It definitely doesn’t drop too much when I’m at 100%, it actually continues to increase slightly and by the last 30 seconds of the final interval I’m hanging on for dear life.
Then just keep doing it. Sorry if you have already tried this, but if you are not sure when you are on vo2max hr effort try keeping the 130% for like five mins. When you see your hr stabilize, thats probably where you’re on vo2max (correct me if i’m wrong). I have experimented a little with this. A few weeks ago i did 3x5min efforts on 120% ftp where i reached a hr of about 180 that held for about three minutes. I did an ftp test with a five min all out effort yesterday, starting off at around 130% (intending to keep it for the duration) but ending up with a 126% average with a hr of around 185 for half the duration. So for my own sake i would aim to reach something above 180 and as close to 185 hr in the future, probably kicking off at a high percentage of ftp and dropping to about 115% or something.
Edit: btw i also tried something similar to what you are doing, except i was doing the overs @125% and the unders @90 and my hr was not near what i was able to on the other sets. Might try your workout just to see what happens.
Didn’t read all the responses…
Being a full time career and family man I try and keep it simple. For me work at VO2max levels is done with both power and HR. There is no formula. I spool up the power at slightly different rate to around 90% HRmax depending on a 3, 4 or 5 minute interval. Throttle the power to feather the HR as time progresses thereafter.
The largest problem as I age is getting good data on HRmax. For example, now 52, been using power since Enve became a company, my power metrics are PR’s but, HR has changed so much in 13 years it’s really shocking.
Anywho…I’ve also learned over the years, training longer VO2 efforts, while in the lab maybe not as perfect, it is vital to feel that level of suffering for relatively long periods of time.
Cheers and thank you @empiricalcycling for adding your insight here. Amazing!
Not really stoked on this study, or the protocol. Nobody trains like that, and the biggest foul is using vo2peak and not tossing those out and only using true vo2max numbers. And how they matched the “work”. Plus we don’t know much about what else anyone was doing besides freeform whatever training. Who the fuck would do HIT 7 days in a row if you’re not in a week long cyclocross stage race anyway? 
They tried to replicate real overload training, didn’t succeed. And a 20k TT isn’t how I’d measure effectiveness of that kind of training anyway. Remember the TP article on “miracle intervals”? That author reported what looked like good improvements, but when you broke it out into watts, they really were the same as you’d expect from someone doing a more “normal” approach. There’s no such thing as a miracle intervals, and HIT stuff is not miraculous. In fact, it’s very easy to overtrain doing too much HIT, say… 7 days in a row.
Also, yikes, don’t take a class on metabolism from Krispen.
In case it’s not clear, I’m entirely unimpressed by 80% of the exercise physiology literature, but quite happy with the lit elsewhere. Just because it gets published somewhere doesn’t mean it’s any good, or that the reviewers understood it or even gave a shit.
Inspired by this thread, I modified the Kaiser workout so the intervals have a hard start, and then decline through the interval.
I’m going to play around with the parameters of this workout and over time dial into a workout that gets my HR up quickly for each interval, then holds at mid 90s% of HRmax.
Interval duration: 3 minutes
Starting power: 130% FTP
Ending power: 100% FTP
Rest duration: 3 minutes
If you are like me, think you’ll find that it requires a much harder start. And I prefer to have a hard start followed by steady state like posted earlier:
as it better simulates getting my turn at the front, pulling into a headwind, and then dropping back into the echelon/paceline.
Yeah - I’m going to play around and see. The “harder” start versions probably create more oxygen debt that needs to be repaid during the rest of the interval. Maybe that’s a better approach to spend time at VO2 max vs the declining intervals in the workout I posted 
I’d start with at least 155% of FTP. For intervals >2 and up that gets me to VO2max in 30 to 40 seconds. After that, high & plateaued O2 consumption can be maintained with much lower power…like 110%. Anyhow, that worked for me.
So…for example w/ an ftp of 225 start out with 30 seconds @ 350 then 3 minutes at 250. Rest, recover, repeat.
I did a workout like this yesterday, for reference.
btw, TrainerRoad library workout Dorr +5 is something very much like what I used for VO2max workouts.
How reliable is the 90% HRM target? Asking because I’ve gone through a bunch of my various VO2 workouts and calculated time above 90%. The difference between same workout indoors and outdoors is striking. Enough so that I am thinking of doing repeat workouts in and out to create a better dataset (increase N) to rule out other effects. Perhaps being sick, over tired, etc.
I go by breathing/feel/RPE and only look at power and HR after the ride. Hard push to create huge demand for oxygen, and then breathing lags behind by 30-90 seconds depending on what I was doing prior.
Haven’t had time to read the whole thread, what different training response would you expect to see using Hard Start vs Steady VO2 max? Would the av power for both be the same?
My understanding is the hard start puts you into oxygen debt and as a result, you spend a higher fraction of the interval close to VO2max.
TBH, I think the difference between steady VO2max intervals (eg 3 mins at 120%), declining power intervals (eg start at 130% and drop to 100% over 3 mins), and hard start VO2max intervals (eg 30 seconds at 150%, 2.5 mins at 110%) likely falls in the “marginal gains” category. The main thing is to be doing the VO2max intervals in the first place - and finding the flavor you like Improves adherence.
@DaveWh Thanks!
as mentioned above the other aspect of hard start that appeals to me - it better simulates what happens on the road during Wed night worlds or a spirited weekend group ride. In other words, I don’t view it as marginal gains, I view it as practicing what happens to me on the road when chasing stronger riders.
Yep - my “marginal gains” comment above was about impact on training VO2max. There’s other reasons why one form of interval is better vs another - eg race specificity.
How long was the hard start and rest of interval (4minutes?)
Usually I wanted to hold 1 minute with power on level with 2-3 minute power. I have done whole block with different lenght of the intervals, this was only one of them as an example (3,5 minutes).
I did 1 minute at 5 minute max followed by 5 minutes around 90% of my 7minute max. Just looking to do shorter intervals
For me personally HR has quite big fluctuations, even with the same external conditions (outdoor, similar temperature, similar terrain,…). Fatigue and activation play a huge role. In my 32’ test, which I did at relatively constant power, I was at >90% HRmax after less than 2min.
[I mentioned time at >90% HRmax above, but just to give a rough idea of the intensity. I hardly use it to control my training. Or at least, not the absolute numbers - the variation during a workout with respect to the “baseline of the day” can be quite interesting (e.g. power/HR decoupling).]