Modifying the stock KTM Duke2 exhaust cans
If you own one of these wonderful, hysterical bikes, you will have noticed two things. First, that the stock exhaust is far too civilised and completely out of character. It is actually so quiet that it is a liability in traffic. Moreover, it chokes that fabulous engine. Second, that a pair of Akrapovics cost an absolute fortune - £600 in the UK.
Whilst there are cheaper aftermarket cans, none are cheap, some are not very well made, and the stock cans are well made and look rather good anyway. This ongoing project started out intending to make up a whole new straight-through core along the lines described at Team Incomplete's page.
However, I didn't like the way they got rid of the restrictor tube by grinding down the end cap to leave a huge hole. I wanted the appearance to stay pretty much stock, and without obnoxious noise levels. As a socially responsible hooligan, I wanted it to sail through MOT inspections and not result in unwanted police attention.
First thing I did was get a second pair of cans to experiment on, via eBay. Naturally I paid more than I intended, paid lots for carriage, and then paid a fortune to Customs, as I bought them from the USA. They happened to be all that I could find at the time. Since then of course, several sets have sold in GB and Europe at lower prices. Going rate is around 100GBP-140GBP for a pair. Considering the amazing quality of manufacture - all stainless and alloy - this isn't bad. Incidentally, the repulsive white-paint Federal 'mess-with-this-silencer-and-you'll-go-to-gaol' warning notice can be removed from US cans with a bit of Nitromors paint stripper and Scotchbrite plastic abrasive pan scourer, without any damage to the anodised surface. Suggesting removal of the notice is probably also an offence worth a few years in Guantanamo.
Phase 1 : The swap-the-end-caps mod
This is terribly easy, and I'd seen it suggested by a few people on the lists as a good idea, so it is where I started. You drill out the old rivets, which is best done with an oversize drill until the caps fall off. The body of the rivet can then be drilled out with a 5mm drill, or sometimes just gently pushed into the interior using a 4-5mm drift and hammer. You can get replacement 5mm Stanley pop rivets from B&Q (in UK), although I welded M5 dome nuts on the inside of the end caps and used roundhead hex screws instead as I expected to pull them apart a few times. They look near identical to rivets. The stock silencers have an off-centre straight-through core that runs about two-thirds the length of the can. This part of the silencer is just like a race can, a perforated straight core surrounded by wadding.
The rear third of the can is just empty space, albeit lined with heatproof material. If you're familiar with 'decibel killer' end caps fitted to some race cans to make them road legal, the stock KTM design is just the same. A narrow-bore restrictor tube is welded inside the end cap and extends backwards into this empty space, adjacent to the exit of the straight-through core. Gas expands into the empty space and has to make a double U-turn to finally exit through the restrictor tube. The explosive pulses which contain most of the acoustic energy are therefore dissipated by expansion, and only reflected energy gets out of the restrictor tube - which is why the thing is so damn quiet.
The restrictor tubes are cranked so they are offset from the core, to achieve the double U-turn. If you remove the pop rivets and swap the left end cap with the right, the restrictor tubes end up pointing into the core instead, which gives a direct gas path. You could just turn the caps through 180 degrees, but they have KTM's part number etched on the underside. Obviously the smaller tube still can't flow anything like as much gas as the larger core, so it doesn't make much difference to efficiency, just sound.
The box spanner and studding was an improvised tool to pull the end caps off. Having tapped it into the end cap, this 'handle' allowed the caps to be wiggled free. The gaskets need greasing before you replace them, tapping them back in with a soft-faced hammer.
Sounds good? Well, no. I tried it and hated it. Yes, there was a bit more volume to full throttle use, but there were several problems. First is that the restrictor tube still does most of the job it's there to do. It functions like a port in a bass reflex loudspeaker, attenuating both low and high frequencies. Just like an organ pipe, its diameter and length determine its resonant frequency, and sound outside its pass band doesn't get out. This rather asthmatic and high-pitched sound does not change when the end caps are flipped. Nor should you expect any change in performance, since nothing substantive is changed - back pressure and resonance remain the same. But at least this means jetting is unaffected.
I might have been able to live with this except for the appalling racket on overrun. Shut the throttle, and you get a really nasty loud rattly, whistling farty noise that sounds a bunch of squirrels are fighting in your exhaust. It's frankly embarrassing.
Phase 2 : remove the restrictor tubes
As the above demonstrates, the restrictor tube is the bit of the design that stops the noise and gas getting out. The narrowness and length of the tube limits the frequencies so you get none of the characteristic bass bark of a meaty single. So long as that tube remains, that can't change.
So the next thing was to get rid of them, but without mutilating the end caps. This looked awkward from the outset, as they're held in place not only by a strap to the end cap rim (see next picture), but a very solid ring of weld deep inside the cap.
My son has access to both oxy-acetylene and a plasma cutter at work, but neither were any use at all because the available nozzles were just too fat to be able to attack the weld. There are several potential solutions. A fine-tip plasma cutter would do it in seconds. A hole saw would also do it, if you have one that will cope with extremely hard stainless steel. Most will not; Cobalt ones might, but I didn't have one. I chose to drill lots of holes around the restrictor tube from the inside of the cap, the traditional bodger's method. I since found out Jon Covington, on the KTM_duke Yahoogroup list managed to do the job with a Dremel and cutting wheels, which has to be easier. He removed the exit part of the 22.5mm restrictor tube with a large drill then used the Dremel to cut around inside the large tube to leave a shortened stub. This will need a very good 1" drill bit. I've never owned a Dremel, but this seems a good excuse to buy one as Jon said it coped well and only used 3 cutting discs, which are cheap.
In this picture, you can see the flange, like a washer, surrounding the weld, and it's that which
needs to be drilled - though keep the drill in tight to the weld to avoid damaging the inside of the end cap. Lots and lots of holes. It isn't fun, not least because that 'washer' is about 3mm thick and of unbelievably hard stainless steel. I told you these cans were well made. Before you can get a drill in, however, most of the protruding tube has to be cut away and the bracket removed. I did this with a hacksaw, angle-grinder, a small grindstone in an electric drill, and bad language.
This is the culmination of a lot of sweat and wishing I'd never started. I wrecked at least a dozen drill bits getting two-thirds around the first cap. That was using lubricant and drilling very slowly, too.
The trouble is that drill bits are nearly all cheap crap made in India and China nowadays. Even the ones that say 'HSS Titanium coated yada yada' are made of chocolate. If they don't blunt instantly, they snap. Eventually I found an old decent quality bit from the 1970's at the bottom of the toolbox and it did the rest of the job and the second cap without problems.
If you're going to attempt this, I strongly suggest you spend some serious money on a few decent bits, if you can find any. Any from a DIY superstore will be rubbish, and even branded like Snap-On are not that wonderful. I'm told Wurth do some very good cobalt-tipped, but they sure aren't cheap at about £20 each. It's best to drill a pattern of small diameter pilot holes of 2-3mm first, then open them out with a 6-7mm bit. A sharp cold chisel takes care of any remaining material.
The finished end cap, after cleaning up the ragged edges (not that you can see them here) with a 1" diameter grindstone in an electric drill.
How does it sound? Well, more like it should, at last. The exhaust note is far, far deeper and much louder - just on the socially acceptable side of utterly obnoxious. If you stand 20' behind the pipes, the power pulses will now part your hair. Cars dive for the gutter as I overtake now: on full chat it sounds like a V1 flying bomb.
As for performance, there is a marked improvement in pick-up from low revs especially. It's like someone has shaved a pound or two off the flywheel. Snap the throttle open and response is astonishingly instant. Midrange also seems very healthy and gutsy, with a newfound entertaining tendency to spin up the rear on exits from slow corners. Most if not all of the gain is below 4000rpm, but subjectively it seems a bit freer at the top end too.
I suspect the restrictor tube was engineered so it did most of its restriction at the noise test revs of 4,000rpm, suppressing midrange power as well. Equally its likely that its resonant frequency was toward peak revs when it will actually assist scavenging. I think you can hear this with the stock cans gaining a harder edge at higher revs. Not having put it on a dyno before or after, I can't be sure.
Phase 3 : the straight-through core
I rode the bike for over a year with the opened end caps, and gradually became unhappy with the noise levels. Passing police vehicles or pulling up next to them at the lights was not very comfortable. The bike was outrageously noisy above 80mph on fast A roads and motorways. Even my wife commented on how she could hear me from a couple of streets away. I started to get paranoid, cutting the engine and coasting the last 50 yards so as not to piss off my sleeping neighbours at 1am.
One time I took it up to the Ace Cafe on a supermoto night and it was much louder than any other bike with race cans, even those with the infamously loud G-force system. But also, it didn't sound as crisp as it should, it burbled and farted noisily on the overrun. That big empty space in the back of the cans worked like an echo chamber. It seemed unlikely to be the best arrangement for gasflow.
I persuaded myself a new straight-through baffle was a good idea. Probably, 2" (50mm) O.D. perforated stainless tube, as used by Team Incomplete, is better WFO, but the exit from the gutted end caps is 36mm unless you grind them back like they did to get a sharp-edged 50mm hole. I wanted to retain the stock look and keep the noise reasonable. Whatever diameter baffle I used would have to exit through that 36mm hole. A 2" baffle and exit has about 7x the area of the stock system with restrictor pipe, a 1.5" baffle and gutted end cap gives about 4x stock. This is not a race bike. I live and ride mostly in London. I measured up the end caps and bought some 1.5" (38mm) O.D. instead, standard 1.6mm wall thickness. This looked to fit just right inside the tubes in the front and back caps with a bit of packing made from offcuts. If the length was cut accurately, the baffle would be clamped between front and end caps and the packing ensure that nothing rattled.
It's awkward and time consuming to measure exactly so I'll save you the trouble : you need 2 564mm lengths, each with one end cut square (for the end cap), the other mitred at about 60 degrees to fit the inlet cap without obstructing gasflow.
The packing pieces need to be about 40-45mm long, cut lengthwise so they can be opened up and sprung over the baffle ends. I made each mitred piece slightly conical (see picture) so it would wedge in tight to the inlet cap, then welded both to the baffle
The 38mm tube matches the exact inner diameter of the exit from the end-caps so fits tidily. It just needs packing at each end to fit snugly inside the tubes within the end caps.
Here you can see the packing pieces have been opened up, sprung over the baffle ends and welded in place. The weld is then ground back to allow the ends to fit inside the front and back end caps. Both ends are tight enough that the baffle can't move.
For packing I mostly used stainless steel wool. This is much better than glassfibre wadding because it should last indefinitely. Do not use ordinary mild steel wool, which will burst into flames and/or rust to nothing very fast. Stainless steel wool seems to be elusive in the UK in less than industrial quantities, and I had to buy from a US supplier, though still far more than I actually needed. Shipping cost as much as the wool.
I had some glassfibre exhaust packing made by BigGun laying around from a previous bike project, so used an outer wrap of that. It isn't necessary, though it may keep the can surfaces a bit cooler. It should last indefinitely because the steel wool takes most of the battering from the hot gas pulses.
If you don't use stainless steel wool, just glassfibre, the cans will need repacking every year or two. Please don't be tempted to use glassfibre sold for building insulation purposes, it's not at all the same material and will be blown to bits within a week with high-speed fragments of glass being squirted out of your exhaust. Not the sort of thing that's good for eyes or lungs. Most bike dealers sell the proper woven stuff, and it's cheap.
Once you've wrapped the baffle, it's reinserted into the casing. It should be a fairly tight fit if you have used enough wadding.
Here you can see how the mitred end of the baffle fits into the inlet end cap so as not to obstruct gasflow. Check it is in this position before you re-fit the outlet end cap, which clamps the whole assembly up tight.
The stuffed can with the baffle in place, just the end cap to put on.
Close-up of the end cap shows how the new baffle fits, looking reassuringly like a factory job. It is an interference fit, not welded, so the end caps can still be removed easily. Of course the caps still have the 'E1' marks underneath, certifying they are legal, decent and honest. But the main thing is that they're quiet enough and look ordinary enough not to provoke a full bottom inspection or rectification notice from the men in blue. They'll just have to find something else to nick you for, admittedly not hard with Dukes.
Subtly different from stock...
The understated home-made bungee hooks are the only other mods to the bike, and were essential for me. Mostly I have to carry a large bag of cameras on the pillion, and had to have some rear attachment points for an 8-leg bungee 'octopus'. It's just a bent metal tube with eyes welded in each end, welded to a flat plate through which the rear exhaust mount bolts fit. It was only supposed to be temporary whilst I figured out a better solution but has proved to be stronger than expected and very useful.
The motorsport.ds Y-pipe fitted well. Unlike some exhaust components I've met their jig assembly is spot on. It looks a hell of a lot better than the rusty KTM piece. One tip for Y-pipers : get a new KTM graphite gasket seal, for where the Y-pipe fits onto the header. I spent ages prying out the old one to reuse it, and it's near impossible to remove without destroying it or lacerating your fingers on the wire-mesh reinforcement.
So what does it sound and go like? Nice! It's much quieter than it was with just the gutted end caps and is unlikely to upset my long-suffering neighbours or the MOT tester. It emphatically doesn't have that stifled-fart quality of the originals, the cans now sound crisp and ...right. The overrun is cleaned up nicely too. Carburation seems spot-on with the KTM gauze lid and jet kit.
The low and midrange punch is perhaps just a little less than it was with the gutted end caps alone, though it's hard to be sure given of the psychological effect of a witheringly loud exhaust, which this no longer is. The Y-pipe may have made a difference here, as they have a reputation for enhancing the top end at some expense to the midrange. But overall I am very happy with these cans now.
They are louder than Akra's with decibel killers fitted, but not as loud as Akra's without - about halfway between the two, which is what I was aiming for. They work well, with no sense that they are choking the engine unlike the stock cans, it is smoother and stronger at low revs and works better right through the range. They're a pound or two lighter without the complicated KTM innards, too.
Was it worth the effort? Yes, definitely. If someone offered me some Akra's free, I'd grab them - but that's not going to happen. These mods cost very little except time and effort, and much of that was wasted exploring, measuring things and thinking about the best way to do things. I removed both end caps believing I'd have to do that to remove the OE baffle. This was unnecessary. You can leave the inlet end caps rivetted to the body, and after removing the 2 bolts that secure the baffle, get hold of the baffle tube with pliers and pull hard and the whole baffle assembly will come out of the outlet end in one piece. A squirt of WD40 between the body and baffle guts will help.
It's all pretty easy for someone who has moderate DIY skills and tools, once you know where you're going. Removing the restrictor tubes from the end caps is a bit of a sod, but once that is done the rest is simple. I welded the packing pieces to the ends of the new baffle, but you could probably just glue them in place with Araldite, or solder them. Once the can is assembled it won't matter if heat destroys the bond. They aren't going anywhere, friction fit should keep them in place.
Materials cost came to about £35 / $50 for the pair, or would have done if I had not had to buy 5 times more perforated stainless tube and stainless wire wool than I needed. The cost saving over a pair of Akra's is enough to buy a Y-pipe AND a Keihin FCR41 carb! I may get around to the carb next... On the other hand, I have another pair of cans and do slightly wonder what they'd be like with 2" cores stepped down to 1.5" just before the end cap. But then I'd have to derestrict another pair of end caps to try it, and I dread having to do that again...
Incidentally, there is a useful guide to carb mods, and swapping the end caps (as I did initially), at the UK KTM forum. The writer quotes me as saying the bike may be a 'bit flatter' at the bottom end, after the addition of the straight through core. I think he has misundesrstood : my comparison was with the (very noisy) derestricted end-caps alone. As soon as those restrictor tubes are gone there is a large and very noticeable imrpovement in low and mid pick-up, If a little of that gain is lost after adding the baffle, it is still much better than with the OE end-caps, flipped on not.