This page describes:
We purchased a 1959 TR3A vintage race car (now for sale) and started racing in 2002. We made many newbie mistakes during the first year, some are too embarrassing to relate. We had developed a little knowledge about cams during the Model T Cam Project in 2001, so it wasn't long before we started thinking about the cam in the TR3. When we got the car, the engine was fairly mild, with about 10:1 compression ratio and with an old performance cam (Crane IMP-304). The cam had a lift of only 0.270, but based on degree measurements, the seat-to-seat duration was more than 310 degrees. To make matters worse, it was not installed correctly. It was retarded 20 degrees.
It was obvious that the original cam in the car had to go, so we began to look at available cams. The table below was created mostly from manufacturer information on the internet. The limited data given by many manufacturers makes shopping for a camshaft difficult. For example, the duration at 0.050 cam lift is the best indicator of the power band location (see Cam Performance). Unfortunately, none of the British manufacturers give this spec. Instead, they use catchy names like "Fast Road" or "Sprint", which are not helpful. To make matters worse, we have found that the actual seat-to-seat duration is usually greater than the specification (see below). As a consumer you should refuse to buy a cam without an 0.050 spec, accurate seat-to-seat numbers and cam lift. If you are given valve lift, ask how it was calculated. Often times it is calculated using gross cam lift and an inflated rocker ratio. Ask for the cam lift at which the seat duration is measured. It should be close to or a few thousandths larger than the valve lash divided by the rocker ratio. Even when the 0.050 spec is given, it is difficult to compare cams with significantly different valve lash (see What's Wrong with 0.050 Duration?)..
In the table below, we have taken the manufacturer data and attempted to put all the cams on the same basis. To compensate for differences in valve lash, we have estimated the duration at 0.040 net lift as outlined in the footnotes below the table. Even though it is approximate, the duration at 0.040 net is the best indicator of the power band location for a given cam. For many of these cams the manufacturer did not specify either a lobe separation angle or cam advance. Many manufacturers will grind a cam at any lobe separation. If the lobe separation was not specified, a reasonable value is listed. The cam sprockets for these engines allow for any cam advance within 4 degrees. Vernier sprockets are available for even finer adjustment. When the manufacturer did not specify the cam advance a reasonable value was listed in the table. A vernier sprocket and a few hours on a dyno is the best way to determine the optimum cam advance. Given the basic data in the table, you can easily look at other values for advance and lobe separation using our Cam Calcuator.
|Stock TR2,TR3 (spec)||250||110||0||15||55||55||15||0.010||Triumph|
|Stock TR4 (spec)||254||110||0||17||57||57||17||0.010||Triumph|
|Stock TR4||264||217||217||110||0||22||62||62||22||0.263||0.253||0.010||Integral, Elgin|
|Elgin 64 5-12 2||258||204||203||110||0||19||59||59||19||0.251||0.240||0.012|
|Revington Fast Road||270||107.5||4.5||32||58||67||23||0.287||0.273||0.016|
|Integral Stage 1 2||276||221||220||110||0||28||68||68||28||0.285||0.273||0.013|
|Revington TR Sprint||280||108||5||37||63||73||27||0.264||0.246||0.022|
|Elgin 70-9 2||280||228||229||110||0||30||70||70||30||0.262||0.253||0.009|
|Integral Stage 2 2||284||230||229||110||0||32||72||72||32||0.295||0.283||0.013|
|Erson 149 2||282||238||235||110||2||33||69||73||29||0.287||0.272||0.018|
|Kastner D||284||109||1||34||70||72||32||0.269||0.257||0.014||TS Imported|
|Revington/Kent TR4-6 7||290||108||5||42||68||78||32||0.309||0.291||0.022|
|Erson 5 2||290||245||237||108||2||39||71||75||35||0.332||0.311||0.026|
|Integral Stage 3 2||292||238||237||109||2||39||73||77||35||0.305||0.293||0.013|
|Elgin 715-18 2||286||245||240||110||2||35||71||75||31||0.305||0.288||0.020|
|Integral Stage 4 2||300||245||244||108||4||46||74||82||38||0.325||0.313||0.013|
|Erson 4 2||302||254||249||110||4||45||77||85||37||0.353||0.336||0.020|
|Revington/Kent TR4-7 7||300||108||5||47||73||83||37||0.324||0.306||0.022|
|Elgin 75-12 2||300||250||249||108||4||46||74||82||38||0.323||0.312||0.012|
|Kastner F||300||111||2||41||79||83||37||0.299||0.287||0.014||TS Imported|
|Integral Stage 5 2||308||253||252||108||5||51||77||87||41||0.345||0.333||0.013|
|Erson 16 2||308||261||256||108||4||50||78||86||42||0.343||0.326||0.020|
|Elgin 77-9 2||308||256||256||108||5||51||77||87||41||0.312||0.302||0.011|
|Kastner G-3||309||104||1||51.5||77.5||79.5||49.5||0.337||0.323||0.016||TS Imported|
|Revington TR Race||310||108||5||52||78||88||42||0.339||0.321||0.022|
|Erson 23 2||312||266||260||106||4||54||78||86||46||0.376||0.358||0.022|
|Integral Stage 6 2||314||269||267||106||4||55||79||87||47||0.387||0.375||0.013|
- Reasonable advance assumed when none listed.
- Lobe separation not listed, assumed to be variable, reasonable value listed
- All cams have identical intake & exhaust lobes (single pattern), except, Cambridge Rally 290 exhaust
- Intake lash listed, exhaust is the same or 0.002 greater
- The difference between net lift and gross lift = Lash/(1.48) + 0.003
- 0.040 net calculated using 0.050 duration, valve lash and estimated velocity of 0.0055 in/deg
- Revington TR and Kent Cams have identical specs
Before selecting a cam from the table, you should first read and understand the Cam Basics and Cam Performance pages. Cams with duration (at 0.040 net) less than 230 should be streetable. A duration of 230 to 240 would be marginal for street use, but might be good for autocross. Vintage race engines with ported heads and at least 11:1 compression, should have a cam duration of about 250 if you shift at 6,000. If you want to spin your engine to 7,000 or more, then you will want a duration of 260 or more.
Since the engine in our racecar was fairly conservative, we wanted a cam with less duration than the original cam. A duration of 235 to 240 at 0.040 net lift and perhaps 290 seat-to-seat seemed like a good choice. We selected a fast road cam from a reputable supplier, but without knowing the duration at 0.050. When we degreed the cam, we found that the seat duration was 8 degrees more than stated, the duration at 0.040 net was about 251 degrees. The lift was greater than specified. This was more of a race cam than a fast road cam. I should have returned the cam because it did not meet the specs. Instead, I installed the cam using a larger valve lash to achieve the specified duration. This was a mistake for such an aggressive cam. On the first weekend, the stock valve train suffered a broken rocker shaft and 5 bent pushrods. On the second weekend (the Mid Ohio gathering in 2002), the lifters became mushroomed and one lobe went away. I admit that not all the problems were caused by the cam, some were my fault. Though short lived, the engine performance with this camshaft was absolutely wonderful.
The engine was then rebuilt. Since I wanted a conservative reliable engine, an Erson #5 was selected for the rebuild. This cam worked well for the three years it was used, but after that, we felt it was time to step things up. We decided to build a new motor with greater performance potential. uncle jack ported the head. A stock crankshaft was modified to accept Chevy V8 H-beam rods. Pistons with a very slight popup were used to achieve about 12.5:1 compression.
For the camshaft, I kept thinking about the great performance I had with the busted camshaft. I wanted to design a camshaft with similar lift characteristics, but a bit less aggressive. Before proceeding, we needed to gather some background data. With the help of uncle jack and Specialty Motor Cams, we ran a cam profiler on a stock cam and five performance cams. To fully understand the discussion of these results, you will need to read and understand the Cam Design page. Also look at What Is an Aggressive Cam? Here is a summary of the results we found:
- The most aggressive cam (max acceleration 0.00074 in/deg2, max jerk 0.00014 in/deg3 ) was the one that broke the rocker shaft and bent pushrods.
- The second most aggressive cam (max acceleration 0.00061 in/deg2, max jerk 0.00013 in/deg3) was the stock cam
- The other four cams were not at all aggressive (average max acceleration 0.00032 in/deg2, max jerk 0.00004 in/deg3)
- Three of the performance cams had valve lash specifications that opened/closed the cam on the flank (average velocity 0.0017 in/deg)
- For three of the performance cams the actual seat-to-seat duration averaged 11 degrees more than the specification. One had no seat-to-seat specification. One was 9 degrees less than the specification.
- When the duration at 0.050 was specified, its value agreed to within one or two degrees.
- Four of the performance cams had a maximum velocity that would allow them to run on a SBC lifter (0.842 in)
To put these numbers into perspective, typical OHV cams have a maximum acceleration from 0.0003 in/deg2 to 0.0005 in/deg2 (see Baykoni), with the lower value typical for your normal production car and the higher value typical of a high performance engine. If we compare these numbers, the stock Triumph cam is more aggressive than those used in most high performance engines, and four of the performance cams are about as aggressive as the cam in your typical grocery getter. The Triumph valve train works with such an aggressive cam because it is well designed and lighter than the valve train in typical American cars. The cam that broke the the rocker shaft is even more aggressive than the stock cam, and is apparently too much for a stock Triumph valve train.
The page, How to Determine Valve Lash, describes how the valve clearance should be determined. It also gives a more detailed description for one of the three cams which specified a lash that opened and closed the valve on the flank. The high velocity impact of the valve hitting the seat could cause valve bounce, high wear and require high spring pressure (see YouTube video of valve train problems)
What did all of this mean? It meant there was room for improvement over the cams available. The fact that four of the cams would run on a SBC lifter, and they were about as aggressive as a production car, made me suspect they were not designed for a Triumph. Now that we have our own cam profiler, we have profiled numerous TR4 racing cams. Every cam produced in the US was designed to allow it to run on a small block Chevy lifter. By using similar parameters as a stock cam, but with greater duration, and use of the entire 0.935 inch lifter, it was clear that a quick opening cam with high lift should be possible.
The Opticam™ cam design software was created in order to develop better camshafts (see Cam Design). For the new cam, we relied on the parameters measured for a stock cam and used valve train dynamics calculations to fine tune the design. A simple SDOF (single degree of freedom) model was used to develop the TR298Dx in 2005. A more sophisticated MDOF model was used in the more recently developed cams. Of course, valve train dynamic models are approximate, so we relied on the calculated performance of the new cam relative to some with known characteristics. Factored into our approach was the knowledge that a stock cam and valve train is good to at least 6,000 RPM, and the cam that broke the rocker shaft and bent the pushrods was too much. The graph below left shows the calculated spring forces and the forces between the cam and lifter at 6,000 RPM. The differences in the spring forces are due to the differences in lift. The spring forces oscillate due to surge effects, but are still greatest at the high lift point (0 degrees). The maximum force on the cam is about 1000 lbs, which seems like a lot for the pushrod to support. If there is no force on the cam, valve train separation or valve toss will occur. At this speed, the TR298Dx shows a small period of separation at -32 degrees. The stock cam shows a larger region of separation from -48 to -40 degrees, but the maximum separation is only 0.001 inches. The graph below right, shows for various engine speeds, the minimum and maximum calculated force on the cam and the velocity when the valve first contacts the seat. Separation first occurs at 5,500 RPM for the stock cam and 6,000 RPM for the TR298Dx. The calculated maximum forces and seating velocity are greater for the stock cam.
These calculations show conclusively that the TR298 cams are milder than a stock cam. The calculations indicate the TR298 should be good to at least 6,500. These conclusions are drawn from the calculations, in practice we have found engines with the TR298 cams pull well to 7,000 RPM. We use valve train dynamic calculations in the design of all our cams. The calculations show that the TR270 cam should be good to 6,000 RPM and the TR310 cams should go to at least 7,000 RPM without valve train problems. For comparison to the numbers listed above, the TR298 cams have maximum acceleration and jerk of 0.00055 in/deg2 and 0.00015 in/deg3. The TR270 has maximum acceleration and jerk of 0.00050 in/deg2 and 0.00040 in/deg3. The TR310 cams have maximum acceleration and jerk of 0.00048 in/deg2 and 0.00020 in/deg3 and is asymmetric with a milder closing to avoid valve seat problems. The two TR298 cams and TR310 cams differ only by the size of the lifter. The "F" cams are designed to run on a 0.875 inch Ford lifter, while the others require a full sized 0.935 inch Triumph lifter.
In 2005, our crankshaft grinder had problems with the crank for our new engine, so we had to freshen our old one . We had a TR298D made and installed it in the old engine. It was first run at the CVAR fall race at Hallett in 2005. It seemed to work quite well. That engine threw a rod at the next CVAR race at Texas World Speedway. uncle jack had been following this development and decided he wanted to try the cam. We had three more cams ground. Over the winter, we built our new motor and uncle jack built his. Due to a combination of mechanical and personal problems, we did not get that motor sorted until spring of 2008. It's performance at the CVAR Hallett spring race in 2008 was outstanding (see Testimonials). uncle jack had several problems that held him back also, but he was running well by the fall of 2006. Jack was ecstatic about the performance (see Testimonials). He was running so well that Tony Drews became our next customer. Tony blew everyone away at the MOTRAH race at Road America in 2007 (see Testimonials). This led to requests from other Triumph racers, so we had 7 more cams ground. Then we had requests for a longer duration cam, for cams that would run on a Ford lifter and also for a street performance cam. These requests lead to development of the TR298F, TR310, TR310F and TR270 cams.
We have now ground over 30 cams for Triumphs. These have all been precision ground with great care, and currently we are using the finest CNC grinding equipment. By examination of the table, it is apparent that our cams are quicker opening and have more net lift than other cams with a similar duration. These designs are better because they take advantage of the specific characteristics of the TR3/4 engine. They are not recycled designs for some other engine. The seat-to-seat duration and other specifications for our cams are honest and accurate. If you find this is not the case, you may return the cam for a full refund.
Until we get a chance to develop the rest of this page, check out the following links: