Monday 1 February 2016

Geoffrey Roe: The Annular Discharge Silencer



The following blog hopefully gives an insight into the research and development work carried out in the early 1970’s into the efficient silencing of motorcycle internal combustion engines.

Dr Geoffrey Roe MA PhD and his colleague Dr Terry Thorpe Bsc PhD were both Academics in the Department of Mechanical Engineering, University of Manchester. They worked as a research and development team within the Simon Engineering Laboratories and made their expertise available to the worldwide motorcycle industry. 

Geoffrey Roe and Terry Thorpe



The team were involved in many diverse aspects of motorcycle design and development, from suspension and road-holding issues to engine design and tyre contact area optimisation. However it is the concept of the annular discharge silencer that concerns us here today.

Geoffrey testing in the anechoic chamber




Geoffrey Roe had long been interested in the understanding and performance of engines and in 1974 published as co-author with Walter Annand, a definitive book on the subject: Gas Flow in the Internal Combustion Engine. In his research he had found that in passing exhaust gas through an annulus and over a series of various capacity chambers, it was possible to effectively silence a single cylinder engine without a serious loss of performance. Furthermore it was possible to ‘tune’ these internal chambers by altering the volume and surface hole configuration, in order to dampen the specific frequencies required.
Geoffrey was a keen motorcyclist and rode a Norton Commando at this time. Terry had made the better choice and rode a Velocette Venom.


In June 1973 Geoffrey published a paper entitled: The Silencing of a High Performance Motorcycle. In it he describes how the noise level of a 750cc Norton motorcycle was reduced from 98 decibels to 86 decibels by silencing of both the induction and exhaust. This was quite some achievement considering the mechanical noise emanating from the air-cooled engine. Later under MIRA testing it was found the antiquated engine produced more noise than the exhaust!

Geoffrey’s initial idea behind his annular discharge type silencers, was that noise reduction by viscous damping could be achieved by passing the exhaust gas down a channel with a high width/height ratio, with height being very small compared to the wavelength of the sound to be suppressed. In order to turn this into practice, a small peripheral (annular) gap between two co-axial cylinders was used. The peripheral volume would also act as a quarter-wavelength resonator (closed end/open end), which was hoped would balance the half-wavelength resonance of the feedpipe. With sufficient width, the flow area of this type of silencer could still be maintained to at least equal to that of the feedpipe from the engine. For perfect mutual cancellation of resonance, the peripheral volume would have needed effectively the same length as the feedpipe (approximately 39 inches), however a silencer of this length is not practical for installation on a motorcycle. For practical installation purposes a length of 26 inches was chosen, with an outside diameter of 3 ¼ inches. The feedpipe and peripheral volume were connected by a short expansion chamber of low expansion ratio and adjustable length. With this initial design a noise reduction of 10 dB(A) was observed compared to the standard Norton silencer.


 
Initial experimental 'annular discharge' silencer produced by Roe


The next development by Roe was to divide the “dead space” behind the expansion chamber into 3 cavities, each tuned by drilling holes into the peripheral gap and packed with fibreglass. Prototype silencers were built to this pattern, although it was found that a gap of ¼ inch in the peripheral volume gave insufficient damping, so was reduced to 3/16 inch. With this design a further reduction in nose was observed; now a 20 dB(A) reduction compared to the Norton part.


 
Prototype Commando silencer produced by Roe



Back-to-back power tests were completed with Roe’s new silencer. It was found that the ‘annular discharge’ silencer gave a large boost in low speed torque. The net effect was an increase in power from 2000-4000rpm, a slight loss from 4000-5000rpm, and a gain above 5000rpm. The increase in low speed torque allowed the Norton to be fitted with a higher gear ratio, which also meant a reduction in noise due to the lower engine speed utilised during noise certification testing.




In October 1973 Norton Villiers filed a US Patent on Exhaust Silencers. Geoffrey is named as joint inventor along with John Favill of Norton Villiers in US Patent Number - US3888332A, but the designs and technical argument submitted are all based on his work at Manchester.


US Patent Number - US3888332A


This patent was subsequently granted and published in June 1975. Initially these annular discharge silencers were fitted to the Norton Commando 850 Mk3 and were very effective in use. Following Norton Villiers takeover of BSA/Triumph in 1973, a similar design of annular discharge silencer was fitted to the Triumph Trident T160.


Norton Commando Mk3

Geoffrey was also interested in all aspects of motorcycle sport and in the mid-seventies his silencer design was approved by the British Speedway Promoters Association (BSPA). Speedway as a sport was coming under severe environmental pressure to reduce noise levels and the Auto Cycle Union (ACU) had finally now imposed an upper limit. As a consequence scrutineers were issued with a sound meter and testing took place on a regular basis, riders had to comply or they were ejected from the meeting.

Quote: ‘SILENCERS are going to be the biggest thing this season and you're going to have to get used to riding with one.’
 
It is no easy task to silence a high performance 500cc single cylinder air-cooled engine and it was though this interest that our paths crossed.
 As had happened in Speedway, the ACU were about to impose noise limits in all branches of the sport and being involved in Road Racing, did not wish to be caught unawares. At International level the FIM had already introduced upper noise limits in an attempt to quell the onslaught of the 2 stroke. In practice it had completely the opposite effect to that intended, the expansion boxes of the 2 strokes were easy to muffle with an absorption tail-pipe. However the 4 strokes struggled to meet the new limits without a serious loss of power and the dominant MV’s were rendered non-competitive and retired overnight.

We were currently involved in a Final Year Project on exhaust silencing with a student from Stannington Technical College in Sheffield. Through this a visit to meet Geoffrey in the Simon Engineering Laboratory was hastily arranged. We spent a very interesting afternoon in Manchester and it was instantly apparent that Geoffrey was both a very able Academic and an enthusiast. He was extremely helpful, offering advice and the benefit of his experience to help us on our way. In return we agreed to make available our findings to him, not to produce the silencers for commercial gain and before we departed a letter was signed to this effect.
We hurried home to start work on our first designs. Clearly we were influenced by what we had seen and heard that day and it was not long before the first silencer was fabricated.


2valve annular discharge silencer


As always time was of the essence and in an equally short period we were ready to test on a cold and bright day at Cadwell Park early in the season. A series of designs were produced, varying both the annulus width and the diffuser angle. The primary pipe length could also be altered easily on the day, so that the effects of any back-pressure change could hopefully be mitigated. All the initial testing was done on a 2valve engine, as the characteristics and setup were well known.


2valve testing at Cadwell Park


Initial results were good, too good if I am honest and as a rider you could hardly hear the exhaust note through your helmet. This made starting and gear changes more difficult but there was a useful increase in bottom end power, which was helpful in pulling out of Cadwell’s hairpin bend. This was no doubt caused by an increase in back-pressure in the exhaust due to the fixed plate at the rear of the diffuser chamber. This acts as an end-stop to the incoming flow, before it passes out through the primary holes and into the annulus. However there was also a slight loss of power at the top of the range which was considered unacceptable. We played with different primary pipe lengths and the alternative silencers we had available throughout the afternoon and wrote up our results accordingly. It was interesting to note that the silencers when fitted made very little difference to the carburation of the engine. This is always a good sign in that the engine is continuing to breathe well and shows that things were not a million miles away. As far as we were concerned it had been a most satisfying afternoon and even after this brief session, from the changes that were made a pattern was forming. Further silencers were fabricated in the light of our findings, and this whole development sequence was repeated many times. Eventually we began to understand what made things tick and it was time to report our findings to Geoffrey.

Another visit to Manchester was arranged and once again Geoffrey made us most welcome. We had gone for the full day this time and for him to make the time available to see us was quite telling. He was exceedingly interested in our results and in what we had found out during testing. As agreed, we passed on this information freely and showed him all the modifications that had been made as the silencer was developed. When three enthusiasts are in conversation about something that interests them, food and drink become irrelevant.

Eventually a break was taken and when we returned to the laboratory it was via his office. This time we were sworn to secrecy before he took us over to his covered drawing boards. It was here we saw the initial designs of a front suspension system for the BMW motorcycle. As you might expect the associated ‘brief’ discussion lasted well over an hour, after which he took us to a second drawing board, which if I remember correctly, was secured by a lock. Geoffrey raised the cover and asked what we saw. It was clearly an air-cooled, OHC 4valve single cylinder head, other than that not much was obvious. He informed us that it was a new design for Yamaha and a development of the single cylinder, 2valve SR500 engine currently in production. Again he asked us what we thought.
Before anyone could speak, Eric my companion informed him that it would not work.
Now Eric was always forthright and often out-spoken, but even I dreaded this last comment, straight away imagining this is the end of a flourishing relationship. Geoffrey asked why?
Eric asked what size the spark plug would be used and where it would be situated relative to the single OHC. Geoffrey produced another drawing from the bundle, at which point Eric shakes his head and says “the plug breaks through into the inlet port”. A close investigation takes place, before Geoffrey backs away wiping his eyes and says “you are absolutely right….how did I miss that?”
So there you go, even the mighty can have a day off.

Before we left, Geoffrey was asked what further developments he would make to his discharge silencer given the chance. Armed with his reply we bid farewell and headed home. Using the information gleaned from our testing and advice from Geoffrey, a second series of silencers were designed and fabricated. 


4valve annular discharge silencer


Thought had been given as to how the incoming charge is diffused into the annulus and the surface effect on flow along the silencer length. The second was a subtle change, the first more dramatic in concept, but the pair combined to greatly reduce the previous loss of performance at the top of the range. Once again significant testing was carried out, but this time using a 4valve engine.

4valve silencer (LH)




As might be expected, the exhaust noise reduction was nothing like as great as when using the original designs inspired by Geoffrey. However this did not concern us greatly as the ACU had set an upper limit of 105 decibels at a mean piston speed of 11m/s. On an 86mm stroke engine this equates to an engine speed of 3837rpm and at this figure we were well within the imposed limit. In practice it is difficult to accurately measure the noise emitted from an exhaust other than in a laboratory. Many factors like air density, moisture content and ambient sound levels come in to play, meaning that any values obtained are highly compromised. However it is possible to make a comparison under similar conditions, as long as one accepts the readings as not being absolute. 


It was an interesting project from start to finish and a wonderful opportunity to work with such an able man as Geoffrey Roe.

11 comments:

  1. Hi,
    I was inspired by a lecture in '73 or '74 by Dr. Roe., attended by the Liverpool University Motorcycle Club. The design was later shown in a bike magazine - and I kept a copy of the article. Understanding the principle of gas Velocity and Pressure relationship, down a pipe with a series of holes... and being shown by Dr. Roe that the CSA of holes per length of travel down the tube needs to follow an hyperbolic function... I followed a sort-of copy of the silencer when I worked on Circuit Breakers at Reyrolle Power Switchgear. The air exhaust (A few litres of air at 28bar) down a pair of 3in diameter pipes was a bit like the single opening of an exhaust valve from a "big single" - at least in my imagination. I managed to get a reduction on 136dBA to 124dBA at 1m from the outlet with the silencers I copied from Dr. Roe's work. This meant we achieved the certification for noise at 30m and 100m from the circuit breaker when operated. I now use this knowledge to design diffuser tubes for gas burners as a hobby.. THANKS Dr. Roe! A Great Tutor and Engineer.

    ReplyDelete
  2. Essentially, the exhaust gases do something very simple: (1) expand into the first chamber, then 2 Pass into the annulus: After that they pass 3 more chambers that contain "absorption" material. Each chamber is tuned for a frequency: based on the hole size and volume of the chamber. If you look inside original part car manufacturers' silencers, some of them do the same science. just the box is a different shape. If you look at the intake of most modern cars, there are "plastic boxes" stuck on the side of the air intakes to do the same "resonant damping" to reduce noise. For the air intake, dampers are fitted at the engine revs and load (i) for the drive-by noise test, (ii) for cruising at 70mph or 120kph. in top gear. These 2 phases are particularly quiet when driving and make the car more saleable. Which is what Dr. Roe was doing. The torque change was a bonus from "resonance" I guess? Where the "noise" dissipated and pressure either sucked-in more mixture, or allowed more exhaust out, or pushed some exhaust back into the cylinder to alter combustion and energy exchange for the gases.... just as happens today in your average car or motorcycle. In fact, Velocette, Villiers et al were all doing this from 1930s onwards, especially with 2-strokes. But - not to ignore 4-strokes - Brough and others used fish-tails the same way. Please correct me if I have got this wrong - I love to learn!

    ReplyDelete
  3. My curiousity was originally aroused in 1972 when I owned an Honda CB77 motorcycle - bought second-hand without baffles. after a few months of "loud" exhausts, I was advised by another motorcyclist (doing a Mech. Eng. degree) that the back pressure was wrong for initial carburation and I should buy and fit replacement baffles. He was right, not only was the bike as quiet as new, but low speed torque certainly felt better and economy improved. I can't say the top speed was changed. I understand the Honda design uses 2 or 3 external annular chambers in the silencers, with the central exhaust pipe passing these. The inner wall of these chambers is made by the baffle pipe, having strategically placed holes that align with each outer resonant chamber. Each resonant chamber - both on the Honda design and the annular disharge silencer - acts as an Helmholtz resonator: I think?
    This is not new technology in principle: (see GM paper https://www.jstor.org/stable/44437706?read-now=1&seq=3#page_scan_tab_contents).
    However, I understand that the Annular discharge silencer has some other advantages. If the main exhaust stream is passing down the centre of the silencer, then there is less opportunity for the gas to lose energy (heat) because it is surrounded by chambers of hot exhaust - acting as insulation. Although some heat is lost by the pressure wave entering then later leaving the chamber, this is small in comparison to the conducted heat from the hot gas, through the metal to atmosphere. But this conducted heat is also small in comparison to the heat dumped out of the tail-pipe. However, with the annualr discharge silencer, the hottest gas of the exhaust stream is changing prerssure to velocity in the thin annulus, of which the outer shell is exposed to atmospheric temperature. Thus heat from the exhaust gas main stream, as well as frictional heating from the fast gas stream adjacent to the silencer outer wall, heats the silencer wall and thus loses this heat to the outside atmosphere. This in turn has an extra sound deadening effect as it is extracting energy from the exhaust stream. While I am not clever enough to crunch the numbers, it would be interesting to know if my postulations are correct? Any thoughts?
    Ken

    ReplyDelete
  4. Further memories of racing bike silencing: Between 1973 and 1979 I was a volunteer marshall at Aintree race course with the dubious pleasure of managing the the exit from pits and start zone (grid positioning of riders). (Waterloo and District MCC meetings). This meant I spent all day alongside bikes that were being revved to warm them up.... while the ACU Noise-meter-man checked random bikes with his meter and notched stick. I think the lads had to rev their engines to 3000rpm or 50% of the red-line on their tacho, whichever was higher. - Or something like that. I was curious to note that from the earliest meetings to a couplke of years later, how silencers appeared first on the tail-pipe of expansion boxes of the 2-strokes, but also the short large cans (annular disharge types?) on 4-stroke exhaust pipes. I didn't have a comparison between unsilenced and silenced bikes, and my hearing today is well knackered... perhaps in part due to the 105 dBA I experienced for most of a 10 hour shift at Aintree? - Those were the days before we heard of Elf and Safety....

    ReplyDelete
  5. This comment has been removed by the author.

    ReplyDelete
  6. This comment has been removed by the author.

    ReplyDelete
  7. Further to my earlier comments (as Ken) - I have just re-read the article: Drs. Roe and Thorpe presented their "motorcycle improvements from research" to the Liverpool University Motorcycle club in 1974 - or 75. This stimulated a design I produced for a 400kV circuit breaker silencer: To silence an air motor that was discharging 28bar of air to atmosphere from a 10in diameter x 5 in stroke cylinder. I managed to drop 9dBA from the "open pipe" to "Silenced" condition - as a comparison at 1m from the exit orifice. (similar to the 1m stick the Noise adjudicator used at Aintree!). But I did not use Helmholtz resonators: The air passed into the central larger tube, thence via 4 sets of holes into the outer annulus, and by "pressure wave chopping" and the different speeds of sound inside the large central tube (from exhaust pipe) and the outer annulus (to atmosphere) caused a higher frequency and lower peak magnitude noise: A rifle shot instead of a cannon "boom". So the circuit breaker could not be heard at the nearest village to the sub-station. (unlike previous designs!). I am not an acoustics engineer, so actually cannot calculate what was happening. But there was some adiabatic expansion cooling of the air blasting down the silencer tubes, as they always has a good layer of condensation develop on the outside of the cold steel tubes after an operation. I suspect the Bike silencers also had this effect, though not so cold as the start temperature was much higher.
    Ken

    ReplyDelete
  8. This comment has been removed by the author.

    ReplyDelete
  9. This comment has been removed by the author.

    ReplyDelete