST suffered from a similar problem that could be fixed by performing what became known as the "Atari drop" (as discussed in this section on wikipedia) although this is described as being due to loose connections, are there any other failure modes that could be remedied in a similar manner, and why?

\$\endgroup\$
  • 12
    \$\begingroup\$ Early designers created products that anyone could fix in their own home! Don't you remember the tube testers available in almost every grocery store? Housewives pulled and replaced tubes, whacked their TVs, and repaired their own sophisticated receivers back then. Most family members were even skilled at RF propagation and could adjust rabbit ears and rooftop antennas quickly for best reception. Sadly, such design and consumer skills have become a lost art today. ;) \$\endgroup\$ – jonk Aug 25 '16 at 6:37
  • 2
    \$\begingroup\$ There was another question that talked about this. It was because the circuits in these TVs had no board, so they were literally boardless circuits that tend to have components that eventually came loose. \$\endgroup\$ – Bradman175 Aug 25 '16 at 7:11
  • 6
    \$\begingroup\$ @jonk Ah, that brings back memories! I worked at a Tandy (Radio Shack to you Northerners!) which had a valve testing rig. Around once a month someone would come in with a small collection of glass tubes with the complaint "one of these doesn't work". I'd have (literally, seriously!) great fun wading through the manuals to get the conversions to set up the rig to test the device (including which oh-so-similar socket to plug it into) to say "Yay" or "Nay" on each of their offerings - including selling them the replacement, of course! \$\endgroup\$ – John Burger Aug 25 '16 at 11:01
  • 7
    \$\begingroup\$ The testing was free - but I got the reputation as the go-to guy to resolve the problem. "Could you come in on Friday when John's here?" To this day I have an affinity for the humble vacuum tube (I'm getting a Raspberry Pi Vacuum Tube audio amplifier shield soon...) - but I'll continue to refer to it as a "valve" for the rest of my life. \$\endgroup\$ – John Burger Aug 25 '16 at 11:03
  • \$\begingroup\$ Not sure on the off-topic. The "reason" does give great insight into early electronic design (point-to-point). On top of that the explanation for pots busting off old rust on the contacts can yield some important design considerations. \$\endgroup\$ – jbord39 Aug 26 '16 at 1:23
46
\$\begingroup\$

This practice is generally known as 'Percussive Maintenance'.

Any touching contacts, for instance in connectors, valves and their bases, and between the wiper of a potentiometer and the track, have a tendency to build an insulating film between the contacts. This happens most readily at higher temperatures, in high humidity, and when there is airborne contamination, especially sulphides. This can introduce higher resistance, non-linear behaviour, or break contact altogether. It can produce intermittent behaviour, changing with humidity, or voltage across the junction.

Sending a mechanical shock through the equipment can move the contacts with respect to each other, disrupting the film, and restoring contact.

In the case of a TV with a loudspeaker in it, sometimes the vibrations from the audio will change the contact state.

High contact pressures, and gold plating, improve the reliability of contacts against this sort of problem. TV valves, because they got hot, were especially vulnerable, and of course they aren't used these days.

\$\endgroup\$
  • 25
    \$\begingroup\$ 'Percussive Maintenance' is a sub-discipline of Impact Engineering. ;^) \$\endgroup\$ – Transistor Aug 25 '16 at 7:04
  • 9
    \$\begingroup\$ @Transistor interestingly, the percussive maintenance departments in larger institutions are nowadays mostly controlled by the anger management divisions. \$\endgroup\$ – Marcus Müller Aug 25 '16 at 10:14
  • 3
    \$\begingroup\$ @Transistor ...carried out using a P.A.T. (Precision Adjustment Tool, aka Hammer). \$\endgroup\$ – TripeHound Aug 25 '16 at 11:50
  • 1
    \$\begingroup\$ When preparing the bill for the client, it should be itemised as 'tapping it, 1 Kalganid, knowing where to tap it, 99 Kalganids'. \$\endgroup\$ – Neil_UK Aug 25 '16 at 16:18
  • 1
    \$\begingroup\$ Percussive Maintenance could also be considered a form of debugging. A good jolt could knock dead bugs out of the circuit. \$\endgroup\$ – Caleb Reister Aug 26 '16 at 5:04
19
\$\begingroup\$

It's worth noting that older TV's were constructed with point-to-point electronics soldered by hand, lacking a firm place for the components to be anchored to, such as this image showing the underside of the chassis of a 1948 Motorola VT-71 7" television.

Point-to-point wiring (image taken from the wikipedia page for Point-to-point construction)

From the image alone it is clear why a solid whack to the box could provide enough energy to move loose components and bring them to rest in such a state they go back to working correctly. It could be down to exposed conductors touching, or a poor solder joint being loose and the movement provided from a solid hit moves the offending components back into a position that they work again - by either breaking connections that shouldn't be connected, or by doing the opposite to the components that should be connected.

\$\endgroup\$
  • 6
    \$\begingroup\$ It would be a good idea to provide credit to the author of the photograph, including a link to the source. Doing so may be a copyright requirement for use of the image. \$\endgroup\$ – Makyen Aug 25 '16 at 10:48
  • 5
    \$\begingroup\$ What a mess!!! I'm surprised they were able to keep those exposed wires from shorting to each other on a very regular basis (if not constantly). I barely see any insulated wires \$\endgroup\$ – jbord39 Aug 25 '16 at 13:47
  • 4
    \$\begingroup\$ @jbord39: That's because there are very few wires there. What you're seeing are the bare component legs. It was common to have screwdrivers with insulated shafts for adjusting trimmer caps and chokes in such sets, to avoid shorting things out. \$\endgroup\$ – TMN Aug 25 '16 at 14:26
  • 6
    \$\begingroup\$ @jbord39 It only looks like a mess to the untrained eye. With proper layout (as we see here) the component leads do not touch one another. They are short, stiff, and soldered to numerous fixed contact lugs. If you look closely you will see that in a few places where a component has more freedom of movement an insulating sleeve has been slipped over a lead. Because everything is open and visible working on these things is a real joy, at bit like working on an old pickup truck as compared to a modern compact car. \$\endgroup\$ – David42 Aug 25 '16 at 16:43
  • 5
    \$\begingroup\$ manufacturing these must have been such an ordeal! On comparably complex manually assembled PCBs, you'd have silk screen to tell the assembly line worker exactly where that resistor had to go – I can only imagine people having a printout of a plan, highlighting the components they ought to solder into place in their manufacturing step, sitting in front of a sorted box of components and praying they don't accidentally mix up these two solder lugs in the lower right corner at each of which 8 component leads end.. \$\endgroup\$ – Marcus Müller Aug 27 '16 at 7:34
9
\$\begingroup\$

Bad connections, corroded vacuum tube sockets making poor contact, cold solder joints and so on could sometimes be temporarily mitigated by the judicious application of 'percussive maintenance' techniques.

\$\endgroup\$
  • \$\begingroup\$ Cold Joints, Dry joints ? Lifted PCB tracks... \$\endgroup\$ – mckenzm Sep 7 '16 at 0:56
7
\$\begingroup\$

This was not just a tube-era practice. In transistorized-era displays (ie, where the only remaining vacum tube was the CRT itself) the target of percussive maintenance that could occasionally get more use out of a failing monitor was most likely magnetic components such as transformers, chokes, the flyback, etc. These tended to be physically large and massive so could develop intermittent connection to the PCB as a result of temperature cycling or shock, and also to be composed of parts such as core laminations which could loose mechanical fixation if their cementing compounds failed and audibly vibrate, often to the detriment of image stability.

\$\endgroup\$
2
\$\begingroup\$

The contacts in the drum/turret tuners were a big contributor. [Turret tuner from antiqueradio.org[1]

\$\endgroup\$
  • 2
    \$\begingroup\$ You "fix" that by rapidly rotating the dial on the front, not by hitting the whole case. Same for dusty potentiometers in other equipment (like volume dials). \$\endgroup\$ – Peter Cordes Sep 3 '16 at 13:30
2
\$\begingroup\$

I am no expert, but I remember when I was a kid, doing this to my PC 486's 14" monitor (brand was Leading Edge). at some points of time, first appeared to be randomly, the display was going bananas, showing something like C64's loading screen colors but with much higher frequency (it did not matter whether I was in the black DOS prompt or a colorful game!). The only solution I knew as a kid was to punch the damn thing, and after few punches everything was back to normal.

Later in time, even kicking the damn thing was not helpful, so I called my dad to have a look, he found out that the 15pin RGB at the monitors end is lose from inside the cabinet, soldered it back again and it was fine and dandy again.

I have no experience with old CRT TV's, but my guess would be the punching momentary fixed the position of some lose passive components e.g. caps/resistors/inductors or wires. not sure even if a calculated punch could tell electrons whether or not show clean or distorted picture on the screen.

\$\endgroup\$

Not the answer you're looking for? Browse other questions tagged or ask your own question.