Tag Archives: Jennings

Reverse engineering

Moving directly on from the last post at the end of January, I thought I’d share a little progress on the exhaust front.
Although I’ve been completely absorbed in doing engines for other people, I’ve been wanting to find some time to go back to basics and map the Gibsons made exhaust that I’ve been using on the Scott racer. Now this isn’t an easy thing to do completely accurately as the exhaust is shaped to wind its way through the frame, but I think I made a reasonable job of measuring it out.
I drew it up with CAD, but it’s actually easier to see on sketch up.

model of current Gibsons pipe

I know you can’t see all the detail but it’s interesting to see the lengths involved and tapers used. A couple of things are interesting, the stinger ID works out to be around 54% of the ID of the start of the header pipe. Now I don’t know much about expansion pipes except what I’ve read and been told, but the stinger ID is a critical dimension and very small dimensional changes can make noticeable differences in the effect of an exhaust. If too small a stinger, excess pressure can cause overheating in the engine. Now back at Cadwell in 2015 I was told that as a rule the percentage difference between the stinger and the header should be around 60%, which would make it ~29.8mm ID. That’s 2.8mm difference, which is not even that small. Also, there’s another thing: this is a 2 into 1 pipe, although I’ve only drawn a single pipe in the example. I’ve a Scott engine behind me on the bench with a similar exhaust duration as mine; mine is around 160° duration and the one behind me is around 157°. If this bike had a pipe for each cylinder, then it would have ~203° of crank rotation from the closure of the exhaust port to the re-opening on the next cycle. Since this is a 2 into 1 pipe, the exhaust has ~23° of crank rotation from the closure of one exhaust to the opening of the other. That’s around 11% of the time at any equivalent RPM or even if there’s a mathematical reason why that’s not completely true, it’s a lot less time to do the same job.
Given that the 60% guideline for the stinger is supposed to be based on a one pipe to one cylinder design, am I to extrapolate the time difference to apply to the cross sectional area of the stinger? Again, I’m sure the maths are well above my pay grade for that, but what’s obvious is that it’s about time and area, and most bikes rev faster than a Scott.
A quick googling leads to me think that an early 70’s TZ 250 would rev to 10,500 rpm, for example. Given that the tuning data from Bell would have been used on actual racing machines in that era, then it’s fair to assume that this was representative of the kind of engine speed that the 60% Stinger ID to Header ID rule of thumb, was worked out for. For sure the exhaust durations are different too, but I need to look at this to make a guess at what the stinger ID should be.
Also, having been able to establish the design and measurements of the Gibson pipe, I found that it’s actually made for an engine that revs quite a bit higher than mine.. like my dad’s! (although it’s even on the high side for that. Working backwards, the dimension between the piston face and the mid point in the deflector cone is 1244.5mm, which according to Bell makes it tuned for a whisker under 5500 rpm. Now Roger’s engine, with his substantially balanced four bearing crank will spin up past 5500rpm, but that’s not a place he likes to ride it anyway. Dobsy used to rev it hard, but even my dads engine is getting a little out of breath at that point.. just a limit to what the ports can flow, even with the differences on his engine. My calculations would suggest that I need a tuned length of around 1530mm for 4450rpm. 5000rpm is about right for max revs, but as I’ve said before, I just don’t think it can breath properly up there. Thats 285mm longer, not insubstantial in terms of getting it to fit on the bike. We might be looking at the muffler the other side… I know that’s been done.

So I bought Bell’s two stroke tuning book as well as John Robinson’s to go with my copy of Jennings and have been working through some possible re-designs. I purchased a TIG welder a few weeks ago and so that’s another step in the right direction. I think hydroforming would be the best way of testing some designs.. the most fun you can have with a pressure washer as far as I can see!

Scott Super Squirrel tuning continued- port calculations

It’s been a dispersed couple of weeks on the Scott front but I have been making progress. I’ve been reading the Jennings book on two stroke tuning from the early seventies. It’s quite well respected and has some fairly straightforward explanations of what modifications are likely to do what. The Scott is a bit unusual though and its extraordinary torque at low revs is something I would like to keep.
Let me say that everything from this point on is me trying to work something out, rather than prescribing a best practice!

It’s difficult to know who to listen to but I’ve used Jennings’ recommendations to establish the important points about the inlet, exhaust and transfer ports, those being the time that they are open and the area that is available for gas flow.
He doesn’t use the entire area, but actually has a method for calculating a mean area which is significantly smaller, but apparently more representative.
His assertion is that regardless of the size of a cylinder it will need a certain time/area to achieve optimum power at any given speed of crank rotation and I’ve created a spreadsheet to assist in calculating what’s happening. Doing the maths, combined with looking at the port timings of other engines, have given me some indications of where to go.

This is a useful chart that was once published in Yowl:

Timing info

From having seen many Scott iron block castings, I can say that the port cores were seldom perfectly aligned. Certainly on my detachable head block the port apertures are at angles to each other and vertically misaligned slightly also between the two cylinders. The difference between cylinder timing durations is not insignificant: 4° on the transfer, 3.5° on the exhaust but only 1° on the inlet.

My own as yet standard timings (from one cylinder) started off as follows:

Exhaust: 159.5°
Inlet: 129°
Transfer: 134°

With those timings, I calculated using Jennings method that the combination of available exhaust port area and it’s time open would provide optimum power at 3500rpm, the transfer at 3250rpm and the inlet at about 3100rpm. Of these, the transfer and the exhaust figures are exactly standard, where-as the inlet differs for a couple of reasons.

1: As we have ported pistons to aid the gas flow from the crank chamber on the transfer phase, the rear part of the inlet gallery is blocked up on my engine (otherwise the port in the piston would communicate with the inlet gallery). This means that I have less actual inlet area available than standard. However, since the standard inlet gallery ports are quite small and each bridged, mine has the bridge removed between two and the port raised (as far as possible) to enable more area. My guess would be, that simply adding up the area without taking into account the effects of turbulence around the bridges at higher gas speeds is ignoring an important factor. Whether the gas speed ever gets high enough (given the amount of ports in the gallery) in a standard engine for this to be a truly limiting factor, I don’t know.
2: Our inlet timing is already extended from standard by the removal of about an 1/8″ skirt at the bottom of the piston. According to the tables available, this amounts to a difference of about 17° in total duration over the standard figure.

The standard port timings are really set up for a low engine speed, and the power and torque curves shown on dyno charts bear this out. Here are some charts to see this:

original Scott dyno test sheet
original Scott dyno test sheet

My Super squirrel racer, running methanol, tested on a dynojet rolling road dyno (albeit with RH blown head gasket) in October 2013.

Super Squirrel racer HP Dyno chart (Oct 13)
Super Squirrel racer HP Dyno chart (Oct 13)
Super Squirrel racer torque curve (Oct 13)
Super Squirrel racer torque curve (Oct 13)

Although the fact that the head gasket was blown means that the results themselves are not reliable, the curves are likely to be, and so I can at least see where (with the expansion chamber fitted) the torque and power is being made.

So you can see that the torque curve gives a good spread of consistent peak torque between 3200rpm and 4200rpm. You can also see that everything stops at 5000rpm, although I’m not sure I understand exactly what that shows. I’m assuming it is showing that the output drops off sharply and not just that the bike was only revved to 5000rpm during the test as you expect the curve to just terminate high if that were the case. This tailing off is definitely something to look at.

This torque curve does show that the output roughly corresponds with the Jennings based predictions. The expansion chamber will alter the results according to it’s own harmonics too, and in a perfect world I’d have run a straight pipe to try to remove that effect… but I only had an hour that time. I may do this in the future.

So, I should have a look at how this curve works within my gear ratios. It’s all a bit more important when you’ve only got three. I know it works really well at the moment, so I’m simply hoping that by not trying to do anything too extreme I shouldn’t have any problems.

I have set up the spreadsheet to show the % percentage of the optimum time area figure I am achieving for any given revs. This enables me to change a port duration figure, or a port width and instantly see its effect as a % of the optimum. Theoretically.
Alongside this, I also have some figures from A.Graham Bells book on two stroke tuning from the eighties which gave details of conclusions drawn from the known port timings for different engine configurations. By extrapolating the results to 5000 rpm, it seems 140° would be in sequence for the inlet, the exhaust would be around 165° and the transfer would seem to already be long at 134°.

So, how so these look on the Jennings spreadsheet?

As things are at the moment, the optimum theoretical rev/minute figure corresponding to my port timings (regardless of exhaust influence) is between 3100rpm and 3500rpm. With the above inlet port duration modification plus a little widening of ports, the figures say that I’ll be able to create the optimum time/area conditions for power at 4000rpm. That would be, around 135° for the inlet, 160°° for the exhaust and to leave the transfer alone.

I’ve got this week to finish the ports, I’ve ordered some new main bearings and next weekend I’ll be rebuilding the engine.