Subwoofer Controller with “Sensurround” Earthquake Generator

October 2018 update

  • An exciting development:  A very helpful guy called James (see comments below) has given me a few great ideas and some modifications to the “Sensurround card” have been made to his advice.  One was a mistake, but that’s OK.  Others were worthwhile.  However upon listening with headphones and viewing the oscilloscope while “twiddling” the adjustments pots, a limitation of my 15 Volt supply has been identified.  The original patent required a 20V supply to the analogue filter section.  With only 15 Volts, the output could clip – not just on adjusting the “output” pot (an attenuator), but on adjusting the frequency pot at the digital section.  Moreover, the irregular and very high amplitude transient low frequency effects that increase upon downward adjustment of the frequency pot of the digital section could clip the analogue output section unless “rolled back”!  Rather than changing the voltage to 20 per the patent (which was too hard), the headroom of the op amp was increased by lifting pin 11 from GND and attaching it to the available -15V output of the PSU!  30V is within limits of the device.  More below including a sound file made directly off the card.

This wonderful box of tricks was put together a few years ago to control the input signal to my home cinema subwoofer amplifiers.  It is mono and houses two ESP power supplies, a 12db/octave booster, an 8-band subwoofer equaliser and an optical output power limiter/compressor (all ESP projects).

The subwoofer output from an Integra DHC 9.9 decoder is connected to the input socket.



A balanced line driver and XLR output socket were added since the photos were taken.

The best thing about the project is the pseudo-random earthquake noise generator circuit used as a “test rumble” feature to impress my friends. 🙂

It is activated at the flick of the left rotary switch which powers the circuit board and triggers an “input selector” relay.


I built the circuit back in 1989 from the diagram provided at Fig. 10 of  US Patent 3,973,839.  This was the first Universal Studios “Sensurround” patent and the circuit is supposed to be what created the earthquake rumble effect in each cinema in which Earthquake was shown in Sensurround in 1974/5.  Control tones on the film switched the circuit’s (continuous) output into the subwoofer amplifier bank (BGW or Cerwin-Vega power amps) during the movie and varied the gain of the preamp as required.

The circuit did not work because Fig. 10 was full of errors – whether deliberate or not I do not know, but I have my suspicions and suspect that the patent was invalid for want of enabling disclosure.  Anyway, it sat in a box for 17 years and I dug it out for this project.  I did not have the expertise to identify the errors, but Rod Elliott who is much smarter than me was able to model it with software and identify the problems.  That got the thing working, but it was very unstable and would run for only a few seconds before stopping.  Applying power again usually got a few more seconds out of it and adjusting the input voltage and changing a few resistors extended the time to about a minute before stalling, but it was pretty hopeless.

With a little lateral thinking I finally figured out what the stability problem was and substituted an unbuffered NAND gate chip onto the card (modern buffered types were not around in the 70s!).  It then worked continuously without falter and cranks out an awesome rumble with great transients extending just into the infrasonic range.   Rod didn’t believe that my substitution could make any difference, but confirmed that it did and here is a frequency response graph of the electrical signal that he took straight off the card:


A video of the raw waveforms straight off the card (digital and analogue):

And one of the analogue waveform after the 12dB/octave booster:

Here is a spectrum of the boosted signal (note the taper off as frequency increases):


It works so well with the big subwoofers that it causes the windows in a friend’s house in the next street to rattle all at once!  When calibrating it, my immediate neighbour thought there was tunneling work under his house!  It is therefore only activated in the middle of a week day when people are supposed to be at work or playing golf or something.  🙂

October 2018 update

This page has been up since 2009, but the circuit diagram was further corrected with the help of James (see comments below).  This is the final version:

Note the 2009 wiring changes around Q1 as suggested by Rod were correct, but my lower input voltage (15 compared to 20 from the patent) meant that the 12V Zener supply resistor value should have been dropped from the original 1K to 390Ω to provide the required current to the digital section.  Some changed resistor and cap values remain.  Each CMOS and the op amp chips also have 100nF ceramic power supply bypass caps and there’s a 100μF electro added across the 12V Zener.

Card with these changes:

Oh – I forgot.  It now has a TL054 in place of the LM324.  It’s a much better op amp, so I thought why not?  Well I wouldn’t have rushed out to buy it, but I had about a dozen just sitting there in the stash.

And dodgey (mostly 1989) soldering, but 5 ceramics are shown:

More importantly, an increase in dynamic range has been achieved by adding a dual rail supply to the output filter section.  Pin 11 of the op amp is now connected to a -15V supply rather than GND.  A 56Ω resistor bypasses the supply.

The above circuit is correct, it works and is 100% stable but only when an unbuffered CD4011 chip is installed.

Here’s an audio capture that I made straight off the modified card with Audacity.  It starts off as a kind of “Galactica fly-by”, but as I turn the frequency pot it transitions into that awesome bone-crunching “Earthquake” with randomly spaced LF transients where it’s best at around the 15 second mark.  The trimmer is turned beyond this sweet spot where it sound’s kind of dumb, but I included it for “full disclosure”. 🙂

rumbles 1  ← Download mono WAV.  Play this at your own risk!  The management assumes no responsibility for the physical or emotional reactions of the individual listener, nor for any damage to anything that it might cause!

Rumbles 2  ← Download stereo FLAC.  Ditto!  ← This requires a proper browser else it might produce a screen of garbled fonts.

Here’s my own FFT showing that the card has around 105dB of dynamic range:


Original Mod 3 Sensurround control box:

44 thoughts on “Subwoofer Controller with “Sensurround” Earthquake Generator

  1. Very good, I have been looking at building one from the circuit diagram in the patent but seeing this tells me it will obviously not work.
    Do you have a circuit diagram or circuit board layout that I can use as I am very keen to build one like the one you have here, to go with the low freq sub I am building.

  2. I don’t have a PCB layout, but a friend is working on one and another friend in the States has an original card in a Sensurround control box that I have asked him to photograph. I will upload the corrected circuit diagram to the page later. Do not use a modern buffered CD4011 chip which causes circuit instability. Use an old inherently unbuffered chip or a new one with the UB code.

  3. O.K thanks, I did try simulating the patent verison a while ago on a computer and it did near absolutely nothing so I thought there must be errors in it like you say here.

    The sub I am doing is designed to have 250W RMS pumped throught it to a minimum frequency of about 23Hz ported so it should get a good amount of bass out of it even if it does not rumble at all like the original. Its only using a 10″ driver as I don’t have the money or space for anything bigger really.

  4. I hope you have some fun with it! Try sealing the port and pointing the driver into a corner for better results.

  5. After I have done the rumble generator and the sub, my next project will be the control box. On the Earthquake DVD, on the mono ‘sensurround track’ all the .1 is is the control tones, so I will build a suitable control circuit to exploit them, which actually will proberly not be as difficult as it sounds as it fairly simple, look on page 4 of the patent. The only bit I will not be able to do is the ‘step gain’ as it raises the main volume and obviously I can’t do that and much to my annoyance, on the DVD Universal don’t seem to have made any attempt to implement that feature, even though all it is is raising the gain by up to 8 dB at some points so they sound louder and fit in with the rumble effects.
    To get the partial surround effect the rumble originally gave properly, 4 subs like the original would do it, but do you think I will ever have that sort of space or money? Pretty unlikely.

  6. Yeah the “Sensurround 3.1” option is silly on that DVD, but it could theoretically be used with a DIY controller + rumble generator. The step gain for both the rumble track and the main tracks ought to be possible to implement by hooking the gain control circuit to attenuating resistors after all the relevant output channels. However, I would not bother as the Dolby Digital 5.1 option on the DVD has the exact same rumble in the subwoofer channel with gain at the right places. You would be better off making a Sensurround Mod 2 or 3 controller to steer the rumbles from front to back in Midway, Rollercoaster and Galactica. 🙂 An original Mod 3 unit came up on Ye Bay a month ago. I was the underbidder. 🙁

  7. Its unfortunate that MCA never gave any documentation on what they do the original mod 1 units to upgrade them to 2 o3 or else building a mod 2 or 3 unit would be easy. Can’t find any diagrams on it anywhere, just the original mod 1 verison on the patent.

  8. I have a friend in the States with all the circuit diagrams and will contact him later for you.

  9. Thanks for that.

    Is it not true though that mod 2 and 3 units cannot run earthquake as they don’t work on control tones?

  10. Alright, my last comment is slightly wrong, although from what I found they were able to print the low freqs on the optical track it still used control tones to steer them from back to front and so on.

  11. Thats right. In Mod 2 and 3, the control tones were used to steer and to particularly good effect in Galactica with infrasonic (only just) signals to simulate a sensation of wind. I think the Mod 2 and 3 still had a rumble card, but it was only used to calibrate the system in each cinema.

  12. From what I have read the rumble card was removed as it was no longer needed due to MCA finding a way to print the ultra low freq on the optical track and practically all the existing control boxes were jury rigged to mod 2 and 3 standards to support the new tracks and dbx NR which made the rumble cards unsupported and as so incapable of playing Earthquake.

  13. You could be right. I am not sure, but I read somewhere that the cards were left in solely as a speaker calibration/setup featrure. I will upload a photo of a Mod 3 recently sold on Ye Bay. The card at the left looks suspiciously like it might be a noise generator because it has a bunch of caps and the white thing looks like the exact same style of trimpot (the circuit has 2) as in my Cerwin-Vega A-3000I amps of the same era.

  14. The original CD4xxx CMOS ICs essentially acted like switches—an output-port signal could be read at the input-port if you wanted. This was multiplexers and some simple parts. Flip-flops, counters, and the like cannot do this.

    The newer parts are “buffered”. That is, the outputs are run through amplifiers, called buffers, to allow those parts to drive more inputs, and capacitive PC traces. The amplification process prevents output-port signals passing through to input-ports.

    “Unbuffered” parts are essentially original parts, made in newer—faster, but with lower voltage tolerance—production processes. Gotta read the fine data-sheet!

  15. That’s all very well, but the circuit worked for up to a minute or so with a modern buffered 4011 feeding itself. Anyway, you can see the “UB” code for “unbuffered” on the “modern” replacement 4011 in one of the photos and it works continuously for whatever reason.


  16. R7 should connect to IC-1 pin 4, not pin 14. This will enable that secondary oscillator for additional rumble.

  17. The R7 fix enables low frequency background rumble that is supposed to modulate
    random earthquake rumble using Q1. Likely you’ll need to swap back R6 and R25 values the way they were in the original schematic.

  18. James thank you very much. I haven’t even looked at this in years! I will ask my friend Rod who identified the error in the schematic to pull up his old simulation and try those changes to see what happens. As it is at present though, it does sound pretty much the same as what’s on the blu-ray.

  19. Typically, the inventor submits a sloppy hand drawn diagram, and the patent agent (who is an attorney) redraws it nicely for the patent. That’s where errors creep in. This inventor didn’t proofread the patent diagrams. This is the reason for the wrong connections.

  20. I asked Rod and he still had the simulation on his PC from 10 years ago and tried your suggestions. R6 and R25 must remain as I have them. Connecting R7 to pin 4 sets up a 7MHz oscillation in the simulator which he says might be a simulator problem as it looks like it should add a low frequency generator. He said I could try the pin swap, but not to do the resistor swap. Any ideas?

  21. Sounds like a simulator bug. R7 should be disconnected from power pin 14, and connected to NAND output pin 4. Other end of R7 stays connected to pins 5, 6, and the cap. Then it should oscillate around 1.5 Hz.
    Just try the pin swap.

  22. I’m a retired Patent Attorney! Of course I NEVER accidentally inserted such errors throughout my career. 😉

  23. Just finished skipping through Midway and Earthquake DVDs.
    Earthquake has Sensurround on 5.1, and Midway nothing.

  24. There were different DVD releases of most of the Sensurround films. I have one of Midway that accidentaly has the control tones on it. Very strange. The blu-rays seem to have good rumbles on the LFE channel on the ones I’ve got anyway.

    I’m kind of excited about trying the new R7 connection. Ever since getting this working, I’ve had in the back of my mind that it doesn’t sound exactly as I recall it from my childhood at the cinema in which the effect was kind of more “random” and “modulated” for want of better words.

    Thanks again for taking the time. It is really much appreciated.

  25. I saw Earthquake as a teenager in 1976, and I remember this “bung, de bung, bung” in the theatre. Didn’t sound the same on the DVD last night, but it still was cool. I think 1st generation rumble like what is in the patent (when working) is probably the real deal. Reading, it seems they had trouble with the rumble generator and later started substituting recordings. That’s probably what’s on modern soundtracks.

  26. YES! That’s it – a “bung, de bung, bung! LOL 🙂 In Earthquake at the cinema (as I recall it from 1975 aged 12) it was a kind of random modulation in the rumble’s amplitude with slight gaps. It was not like that in the later films in which the generator was not used, but rather DBX II dynamic range expansion was applied to recorded LF effects. It is also not like that on the blu rays! On all the consumer releases of Earthquake, the rumble is more akin to a subwoofer calibration signal. This is getting really interesting.

    What is your background? How do you know about the R7/pin 14 error, or did you just work it out?

  27. “Well, you have to know these things when you’re a king, you know.”

    Just looked at the schematic, and it was a one shot without the R7 fix. So it seemed clear it had to look like the other NAND oscillator, and Q1 was a modulator. The secret of Earthquake Sensurround sound is two separate oscillators mixed together.

    I also know why the early Sensurround generators failed, and why your buffered chips failed. You’re doing naive audio frequency style wiring. No circuit board with RF ground plane. These CMOS chips switch at RF rates. Unbuffered CMOS has relatively low gain, so are more forgiving of such wiring style. But high gain buffered CMOS breaks into MHz RF oscillation fairly quickly. The way to fix this is to add 0.01 uF (or larger) ceramic or mylar RF bypass capacitors across power and ground of each CMOS chip. Pins 7 and 14 of CD4011 and CD4030, and pins 8 and 16 of both CD4015s. Keep leads short and the cap near each chip. It will be much more reliable after doing this.

    Likely the original Sensurround rumble generator boards were laid out by audio guys who didn’t understand RF, so they were unstable and failed in the field. This is why they went to the bland recorded sound in later generations.

  28. For those following this blog, this is what Ian and I learned:
    Good layout practice requires CMOS chips have local RF bypass caps from power to ground to avoid RF oscillation. Ian didn’t test it, but likely the circuit would now work with buffered CMOS.
    Running on 15V required changing R24 to 390 ohms, and adding a 100uF cap across CR1.
    My idea about moving R7 to make a second oscillator was totally wrong. All it did was cut the earthquake sound on and off at half second intervals.
    Reading the patent, and then the pseudo random noise generator paper it refers to written by Don Keele, it became apparent that R7 original connection was correct. The NAND gate pair and Q1 are used as a one shot on power on to seed the random noise generator. You can see this as the push button on page labelled 36 of Don’s paper.

    Sensurround patent:

    Don’s paper:

  29. Just an update and thanks to James. It’s working flawlessly with greater dynamics than ever with the addition of a -15V rail to the quad op amp. That allows the frequency trimmer at the digital section to be turned down where surprisingly the random transients increase significantly in amplitude. With the previous single rail supply to the op amp, the output could clip at the low frequency range, so you’d have to raise the frequency to eliminate it. The clipping was not visible on the oscilloscope, but audible with headphones.

    I saw all of the Sensurround films in the day and now realize that Earthqualke had the frequency set very low, whereas Galactica (which did not use the generator in the cinema, but recorded it on the soundtrack for DBX II expansion) was at a higher frequency set point for the general Galactica fly-by scenes. Only for the big super explosions (like when a Cylon Base Station went “kaboom”) did they lower the frequency of the recorded generator.

    Back to the circuit however, despite the addition of 100nF ceramic bypass caps across the supply pins of each CMOS chip and a 100uF general bypass electrolytic to the digital section’s 12V supply, I’ve confirmed that the CD4011 absolutely must remain as an unbuffered “UB” type such as the old Motorola one installed with the “MC14011UB” code. On Semi seems to do them now. I tried a standard “HFC4011BE” from ST Micro today and the output lasted no more than half a second before stalling – that’s if it started at all. See TI application note here:

    Also as to the addition of the decoupling caps, they made no apparent difference as there wasn’t any HF oscillation in the first place. HF oscillation would have shown as blurred traces when ‘scoping the digital section. Nonetheless adding them can have done no harm.

    Oh and BTW – nobody follows this blog. 🙂

  30. I have been lately!! Thanks for posting the Audacity File. I have gotten hold of Cerwin-Vega Sensurround Subs, I plan on using them in a play to reproduce the earthquake feel in a production.


  31. Better put up a “Your attention please ” Sensurround warning to get a laugh out of the audience. 🙂

  32. I’ve seen some Sensurround film can that appeared in a recent video of a closed down ABC, UK cinema I doubt if the test film tones in the can is even playable given the state of the room it was found in.

    I heard rumors that cinema placed the CV Sensurround speakers behind a false wall after it was retired. Its just a rumor I heard during the 1990’s a projectionist told me this.

    Sensurround would need the actual speakers as its my belief they would give a certain frequency signature.

    What EQ did they use with Sensurround if any was used in the cinema? Or was it just passive crossover networks with bass bins and HF horns?

    The Laserdisc and (bluray is okay for 5.1)

  33. I don’t know about signatures, but do know that in 2006 the Schauburg Cinema in Germany had a Sensurround festival in which large banks of ordinary JBL speakers were used. To the original effect? I doubt it. As to other questions, the patent document might have your answers.

  34. A Sensurround unit was sold year before last at a steep price $10 grand that I was thinking of but was bidding war I didn’t want get into. I think the owner should have kept it rather than selling it.

  35. Hi, I have a question. I also constructed a generator according to the original diagram. Why is the value of resistors R25 – 33K and R6 – 2.2K in your connection? In the original connection, it is the opposite. I supply 24V circuit and it works fine. The only error in the original diagram is the connection of the base – Q1 from the supply voltage (CR1), which I corrected.

  36. Hi. I have a very limited understanding of the circuit. Changes made around Q1 including the resistor switch that you mention were made in accordance with an analysis done by Rod Elliott over 10 years ago. I could phone him if you like, or you could ask him directly via his email address on the ESP site. I’d guess he probably still has the circuit in a SIMetrix file. If you design a PCB for this, I’d be keen on buying a couple from you.

  37. Ok, I understand. I thought you knew more about the changes in the circuit. With the value of the resistor R6 – 2.2K, more current flows into the base Q1. If it survives, it doesn’t matter. Works only in switching mode. I did not use the original (2N3643), but replaced it with type BC847, which I had in stock. In the original circuit, I only changed the R6 to 47K. If you are interested, send a question to my email. I can send back a photo of my test prototype PCB.

Leave a Reply

Your email address will not be published.