Tube Television Tennis

'Pong' made with Vacuum Tubes

Similar projects worth following
I was given a draw full of old radio valves by a friend and he half-jokingly suggested I could make a pong game out of them. I didn't think it would be straight forward initially, but rethinking the way the circuit works made me change my mind. The circuit is made up of astable, monostable and bistable multivibrators.

The whole thing is made up of multivibrators (aka one-shots and flip-flops). The idea is I start with sync oscillators for horizontal and vertical and trigger one-shots off them, then use diode logic to combine the output pulses of the one-shots.

For example, to generate a vertical line, I start with a horizontal oscillator which produces a pulse acceptable to synchronise a television. The pulse triggers a one-shot who's delay can be varied with a voltage. When the pulse ends, it triggers a very quick one-shot which generates the line on the television. Varying the length of the intermittent pulse moves the position of the short pulse on the screen.

The same can be applied with a vertical timing. Using diodes, the outputs of the one-shots can be combined to produce a paddle or a ball, depending on the length of the pulses, by logically ANDing the outputs.

Currently the circuit works off a 100V regulated power supply (using tubes again) and draws about 30mA. I'm expecting the completed unit to draw over 300mA on B+ and 6A for the filaments.

I'll update the schematic as I go at

Some reading material I found useful:

  • Steady Progress

    256byteram08/30/2014 at 09:59 0 comments

    Two of them now!

    It's a little wonky but I haven't filtered the B+ supplies from the various sections. The two objects are completely independent but I wired the controls together to see how well they track each other. So far the left one (the original) is slower than the right one. This is because of small differences in the components used, including the tubes. 

    I measured the current consumed on the B+ rail. At 82V it drew 44mA. That's about 3.6W. Remember however the voltage regulator dissipates power and so do the filaments on the tubes. There's about 100V across the regulation tube so it's throwing out  4.4W and getting quite hot. In fact the tube I was using (a 6GV8 triode/pentode) was getting a little too hot so I wired in a separate pentode (6CM5, a common Australian horizontal output tube) in place of it, keeping the triode in circuit.

    The filaments draw 3.75A, excluding the regulator filaments which have their own supply. The voltage is 8.2V, higher than what I'd like it to be. The transformer has taps on it for different line voltages. That makes the power consumed by the filaments 30.75W, and the overall consumption about 38.75W. Given the television chassis the mains transformer was taken out of would have taken more than 100W, I'm quietly confident the transformer won't let out any magic smoke in a hurry.

  • Chassis

    256byteram08/28/2014 at 12:04 0 comments

    Of course, having everything attached to a chassis of some sort is much easier than mucking about with loose components. Here are some pictures:

    The three tubes to the left generate horizontal and vertical sync signals and amplify the video signal. The four to the right generate an object on the screen, which can be positioned with the two pots to the right of the chassis.

    There will be four objects in total: Two paddles, the ball and the playing field. The ball is logically compared to the playing field; if it falls outside the field vertically, its vertical direction is changed. If it hits a horizontal edge a point is awarded to the other player. If the ball hits a paddle, the horizontal direction of the ball is changed. The spaces that haven't been completely drilled out will be used for game logic (bistable flip-flops etc). I didn't drill them out because my drill battery went flat...

  • Vertical Sync

    256byteram08/20/2014 at 14:28 0 comments

    I scrounged enough working valves (barely) to get vertical sync and positioning working, just while I wait for things to come through the post. The valve count has risen to seven. This is just to prove to myself the idea will work in some way. Getting a decent picture out of the diode logic that connects everything is tricky, as can be seen by the vertical line that follows the paddle. There was much fiddling with resistor values.

  • Beginnings...

    256byteram08/18/2014 at 13:38 0 comments

    So far I've got a vertical line on the screen. I've only got a few valves of the right type so I'm limited with what I can do so far. I've got 20 old Remington Rand tubes in the post (type 2033's, similar to the 6J6) so I'll make a chassis and add things to it when they turn up.

    Here's a video of my as-always messy setup:

View all 4 project logs

Enjoy this project?



helrshes wrote 03/26/2021 at 12:14 point

yes its really interesting project I am also working on a tennis shoes project here

  Are you sure? yes | no

Mirkóczki Máté wrote 01/28/2021 at 16:54 point


The link for the schematic are not available.

Please send a new one.

Thanks! :D

  Are you sure? yes | no

greenaum wrote 04/19/2020 at 18:25 point

Do you find yourself doing the "bip!... boop!" noises as you move the paddles? I would!

Please don't abandon this! It might well be unique in the world. You're on the right track, just need a couple more object generators and a few diodes to connect them together. I imagine you'd detect collisions between bat and ball, by seeing if both circuits are outputting ("drawing") at the same moment. Then latch that for the frame.

It's also charming to see such a thing, on an old B&W set, being filmed for Youtube on a camera, through a computer, that contains more than a billion times as many components (inside the microchips) as the Pong has, and yet is more reliable!

I'd love to see this finished, and I bet it would get a lot of attention at retro-gaming exhibitions and parties, even if it's not as old as it looks. It also would show that you don't need thousands of parallel processors spitting out raytraced graphics to have a game that's fun. This one would have, what, 3 polygons? All rectangles. And a few more if you added a playfield.

You seem to have done most of the work, with creating the chassis and the power supply. Since you have a bat that can move X-Y, shorten it a bit and that's your ball circuit. You're nearly there! Just needs those last few valves!

I'm fairly sure I saw a diagram for the General Instruments Pong chip online, some time in the past. Though you don't really need one, it's easy to deduce just from observing the game and realising it is mostly just timers, triggered accurately off other timers. I once wrote a software "emulator" of Pong, using the principle of timers (except I used counter variables cos it was a program). Using timers and checking them on each "scanline", rather than the more conventional approach of just using sprites and conventional game logic. It was a software implementation of exactly what you've discovered.

I'd compare the current scanline number against the controller for the left bat, say. Then if it was equal, set off another short timer (counter) for the horizontal portion of the left bat. As long as this counter was greater than zero, the bat would be drawn on that line. The counter would decrement each line, like a timer would decay.

I only did that vertically, didn't bother implementing it for horizontal drawing, that would have been a needless pain. But it produced a very accurate looking Pong game. And even had the same, I don't know how, weird "bug" of when you move the bat off the top of the screen, it sort-of pops up at the bottom again but much too large. Something to do with overflows I suppose but looked a lot like the same sort of error you'd get in the real analogue thing.

I think there's plenty of science museums, etc, would gladly take your Pong if you didn't want to keep it yourself. It's that cool! 

  Are you sure? yes | no

Dr. Cockroach wrote 07/02/2018 at 22:35 point

Oh man, I love it. My hats off to your work on this :-) I started out with V-tubes back in the late 60's but have not used them since the 80's :-D

  Are you sure? yes | no

robot797 wrote 01/11/2016 at 21:29 point

how is the progress so far?

  Are you sure? yes | no

robot797 wrote 08/19/2014 at 23:24 point
thank you
i will follow this project
i am realy intreged

  Are you sure? yes | no

robot797 wrote 08/18/2014 at 22:31 point
will the circuit be afailible when its done?

  Are you sure? yes | no

256byteram wrote 08/19/2014 at 08:53 point
I've added a link to the current schematic in the project description. It's only a draft but I'll change it as I go.

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates