söndag 24 november 2019

The Gamified Timed Static Contractions Extended

Quite a while ago, when I had hit a stubborn plateau with my counter-weight assisted Chin-Ups, I decided to replace them - at least for a while - with "gamified" Timed Static Contraction (TSC) Pull-Downs and Pull-Overs as per Drew Baye's book "Timed Static Contraction Training", since it had worked very well for me with my gamified TSC Squats.


Above, you can see my combined rig for the Gamified TSC Pull-Downs and -Overs. The Phidgets' load-cells are the same as I use with my Squats, as are the off-picture bridge to my laptop. The DIY handles (PVC pipe, chains, and carabiners) I use with a underhand grip for the Pull-Downs are old ones from another, ancient project. I added the webbing loops for the Pull-Overs (left in Swedish is vänster and right is höger, hence the V and H - they need to be marked as if they are switched, they won't be level with each other, due to differences in the hooks and chains...).

Only this week, I also got around to craft myself a piece of webbing for Chest-Presses - yepp, I've kinda-sorta hit a plateau with my Dumbbell Bench Presses lately...


It was tailored to be of the right length for Baye-style TSC Chest-Presses but - by sheer luck - it also work for me for his TSC Compound Row. I wonder if that is generally true for the majority of people or if it is more specific to my anatomy?

On the photo, you can also see my newish Phidgets Wheatstone Bridge and Vint Hub taped together with silver tejp. My old bridge, that could be connected directly to a computer, was dead as a Dodo one morning workout (not a good workout, spent waaaay to long trying to revive the Phidgets bridge...). The new Vint Hub also forced me to upgrade to the Phidget22 library and - since it doesn't have any official Ruby support - hunt down a third-party integration in order to adopt my existing Ruby script to work with the new Phidgets library version and hardware. On the plus side, it seems like the new ones are a bit more stable.

With equipment to do gamified TSC Squats, Pull-Downs, Compound Rows, and Chest Presses, I'm only lacking one for Standing Presses - then I would be able to do a complete Big Five workout of the Doug McGuff/John Little "Body By Science" variety but in a Baye:esque TSC interpretation. Perhaps I should dig out the old sewing machine again? ;-)

tisdag 1 januari 2019

Gamifying Timed Static Contractions with Phidgets Load Cells, revisited

So, I've been on my "gamified" Timed Static Contraction (TSC) Squats I outlined in an earlier post almost weekly for half a year now and thought I should to a write-up on my experiences so far.

The Executive Summary would be: has been overall successful but there is still some room for improvements.

I've modified one of the example Ruby scripts that ships with libphidget21 to output measurements from both load-cells as soon as they're available for 95 seconds and are using the following scheme,

while (1) {
        replot
        pause 0.2
}
to five times a second re-plot the graph of my load-cells output in a Gnuplot script, to get the interactive feedback needed to be able to adjust the force I'm exercising against the unyielding band to stay at or near the target level.

The very first time,  in mid June, I didn't have any target resistance to try to meet but just used the set-up to measure how I really did on a conventional, "blind" TCS Squat:


You can see that I've opted for 45 seconds, 30 seconds, 15 seconds phases rather than the more normal 30-30-30 ones. That can always be a future progression, to use equal length phases. Basically, I just scrapped the initial pair of phases on this try-out and only estimated the last phase to have come in on around 50 units on both load-cells (hopefully, the S-cells should measure kilograms but let's just speak about generic units in case the calibration is off or the cells inaccurate). Hence, the following week, I set the new target to 55 units for the final phase and half of it (27.5) for the first pre-exhaustion phase and three quarters of it (41.25) for the second:


Now I had to, over a couple of seconds, ease into the target resistance of each phase and stay at it, as indicated by the horizontal blue target lines. As you can see, the green load-cell tracked the targets better than the red one. Later in the year, I would have considered this workout a bust due to the wanting red line, but as this still was during the trial weeks of the Phidgets TSC Squats, I allowed it and set the target of 60 units for the next week:


Still questionable results for the red load-cell. However, I also noticed that my stance was way too high in the last phases, so for the fourth workout, I shortened the loop around my waist with about 10 centimeters, but still quite optimistically/aggressively increased the target with another 5 units to 65 for the following workout:


This time, the overall result was more on track, despite the shorter loop/lower stance. Let's fast forward to the seventh workout, the first time I clearly failed to match the target of 80 units:


Following that, I redid 75 units successfully next week and but then failed again at 77.5, so I then settled for 1 unit increments - 76, 77, 78, ... - which worked well until the 15th time and the target of 81 units:


Here it wasn't really I that failed but my equipment. One cable attaching the red load-cell to the bridge broke... So I had to re-attached the cables properly to the bridge and then secure the cables with duct-tape to prevent future material exhaustion in the cable cores... The week after, I successfully passed 81 units and have been increasing one unit a week ever since, up to the last workout before the holidays, the 23rd one with the Phidgets, where I did 88 units:


So what will the imminent improvements and/or tweaks be?
  • One trivial one would be to label one of the load-cells "left" and the other "right" to always use the same one with the same foot/leg - but in practice, I don't think it matters that the cells might switch foot/leg from workout to workout. After all, they're supposed to be calibrated the same.
  • I'm still in a tad too high stance in the last phase, so I need to film myself in profile during the squats and shorten the band until my thighs stay at a 45 degree angle to the floor throughout the third phase (but let's wait until after I've hit 90 units with the current length).
  • The rubber training mat I'm using to stand on actually doesn't offer enough friction for my feet to stay put during the high exertion during the third phase. When I focus on maintaining the full pressure into the band loop, my feet tend to slide a bit closer together, making the exercise easier. To combat this, I either need to switch from the training mat to something more rugged and rigid - or place a piece of plank sawed off to keep my feet shoulder-width apart when placed between them (I think I'll go for the latter - also once I've hit 90 units with the current suboptimal set-up).
  • Since my load-cells only goes to 100 kg and I never imagined that I would ever surpass 2 x 100 kg, I might have to progress to a deeper stance than thighs 45 degrees to the floor - or splash out and get new load-cells that goes beyond 100 kg (Phidgets' next larger ones go to 500 kg which surely would be a total overkill).
  • Naturally, another way to make the current load-cells last longer is also to increase the length of the third phase - that would also decrease the level of resistance I would be able to handle. The 30-30-30 second protocol Drew Baye recommends would be the natural choice, but another possibility would be to opt for maintaining the same, more conservative load for 90 seconds - i.e., no pre-exhaustion phases but just one marathon session where the challenge would be to struggle to keep the load at the target for the complete duration. Actually, it would be something fun to try but it would be hard to guess what target load to go for? Perhaps 70% of the current third phase target? Or is even less a must? Or is more possible to handle?
  • I also need to simplify re-checking the current calibration of the load-cells. Currently, I basically just run the normal workout script and hang a few dumbbell weight-plates to each load-cell while they're hung from my chin-up bar. So far, with 30 kg in weight-plates, they've always been less than 1 kg off  - but I really should check closer to the max weight, using my own body weight or so. (This is why I've written "units" rather than "kilograms" in this post - I fear they're not 100% accurate, but as long as the trending is correct and they're not too far of each other, they still work as long as one consider them as measuring generic load units rather than exakt SI kilograms).