Centaur Valve Timing

I'm in the process of setting up the valve gear of my Centaur gas engine. Having made the cams to drawing, I can't achieve the specified valve timings. Both valves have opening periods that are too short.The drawings are shown below.

The inlet and exhaust cams lift over 220deg and 240deg. resp. and slightly less if tappet clearance is present The timing diagram calls for 240deg. and 250deg.Which is correct, does it matter much or have I got this wrong? Since ETW was a very respected designer and the Centaur is a long established design I'm surprised to find this apparent discrepancy. Comments would be appreciated.

He does mention in the notes that if you set the opening points exactly to the drawing you will find the valve close a little early.

**LINK**

Page 504.

I think you may find the discrepancy is down to the profile of the cam follower. If it were a small diameter roller, it would follow the cam profile fairly closely. But with the Centaur's flat profile cam followers, the edge of the flat camfollower contacts the ramp of the cam before said ramp hits the centreline between cam spindle and cam follower. Likewise, the cam follower does not fully close until the trailing ramp has moved past the other edge of the cam follower. Hence the actual valve opening duration is slightly longer than the cam iteself would suggest.

If the cam follower is halfway in between and a flat follower but with a bit of a radius on it, the result will be halfway in between. On some old motorbikes with radiused "flat" followers, a performance upgrade was to fit followers with a larger radius or even flat, in order to both advance the valve timing and increase the duration of opening time. Even a full sized roller follower will extend timing duration beyond the simple profile of the cam to some degree.

Best way to check it is put a large degree wheel on the finished engine and a dial indicator on the pushrod and dial it in exactly.

PS Looking at pics of the Centaur, it looks like it has in fact large diameter roller followers, which would add to the timing duration to some extent.

Phil Irving (designer of the Vincent motorbike) goes into this stuff in some depth in his book "Tuning for Speed" and how to map out valve lift with a roller tappet. It is not a plain matter of cam lift equals valve lift. Quite complex to map it all out, by drawing a roller moving around the profile of the cam in one degree intervals, then plotting the path of the centre point of all those circles. It does add degrees to the valve opening duration.

Edited By Hopper on 15/06/2022 12:29:32

You can see how different cam follower shapes affect actual valve timing, vis a vis cam profile.

in (a), a perfect point follower (never used in practice due to wear) follows the cam's timing exactly.

But in (b) with a roller follower, the follower has already started to open the valve, even though the ramp is not yet in line with the centre line. The cam is in the same position as in a, but the roller is contacting way up towards the point of the cam lobe.

In (c) with a flat tappet, the effect is more exaggerated. The cam is in the same postiion as in a and b, but the follower is riding on the very point of the cam already, well and truly open. The contact point is a way over to the right of the centreline of the camshaft, not in line with it as the pointer follower in a was.

That effect would give you your extra 20 to 40 degrees of valve opening duration between the gain on opening and the mirror effect on closing.

So if you took out your roller tappets and put in flat tappets, you would get earlier opening, later closing, plus faster opening and more time with the valve close to fully open, but the same lift. Cheap bolt in horsepower. Probably not somthing you want to do on a model stationary engine! But vintage motorbike guys were known to do it back in the day.

Edited By Hopper on 15/06/2022 12:46:37

This is a welcome post for me as I have the same problem and have built two working i/c engines while trying to work out what is wrong.I had come to the conclusion that the cam radious is to short and after the next project finishes I am going to make new cams.The explanations given in the posts gives me hope to get it going.

Frank

Thanks gents for quick replies;

Jon, I had read ETW's build notes re lost motion in the valve train, i did mention tappet clearance in my post but the timing differences I'm seeing are too large for this I think.

Hopper, My mearurements are with a scale on the flywheel with a dti on the rockers as you suggest. I agree that the follower geometry will affect the measured valve lift for any given intermediate position of the cam lobe, but, for a roller follower at least, I can't see it altering the angular positions of the start and finish of the valve movement. Those positions coincide with the roller crossing between the base circle and the two cam flanks, which are tangents to that circle. This angular distance is fixed by the cam shape itself. In your post you seem to suggest that extra angular distance might result, whereas I'm seeing a shorter opening period, this with virtually no clearance in the train.

Bricky, you say your engines work, why are you considering new cams? I'm thinking along that line with the Centaur.

Clive

Deleted as not relevant

Edited By David Jupp on 15/06/2022 14:46:11

Edited By David Jupp on 15/06/2022 14:46:32

Clive I have made two engines designed by Brian Rupnow,they work but the first engine I built is the centaur which I can't get to work.I have good compresion and useing a large perspex timing disc I finally noticed the exhaust valve was not staying open long enough and I am going to try a cam with a longer opening time.

Frank

Frank, thanks for follow up, seems we're both in the same boat. With me it's both valves, and the inlet shows greater difference from spec. than the exhaust.

So many tolerances will affect the valve timing. For a slow speed engine the timing is not very critical, with good compression and good ignition it will run.

Indeed. I had to check Phil Irving's book and you are correct. He in fact says follower radius influences rate of valve opening and closing but not the exact point of initial opening or closing. As you say, the point of contact at initial opening is still in a straight line between cam and follower centrelines. My bad. Sorry, I hadn't realised you had actually measured real-world valve timing with a disc.

But you would think it should still run at least to start up with the cam timing you have. On those old low revving stationary engines, it would seem cam timing was hardly as critical as on a Manx Norton or Vincent etc. The early closing of the exhaust valve would reduce valve "overlap" when they are both open around TDC, but that is usually not a critical factor until the engine is at speed and the action of the exhaust pulses down the exhaust pipe provide a "suction" effect to pull more charge in through the open intake valve. Plenty of old engines run quite well with zero degrees of overlap there.

Maybe play around with slightly varying the overall cam timing to find a compromise that works? Or varying the ignition timing also? Spark right on TDC should be enough to get it started. I would be real surprised if the smallish variation you have in valve timing were enough to stop it from running at all. (But of course I have no experience with the Centaur so am only guessing from general principles from years of mucking around with much larger engines!) Or maybe try a commercial carburettor off something like a lawn string trimmer and see if that helps?

Looking further at your timing diagram in the OP, is is very surprising to see that kind of valve overlap on a low-revving, low-performance stationary engine. According to this chart, your Centaur has more valve overlap than ETW's Kiwi engine, which has 10 degrees inlet lead and 20 degrees exhaust lag vs your Centaur chart showing 15 and 20 respectively. That's according to the Kiwi drawings I have sitting under the bench awaiting the day castings become affordable.

This overlap is surprising on the slow revving Centaur because the Kiwi was a high-performance, high revving engine developed on a dynamometer with Prof Chaddock to power racing model boats at full tilt. So it seems possible that ETW's drawing of the Centaur cams is correct but his timing chart is out, perhaps confused with one from a higher performance engine?

So you might try sticking with your existing cams but repositioning them so the intake opens later and the exhaust closes earlier, pretty much eliminating the overlap when they are both open, and both valves sticking to the timing diagram at the other end of their duration, ie intake closes 45 deg after BDC and exhaust opens 50 deg before BDC.

Plenty of low-revving low-performance engines (including some Harley Davidsons of the 1990s even!) run with little or no overlap, or even negative overlap where the exhaust valve closes a few degrees before the intake opens (1990s Harleys to meet emission standards) so your Centaur should run with that sort of valve timing.

Lack of valve timing overlap should also give you increased effective compression, which can only help too. Valve overlap really depends on a long exhaust pipe to provide the extraction effect for exhaust gasses, then a tuned length pipe so a reversion pulse wave pushes any excess intake charge back into the cylinder from the exhaust port just before the valve closes. You are not going to get any of this stuff going on at low revs and with the Centaur's stubby exhaust pipe. So I am really suspicious of that timing diagram with all that overlap in it.

ON the other hand, many dozens or hundreds of these engines have been built in the 65 years since its inception and there seems to be no mention of them not running with the standard cam set up. So maybe the problem is elsewhere?

Edited By Hopper on 16/06/2022 06:12:23

Edited By Hopper on 16/06/2022 06:18:21

PS Scratch the comment about increasing effective compression. Wrong end of the intake valve opening. Doh!

When I built my Centaur I had the same problem with the valve timing, after many hours trying to get the timing as the drawing showed and using a degree disc on the flywheel I gave up, I then set the timing to what I thought would work, I set the inlet to start opening just after TDC, the exhaust to open at BDC, I don't know at what degrees the valve timings are all I did was position the crank and watch the valves and adjust the cams to get what I wanted.

It's a slogger of a engine and as Hopper says it doesn't want valve overlap.

The result here LINK

MichaelR

Edited By MichaelR on 16/06/2022 09:07:16

Beautiful engine MichealR. I must stop watching these vidoes or it will cost me a lot of money to ship one of those kits to the colonies.

I think you are on to the right solution with the cam timing though. Get back to basics, intake opens on TDC and Exhaust on BDC. All the rest is a bit moot for a model engine running at no load for demonstration runs. Not like it needs to develop power to saw logs or cut chaff or whatever like the original full sized engines did.

Edited By Hopper on 16/06/2022 09:55:28

For engines of this type if the inlet opens at TDC and the exhaust at BDC then it should run UNLESS the ignition timing is out - and sometimes even a VERY small error will be enough to stop it from running. I well remember a briggs and stratton engined rotary grass cutter that was stopped by a rock and proved immpossible to restart, the fly wheel is retained by an ali key. The sudden stop had put a tiny step in the key of a few thou that put the timing out, Filed a new key and off it went again.

There was a very interesting series of articles in ME a few years ago on IC engines, I've worked on IC engines all my life but I learnt quite a lot. Noel.

Just had a look in my 1923 Dykes Automobile Encyclopedia, the universal sort of Haynes manual of its day, and every car and stationary engine timing chart in there shows negative valve overlap. The exhaust valves are shown typically closing about 5 deg after TDC and the intake does not open until about 10 deg after TDC. So again I would supspect that ETW's chart showing 35 degrees of overlap is most likely wrong for this honourable old banger of an engine. Again, it would seem to point to stick with the cams themselves as drawn but position them to open and close respectively right at top dead centre as a starting point.

Or alternatively, position the cam lobes to open and close at the 45 and 50 degree points respectively on the bottom of the ETW chart and let the overlap fall where it will at the top, by my rough mental reckoning exhaust closing about 10 deg after TDC and inlet opening about 5 degrees after TDC, giving a much milder 5 degrees of overlap. Seems more appropriate than the charted 35 degrees.

Edited By Hopper on 16/06/2022 13:13:37

Thank you all for these further thoughts and comments which reassure me considerably. This is the first IC engine that I've built so I have no real experience to draw on. Whilst I did think that the valve overlap etc seemed strange, ETW's reputation made it difficult to ignore "the words and music" despite the inconsistecies. I will now set the valves to what seems sensible positions.

A good tip about ignition timing.

MichaelR; a very nice engine, hope mine turns out somewhere near that result.

Once again, many thanks,

Clive

Excellent, pity the castings are so expensive

You're welcome. I hope some of it helps you get it going. Please do let us know how you go.