Functional threshold speed training

[bikenrrd]

While messing around with heart rate to power and speed to power (on a turbo) calculations, (in this thread: https://yacf.co.uk/forum/index.php?topic=74756.0) I realised that instead of doing all this I could just use speed to train (on the turbo). The response of my turbo (an Elite Powermag) is so linear, with respect to power vs speed, that if I am doing x km/h on the turbo, I know I am doing y watts.

So, instead of using lactate threshold heart rate or functional threshold power, I think you should be able to use function threshold speed.  This is probably not an original idea.  The process would go something like this:

Set up the turbo trainer so that the rear tyre is pumped up to a certain PSI (I use 100).
Set the resistance level to something that will let you do your maximum pace over 30 minutes without spinning out.
Always use this PSI and resistance level when future training.
Do the 30 minute threshold test - 20 minutes warm up, 30 minutes at a pace you can sustain over the 30 minutes.
Record the speed over the last 20 minutes.
Your functional threshold speed is 95% of the average speed over the 20 minutes.
You can then just plug in your speed into one of the many online power zone calculators and use these zones for training on the turbo.

Much cheaper than a power meter!

[vorsprung]

this is true

You must have a fluid turbo though

[bikenrrd]

this is true

You must have a fluid turbo though

I thought the fluid turbos had a non-linear response?  The curves I've seen on the web look exponential.  I have a magnetic turbo, which has a very linear response.

[Chris S]

Interesting. I'm looking at ways of improving my climbing - and as I run out of puff long before I run out of legs, FT improvement is something I could probably do with.

I have a cheap Cycleops linear Mag TT that would do the job, and a cold draughty garage to set it up in. Hmm...

(goes off to ponder bikenrrd's protocol)

[Ningishzidda]

As long as you absolutely know the kmh vs Watts relationship of the dyno rolls, set up the laboratory as repeatable as possible to get reliable comparative results.
Warm up by the same routine each test.

It don’t matter if the Velocity/Power curve is linear or follows the RoadLoad polynomial, when you read off your average speed for 20 minutes, the intersection on the graph tells you your power.

Good dynos have a polynomial relationship CURVE. Better dynos allow you to trim that curve to simulate the bike on the road. Even better dynos have a flywheel that is sized to simulate the inertia of your bike, and electrically simulate the RoadLoad curve.
The bestest best dynos, electrically drive the rear wheel round to simulate the bike freewheeling, and do all the road simulation electrically.

[Ningishzidda]

Pulled off the Kurt Kinetic website.
-------------------------------------------
Try It Yourself

Figure out your Functional Threshold Power. Be sure your body is rested, this is like a short race effort. Allen and Coggan's protocol is as follows:
1.   20 minutes easy warm up
2.   3 x 1-minute wind ups with a minute rest between (100 RPM pedal cadence)
3.   5 minutes easy
4.   5 minutes all out (hard at first, but not so hard that you can't complete the effort)
5.   10 minutes easy
6.   20-minute time trial effort (like the previous 5-minute all out effort, keep in control, hard but steady, you don't want to over cook it and die at the end)
7.   10 to 15 minute cool down

Rob Hayles gives a suggestion to what “easy” means, and it is breathing with your mouth CLOSED.

[bikenrrd]

I forgot to mention:
Repeat every 4 weeks to measure your improvement.  You should see an improvement and will have to adjust your zones accordingly.

As long as you absolutely know the kmh vs Watts relationship of the dyno rolls]

I don't think this is important, so long as the km/h vs Watts relationship is consistent, i.e. for speed x, your turbo always produces speed y.

set up the laboratory as repeatable as possible to get reliable comparative results.
Warm up by the same routine each test.

This is very important!

It don’t matter if the Velocity/Power curve is linear or follows the RoadLoad polynomial, when you read off your average speed for 20 minutes, the intersection on the graph tells you your power.

The above protocol assumes you don't have access to the graph.  It's not using power at all, just the implied relationship between power and speed.  This is why having a linear response turbo is important. 
If your FTS (functional threshold speed) is 20 km/h, then 110% of this on a linear turbo is 22 km/h.
Suppose we do know the Watts per km/h and at 20 km/h it is 150W (7.5W per km/h).
Then, with a linear response, at 22km/h (110%) it is 165W.
If the speed / power curve isn't linear - for simplicity let's say it is 0.37 * (km/h)^2, then at 22km/h you are doing 179W.
So you can't use simple percentage zones like this.
If you do have the power / km/h curve you can read off the speeds from the power.  But let's assume we don't!

[thing1]

This is probably not an original idea.

Indeed, this is pretty much spot on how the Virtual Power feature on Trainer Road works: http://www.trainerroad.com/features/virtual-power
And using that you don't need the indoor trainer to have a linear speed:power curve so long as it is a model they've sampled the curve for. The main requirement is that the trainer doesn't change curve as it heats up: they recommend the Kurt Kinetic Road Machines being good in this respect.

DC Rainmaker has a good rundown on it here