Equipment - Homemade Rocking Trainer

Last night I did my first ride on my home made rocking trainer. The goal behind the rocking trainer is to let the bike tilt left and right, like it does when riding out on the road.

The two obstacles (well three?) to a rocking trainer are:

1. Does it rock realistically?

2. Does it allow you to ride it without tilting inadvertently?

3. Does it provide enough resistance?

Realistic Rock?

The first is the main goal of the trainer, to be able to rock the bike realistically. What a lot of people don't realize is that when you rock your bike side to side while out of the saddle your bike actually tilts around the bottom bracket area, relative to your path. Meaning, yes, your bike tilts left-right with the tires as the axis, but when taking into account the slight wiggle you get when rocking the bike back and forth, the bike is really going in a straight line from about the bottom bracket area.

This means the tilt axis needs to be higher than the tires, closer to the bottom bracket. Your body is mostly stationary during out of saddle efforts (if your body moves then that's not the best), you're basically standing on the pedals, and the two pedals share one thing in common - they're anchored at the bottom bracket axle. This makes the bottom bracket the ideal pivot point.

If the axis is too low then you get the "my tires are stuck in a groove and I'm going to fall over" feeling", which is what it would be like if you actually pivoted around the tire/road junction.

Problem is if you have the axis too high then it gets a bit weird. Imagine if your head was the pivot point - the tires would slide back and forth a few feet as the whole bike/trainer assembly pivots around your head.

Can You Stay Upright?

The second is sort of crucial. Bikes stay upright because you can steer the bike while you move forward. Basically you catch yourself falling by steering into the fall, like if you're tilting right you steer right and now you're not falling, you're just turning right.

With a trainer there's no forward motion, meaning your bike isn't moving forward and therefore you can't really steer into a turn because you're not turning per se. Therefore it's a bit trickier to keep the bike from tilting.

The thing that helps to tilt the bike is to be able to steer the front wheel. The head tube angle and fork rake encourage the bike to lean if you turn the bar while the bike is stationary (by lowering the front of the bike). That allows the bike to tilt a bit where it wouldn't otherwise tilt.

My homemade rocking trainer.

The brown thing under the wood is a spacer to raise the left side just a touch.

Eventually I think I'll use wedges between the two pieces of wood.

Due to the forward hang of the trainer the base has to extend forward a bit. I made my base out of wood because I could, and I don't have welding skills.

Note that the folding arms and their mounts are removed from the front of the trailer (normally under where the trainer clamps the wheel). This was an all-or-nothing experiment.

I had this trainer frame after warrantying a blown up electronic trainer through CycleOps. Although I requested they send only the resistance unit (a Fluid2 since they discontinued the electronic unit at the time) they sent me a complete trainer. I suppose it's probably easier logistically to do that instead of stock a bunch of resistance units. Anyway end result was that I had an extra trainer frame. They don't wear out so I decided to use the spare to see what I could do with it.

View from above.

I've been thinking about the trainer and how to make it better. I think one thing would be to make the pivot point (where the plates are, see pictures below) below the bike, not behind it. This would support the bike better. Right now the bike dips a bit when I pedal, more when I pedal hard. It's to the point that I expect the tire to hit the floor/mats at some point.

Different view of the rocking area.

Although I asked my car friend to cut down the plates a bit, he ended up welding the plates in a different order. I wanted the top plate to be in the middle, so it could clear the other two plates while it rocked. It's okay, nothing hits so far.

(I gave my car friend a piece of plate steel in exchange for doing this work - the plate steel was probably 2'x3' and had to weigh 50-80 pounds... it was really heavy but it's stuff he could use for his car restoration hobby.)

Plates and bushings.

The plate angle was designed to allow the bike to pivot around 1/3 up from floor relative to BB.

The guy who welded everything painted it also, without me requesting the paint. Very nice.

You can see here that the plate welded to the trainer frame (middle one) is wider than the top plate. This was an error. The guy helping me out inadvertently swapped the two plates. If the narrower top plate was welded to the frame there'd be more room for it to rock.

I have to increase the height of the bushings as the trainer doesn't rock enough so I don't think I'll run into interference problems between the middle plate and the bottom one. If I do I'll have to try and cut down the middle plate. It's very tough to cut - the guy who welded it actually asked someone else with a plasma cutter to cut out the pieces for him. I don't have a plasma cutter. I may visit a machine shop nearby and ask them to cut the piece down.

A stock Kinetic Road Machine resistance unit.

The tire-roller junction is the weak point of the whole set up. I haven't done an all out sprint yet, just a few rolling jumps to make sure nothing would break, maybe 600-800w each. Nothing like 1000-1200w, and definitely no 10% jumps. However the tire slips if I do any kind of a super hard jump.

I got the KK Road Machine unit from a fellow racer. Note the spring - it's not curved around the corner. When I got the trainer the spring was curved around the corner. I didn't realize it but the various pieces of metal holding the old trainer were slightly bent. This meant I couldn't get as much clamp power to hold the roller to the tire. Now it's much better.

Note: This is a Kurt Kinetic resistance unit. The frame is a CycleOps frame that used to have a Fluid2 resistance unit on it. The frames and resistance units are interchangeable based on what I have experienced. Apparently KK has some deal with CycleOps where the frames are the same. Different mechanisms for adjusting tire tension (I prefer the CycleOps lever system although I haven't tried the two companies' newest systems) but obviously the resistance units fit one another's frames.

For the front a heavy duty lazy Susan with the front wheel holder.

This allows the front wheel to turn easily; I've turned it a bit to illustrate that.

I need to mount something to the metal lazy Susan, probably a thin piece of wood with the tire block mounted to it. The lazy Susan is an industrial strength one rated to a few hundred pounds, not something for the coffee mugs in the kitchen cabinet.

If I do that I'll need to raise the rear a similar amount, to keep the bike level. This isn't a bad thing since it would increase tire clearance to the floor.

How Does It Work?

Right now it's not working really well.

1. The rocking motion is too limited side to side. I feel like I want to move the bars another 4-6 inches to each side. That might be an exaggeration because I didn't use a tape measure or anything, but I don't think so. For sure the bike still feels too rigidly upright.

2. The bike doesn't tilt easily enough. I have to force the bike down to the side. It doesn't drop down naturally. I'm not sure what the solution is - maybe some bushings to the side of the main axis, soft ones, with the larger central bushings cut down a bit? I don't know. I'll have to think about this for a bit.

3. The rear tire gets very close to the floor.

4. The skewer has to twist in the trainer. Since the trainer frame is tilting forward on each downstroke, however slightly, the skewer twists inside the cups. Ultimately I'd like to have some kind of bushing type material or a separate cup for the skewer ends that rotates on its own.

5. Lack of peak power. The tire-roller interface is still the weak point of the whole set up because it slips at about 1100-1200w. This limits any peak power kind of efforts.

Going Forward

My ultimate trainer set up would be a reduction gear trainer like the one here, but that's a pipe dream for now. The challenge here is to mimic the inertia you have to overcome when accelerating without drivetrain/tire slip as well as rocking the bike like you do outside.

With the Kinetic Road Machine resistance unit you get some of that - the inertia and the exponential resistance. However there are two limitations.

First and foremost there is a definite element of drivetrain slip, aka tire slip. If I do a max effort downstroke the tire slips every time, so I've learned to tone down my initial jump by 300-400w. That's not a good habit to get into and it also prevents me from exploring max effort jumps indoors. I'm pretty sure this is part of the reason why my peak power is a bit lower nowadays - I've trained myself not to make big jumps nowadays. In 2008-2010 I was training outside a lot, I was doing many massive jumps, and I regularly hit significantly higher peak power numbers, 1400w whenever, 1550w on a great jump. Nowadays I think it's a big deal if I hit 1250w and many of my jumps are more like 1100w.

The other is the incorrect rocking motion. Without the bike tilting as much it just doesn't feel the same as it does on the road.

I can't do much on the drive train slip limitation. I can, though, work on the rocking bit, so that's my homework going forward. I'll also finish up the front tire mount as well.

Equipment - Modded CycleOps Fluid Trainer, Step 1

Along the theme of "stuff I want to get done this winter" is this idea of modding a CycleOps Fluid Trainer.

One of the biggest shortcomings of a normal trainer is that it doesn't rock side to side. Way back when there was a mild attempt to make such a trainer, manufactured and sold by Technogym. A friend had one and I tried it briefly. To use it you took the front wheel off and clamped the fork into a spring-mounted "fork mount". Two springs, about the size of a coil spring you might find inside an old suspension fork, allowed the "fork mount" tilt from side to side.

I don't remember much of the device except that the springs forced the bike back to vertical with too much power - I felt like I was on a coil spring playground ride rather than on a bike on a pretend road.

Kurt Kinetics has a better design. Theirs rocks from side to side at a natural height, about halfway between the hub and bottom of the tire. Ideally I think a bike would rock around about the bottom bracket - if you watch a rider coming directly at you while rocking the bike you'll see the bottom bracket follows essentially a straight line, the tires carving small arcs under it, the rest of the bike waggling above the bottom bracket.

Having said that I decided that modding my CycleOps to resemble a Kurt would make sense. I'm sure they tried putting the pivot a bit higher and found whatever they found and therefore decided to put the pivot a touch lower.

Having tried one as well as watching numerous other try one, I think they got it right.

In addition the CycleOps and Kurt basically share the same "frame", meaning the trainer frame. There's a connection between the two companies that, although I don't know the details, results in the fact that both companies use what I consider to be the best (aka most rigid) trainer frames around.

A final vote in favor of using the CycleOps frame - I have an extra one. Yes, an extra frame. I had a Fluid trainer and a Power trainer, both by CycleOps. The power trainer, which used a proprietary head unit matched with some electric motor/generator resistance unit, literally started smoking one night, the smell of burning electrical stuff filling the room. CycleOps, to their credit, sent out a Fluid unit as a replacement. Although I asked for just the resistance unit they sent a whole trainer, including the frame. This left me with an extra frame with no resistance unit.

In the meantime I'd switched resistance units to the former-power frame (and for the life of me I don't remember why I thought this was a good thing). This left me with a gray Fluid frame for modding purposes.

A scrap piece of metal gave me the raw materials needed to mod the frame. The thick metal plate would have felt at home as a side skirt on a WW2 tank. It weighed a good 60 or 70 pounds and it was only about 15"x15". Let's put it this way - I had a hard time carrying it on my own.

I recruited the same guy that painted my red frame. He can do some very basic welding and in fact I'd been thinking about having him fix some of the white van's rusty areas. He, in turn, recruited a local metal artisan to cut the metal into smaller pieces with a plasma cutter. That artisan, incidentally, covered his whole house in metal, and who made a local bike sculpture. With the raw plates in hand (the rest of the plate was essentially payment to the artisan) the painter guy could start his work.

First, though, I had to tell him what I wanted from the project. The painter is not a bike guy, and in fact he lights up a cigarette if he doesn't have one already in his mouth. He understands mechanical stuff but really doesn't understand the bike riding part of bike riding.

I tried to get some angles and fit type things in place. I planned on using a 2x6 as a wide base for the trainer, with wooden extensions reaching forward. The regular folding legs won't work because they'd lock the trainer and prevent it from rocking. I know that if I had a 2x4 under the front wheel it's about the right height off the floor, so I figured that if I "fitted" everything with the bike flat on the floor then I could raise/lower it in "2x4" increments.

Thoughts on height, plate angle.

Regular folding legs are the lower tubes, with the black caps on them. They'll go away.

You'll see that the bike's rear tire is sitting on the floor. I don't have the rest of the bike in the frame, just the rear wheel. I wanted to get an idea of where the trainer would sit, what angle the arms would hit, what sort of angle I needed on my "rocking plates".

Gusset shape, if needed.

I had no idea how strong the welds would be so I figured we'd need a gusset plate. I knew that I had given the painter an extremely heavy piece of steel, significantly thicker than the plates used in the Kurt. Plus if the thing broke I'd just topple over, it's not like I'd be going 60 mph in a tuck.

Probable placement of plate.

After a lot of debate I decided that putting the plate under the U-tube would work best. It gave a lot of surface area to the weld area, it would clear the controls of the resistance unit (the spring loaded lever thing), and it gave me enough height off the ground to give me room for the additional plates necessary to create the non-rocking part of the frame.

The guy welding didn't do the gussets immediately and I told him, after checking things out, that they didn't seem necessary. The welder guy did paint everything so it looks semi-pro. He also shaved the original leg mounts, for the folding legs. Although I wasn't keen on that it does clean things up. It also commits me to trying this out.

Bolting things together, the various plates in the right order.

Wood works really well although it looks pretty ghetto.

I didn't get a 2x6 piece of wood but I had 2x4s left over from my garage organizing binge; I decided to use them instead of going out and buying another piece of wood. This seems to be totally fine, very rigid and secure. I used galvanized carriage bolts, galvanized washers, and stainless steel nuts, all in interest of their anticorrosion properties.

(2x4s are 1-1/2" thick so I bought 3-1/2" carriage bolts to give me 1/2" for threads and washers.)

The carriage bolts come up from underneath - I use a gym mat type thing under the trainer so the rounded head underneath won't hurt anything, and on a rug it won't hurt either. This allows me to periodically check the nuts on top without having to tilt the whole trainer on its side.

Now one error the welder made, not realizing how things were going to work, is that he welded the wide plate to the trainer stand. I meant to have the wide plates sandwiching the narrow plate, so there's more room for the rocking motion. He also assembled them in the wrong order, which to me illustrated that he had not fully realized the idea of the whole thing. This is my bad.

My first trial ride ended unsuccessfully inside of two minutes. I used OEM spec rubber spring tower bushings from the now-gone 1993 Honda Civic. It's a light car, 2000 lbs or so, and the bushings are pretty soft. I bought polyurethane bushings for the car but I actually installed them in the car. I thought I had extra bushings and went looking for them. See, back then I bought the shock install kit as well as a "full suspension" kit, but apparently the full suspension kit didn't include shock bushings.

Note the downward tilt (the rest of the trainer is to the left).

Rubber bushings, not polyurethane.

At any rate the rubber bushings allowed too much tilt to the front. This exacerbated something that already happens to the Rock N Roll. The tilt allows the angle to change between the trainer support arm and the bike, forcing something to flex a bit. It appears the skewer moves within the locking arm but the skewer could easily rotate on the frame. The latter would prematurely wear out the dropouts on the frame, not a good thing.

The solution, at least temporarily, is to insert a nylon washer to act as a (bearing) bushing. A better solution would be to use a thrust bearing, typically used in a clutch assembly to allow the clutch to slide back and forth while allowing the shaft to spin. Thrust bearings allow the shaft to rotate (or something to rotate around the shaft) while supporting mainly a side load. Modding the skewer holder to accept two thrust bearings is a bit much right now but at least I have something to think about.

My plans for step two include a couple different things.

First, I need to get much stiffer bushings, probably polyurethane, hopefully a bit taller. A friend who is revamping his car's suspension my have some used poly bushings for me but if they don't work out then I'll just go and buy some poly bushings. I may get larger ones if I buy them, like bushings that fit between a chassis and body (typically in trucks).

Second, I'd like to tackle that pivoting issue with the skewer and skewer holder. There's a great site McMaster Carr and they sell all sorts of hardware. I've bought things like sway bar clamps, stainless steel license plate screws, and even suspension nuts (metric fine thread), and they have an assortment of thrust bearings. I think just having a thrust bearing between the skewer and the skewer holder will work out fine.

For for now that's where I stand. Everything has worked out well so far except for the too-soft bushings. After I fix that I'll see how the trainer actually works when I do out of saddle efforts.