Project - Machine to measure punching power

bannor32

Hi there. I've been doing some research trying to come up with the best (and most economical) way to build a machine that measures the punching power of a boxer. What I want to do is have them hit a speedbag, and be able to get a numerical readout on a·4 digit LED·(the harder they hit the higher the number). This number by itself will have no meaning - it doesn't have to reflect pound force or anything like that - but will be accurate in relative terms to someone else hitting the speed bag. For example Mike hits the bag and gets a reading of 109, and then if Bill hits the bag 3 times as hard he should get a reading of 327. From asking around the best method I'm told would be to attach an accelerometer to the back of the speedbag (or insert it in the center) to measure the force of the punch. Once the bag is struck, I would like it to take a continuous sample over about a two second period and the number it ends up displaying would be an numerical interpretation of the peak·acceleration measured over that period. A second 4 digit led display would display the highest score to date. I have been searching the web exhaustively for a simple/cheap solution to this, and came across the·Basic Stamp 2. I don't have much experience with electronics, but have some programming background, and from what I've read so far·the learning curve is pretty gentle compared to some other solutions.

Would·the·Basic Stamp·be able to achieve what I am trying to accomplish?

Here is an example of what I am looking at doing (only mine would have a much more simplistic physical design):

http://www.coinopexpress.com/products/mach...rs%29_5932.html

Again, I am looking for simplest method possible for getting data from the accelerometer directly to a·microcontroller and being able to manipulate that data to output a numerical reading that displays relative punching power. Any advice would be appreciated.

Regards,
Dennis

Beau Schwabe

I think that with an accelerometer on the punching bag, and the known weight and 'pendulum length' of the bag, you could determine
how much force the bag was hit with a few calculations. There are several examples on how to interface an accelerometer to a stamp.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

---

Beau Schwabe - Mask Designer III

National Semiconductor Corporation
Latest Company News
(Communication Interface Division)
500 Pinnacle Court, Suite 525
Mail Stop GA1
Norcross,GA 30071

---

bannor32

Edit: Here is an updated link to the boxer machine I had referenced above:

http://www.coinopexpress.com/products/machines/sport_games/Boxer_Punch_Machine_(Plain_Colours)_5932.html

allanlane5

I would think the "cushiness" of the bag would throw off your figures a little -- the accelerometer approach would work fine for 'in-elastic collisions', where all the force of the blow went into accellerating the bag. A boxer punching a bag has very elastic collisions -- the gloves and the bag padding.

Having said that, when boxers punch hard, even if there is a lot of padding there's still some residual accelleration. Come to think of it, it is that residual accelleration applied to the other boxer's head that is the goal after all.

Conclusion -- yes, this approach may give you some useful numbers. Note the 'pulse' of accelleration may be quite short -- on the order of 10's of milliseconds. I would assume the goal is to apply as much force over as little time as possible. Force = Mass * Accelleration, so if you know the mass of the bag and can measure the accelleration you can get the force.

P.S. Yes, the BS2 can be used in these conditions.· It cycles fast enough and measures pulses as part of its native 'PBasic' language.

Post Edited (allanlane5) : 4/7/2005 6:36:14 PM GMT

Beau Schwabe

allanlane5,

I would think that the moment arm length or pendulum of the bag will also need to be taken into account with the mass

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

---

Beau Schwabe - Mask Designer III

National Semiconductor Corporation
Latest Company News
(Communication Interface Division)
500 Pinnacle Court, Suite 525
Mail Stop GA1
Norcross,GA 30071

---

Larry~

Something I ran into with a similar project was where the bag is being hit

a shorter person with the same force hitting lower on the bag may make the bag move

differently than a taller person hitting higher on the bag ( bag hanging )

bannor32

Thanks for the information. I don't necessarily need to measure the actual force being generated. Basically I just want to produce an arbitrary number that will be consistent in relative terms. I want to be able to allow competition to see who can produce the highest number when they hit the bag. If one guy hits the bag 1.5 times harder than another guy the number displayed should be 1.5 times as high. As long as that remains consistent that is all I need it do be able to do.

KenM

Create a solid link in the path of the chain/rope that is suspending the bag. Outfit the link with a strain guage?

Comments?

steve_b

Having had "my Smile handed to me" on a few occasions...there's a difference in a quick jab and a power punch.
So the weight of the bag plays a bit in this.

Someone like Mike Tyson who, doesn't just bite off your ear but, punches 'through' you wouldn't necessarily have a fast punch but would do more damage. Whereas a Sugar Ray Leonard or Billy "the kid" Irwin (canuck fave) would have a fast punch with less power but enough to move the bag and score points.

So, you might want to have a boxing game and then like a Home run derby (or a "knockout" derby) for the hardest punches.

I played a game similar to this a ways back....and there was a screen in front of you with a boxer dancing around. The idea was to hit him when he was vulnerable. The trick was to hit him hard and fast as soon as he became open, to score the most points.
You might want to use a reaction timer as well as a green light that's on for a certain amount of time. The faster they hit the bag when the light goes green...the more points they'd win.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·

Steve
http://ca.geocities.com/steve.brady@rogers.com/index.html
"Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."

glaso

I think the accelerometer signal will be plenty of noise, and it will be very dificult to "clean" because of the short period of the signal.

A better solution will be to measure the velocity of the bag. This can be done attaching a solid link to the bag and attaching a pot to it on the other end, then we ad a pull down resistor and measure with the command "pulsin" the time which takes the bag to do the full arc.

Bean

I would use an optical sensor to measure how long it took for the arm attached to bag to go through the sensor. This will give you the bag speed, without cables and such.
Bean.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"SX-Video Display Module" Available Now.

www.sxvm.com

"A problem well defined, is a problem·half solved."


Post Edited (Bean) : 4/8/2005 12:45:35 AM GMT

steve_b

I'm trying to remember how this other "Punchout" game was rigged!

I think they had an encoder attached to a level (of which you hit one end). The lever had a large coil/spring that would make it harder to push the further you got.
So, as the lever got hit it would obvious increase a POT or increment an encoder position.

This also allowed for the bigger slower guys to push in the the spring more and win more points....
You could have a slotted opto to measure speed (timed break of light beams)....and also have an encoder for the force....

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·

Steve
http://ca.geocities.com/steve.brady@rogers.com/index.html
"Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."

Tom Walker

bannor32,
As L_Gaminde, Beau, and allanlane5 pointed out, there are several variables that will come into play. And, unfortunately, they will all be working against you in trying to come up with a number that means "anything" in relative terms. If you imagine a really powerful hit striking somewhere near the top of the bag, and then imagine a somewhat softer blow impacting near the bottom of the bag, the physics of it may very well register a larger number for the second hit (assuming the bag was suspended from its top point, that is). This can be negated by having a very small target...and a heavier target would take care of some of it.

Ain't physics fun?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Truly Understand the Fundamentals and the Path will be so much easier...

bannor32

Wow. Lots of interesting suggestions. If I decide to go with the accelerometer method does anyone have a ballpark idea of what range I would need? I was looking at one that did +-10 g. Would this have enough range, or should I try and source out a high g accelerometer?

Larry~

On second thought this may be usefull
if the bag had a differently colored area
that was the punching area and was able to
be moved up and down you could set up the bag
for anyone, you could also use this to show that
different punching hights relative to there waist or
sholders gives different punching power and could
use this to determine each persons sweet spot

Paul Baker

Well according to physics F=ma, so depending on the mass of your bag you could calculate the acceleration undergone by the bag given a punch force in newtons (N). Karate chops that break boards and concrete is in the couple of thousands of Newtons, I would imagine this is around the upper level of force which can be imparted by a bare hand. Assuming a 1 kg bag (2.2 lb) this would mean an acceleration of ~2000 m/s², or ~200 G's. Since this is an elastic collision some of the force will be imparted to the deformation of the bag so your actual measured acceleration will be a little less.


I am slightly concerned with the idea of an accelerometer. Since this is a tightly teathered object, the window of accurate acceleration measurement will be quite small. To get an accurate reading you will have to do a multi-sample during this time window. Without going into the physics of determining what this time window would actually be, Im not sure if the stamp will be fast enough to get reliable readings with every punch.

Post Edited (Paul Baker) : 4/7/2005 8:57:59 PM GMT

Guido

What about a pressure senor somehow hooked up to the inflation port?

Klutch

greetings,

as luck would have it, i am also in the early stages of a similar device for a martial arts instructor friend of mine.· after much hunting on the web a week or so ago·i came up with a product close to what i want, and i think exactly what you want:· http://www.imptec.co.uk/html/product-information.htm

"When strapped to a kick bag, a GOOD strike ON TARGET will show about 1.00 Imptec units
and very good strikes will register over 2.00 Imptec units."

"The unit measures from 0.00 to a maximum of approx. 4.00 units in 400 measurement steps."

searching a martial arts forum bounced me to something that alluded that they use a simple accelerometer, +/- 10 g i believe.· the imptec·thing is their arbitraty impact unit that i feel is basically a slight of hand to avoid the mish mash of potential·issues with lever effects·as beau mentioned, deformations of the bag, mass of bag, contact time (long for a boxer, very short ~5 ms karate strike), etc.·thus boiling it down to simply a scaling of acceleration.· i personally wanted an actual·force·output as it seems more meaningful, but i dont think it is as simple as F=MA due to those issues, could be wrong though.

back to the stamp, i havent given it too much thought yet, but·i·suppose you could·continiously scan with·PULSIN to gather the·duty cycle from a +/- 10 g·accelerometer like the memsic MXD2002 or analog devices ADXL210E, if PULSIN returns a value > 0G (> 50% duty cycle) + threshold·there must have been a hit·that needs to be scaled to be output to an LCD/LED, ignore all < 50%·duty cycles·as that is the bag swinging back.··the threshold is a magnitude "lockout" to prevent registering "false hits" as·the bag oscillates back and forth.· quickly looking at the datasheets for those accelerometers, they·are good for·50,000 (MX) and 1000 (AD) g's of shock survival·and·are available·up to·10ms duty cycles which should be no problem for BS2(PULSIN res @ 2us)·to measure, process, and display prior to another strike???·

food for thought i hope, input/comment??·

ross

ps - another website i had found·showed various·applications·that their accelerometer was used for, one was to test·a 5x world karate champ, as i recall·he·kicked w/ ~2300 lb force and punched w/ ~1900 lb force!

KenM

Lets consider another approach.....

Who is going to view the "punch stats"? The fighter, trainier or both?

As you mentioned, you are not interested in a "real" number, but at least something that is relative to the punching power.

Parallax sells the flexiforce pressure type sensors.

·http://www.parallax.com/detail.asp?product_id=30056

Can we assume the area of the glove that will strike the bag is relatively small....?

Maybe mount the flixeforce sensor in/on the glove itself and run wires to the monitor unit (could get messy though), or use an inexpensive RF unit to transmitt the data.

http://www.rentron.com/remote_control/TWS-434.htm

?????

bannor32

Ideally I would like the person who is punching to be able to hit the bag with a bare hand. It is intended more as a novelty device than any sort of sports training/data device. You could put it on display anywhere there is a gathering of people and anyone can walk up and try it, and see how they compare to their buddies. Above I referred to the person who would use the machine as a 'boxer', but that was just to help convey the type of machine I was visualizing. The link I provided above to the punching power machine available from Coinopexpress is pretty much exactly what I have in mind. I came across one at the arcade at the Luxor hotel in Vegas, and people were playing it like mad. I gave it a couple tries, and it was pretty cool. I looked it up when I got back, and was shocked to see the price tag attached to it (it is a pretty simple machine) $4500! I figured I could build one for a tiny fraction of that price, and that has led me on a chase to try and find a reasonably flexible but economical method of gathering the data I need, and the Basic Stamp 2 seemed to fit the bill. I'm a bit overwhelmed by all the great ideas that are being thrown out here, but it is leaving me more uncertain than ever whether or not this will work like I want it to.

Tracy Allen

How about a flat target, then, instead of a punching bag? Fewer degrees of freedom. An economical sensor for something like that might be a peizo film. The MSI technical manual is pretty interesting, including mention of impact sensors for baseball:
www.msiusa.com/piezo_documentation.htm

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

StarMan

I agree with Tracy.· A single degree of freedom·would be·easier.· A velocity sensitive pad such as those used in electronic drums (and keyboards for that matter) comes to my mind.· The harder you hit, louder the sound.· Instead of sound intensity, output a number.

You can put as much padding on the sensor as you want so it won't break anyone's hand.

Post Edited (StarMan) : 4/8/2005 6:04:43 AM GMT

bannor32

I had considered the option of having a flat padded target, but a couple of problems kept me from delving further into that possibility. First of all I was worried that the sensor would be too likely to get damaged, considering it is measuring violent impact (though Piezo film wasn't an option I was considering at the time), and my other concern was potential injury for participants. With a flat target that wouldn't have as much "give" as a speedbag would people could damage their hands or wrists by striking the target too hard with improper form.

m_fabio2

Didn't they do this in Rocky 4 ?
·

steve_b

Hey guys....
I have to say I'm still quite the kid and love arcades! My first job out of college was in a "family entertainment center" (arcade! lol).

Part of the fun in playing these types of games is finding their "sweet spots"....this is what enables a little kid to beat a giant. If the kid accidentally hits the sweet spot while the big gorilla is pounding away with full force. I wouldn't worry too much about the imbalance in energy transfer from hitting the top of the pad or the bottom of the pad.
If you wanted to void that (sorta) then don't use a 'swinging' target...or a levered target. Use what equates to a piston. So that it's horzontal to the strike path....but I'd leave it so that there's some imbalance!
You all remember the Hercules hammer thing....hit the pad and it sends the weight flying up....and if you hit the bell you're hercules! Well, if you hit further out on the levered pad, then the energy transfer was greater and it went higher!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·

Steve
http://ca.geocities.com/steve.brady@rogers.com/index.html
"Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."

StarMan

To expand on Steve's thought, you could mount a pad on a cantilevered linear slide system (a couple of rods and slide bearings) with a return spring and a dashpot (shock absorber).··A fluid pressure sensor could easily detect changes in the pressure inside the dashpot.

Check out a company called Pacific bearing for rods and bearings.· I believe Mc Master-Carr carries some of their stuff.

bannor32

I've been looking at the Flexiforce sensor that KenM mentioned above and I am wondering if this will work if inserted into the face of the speedbag. Mark the bag with a target (strike zone) overtop of the area where the sensor is. That way accuracy would factor into getting a higher score. You wouldn't have to 'try to find the sweet spot' - the target would point you right to it. I realize that because of the free swinging nature of the speed bag that most of the applied force would disperse when the bag starts to move, but that moment the fist makes contact and "deforms" the bag it should get a reading of force, shouldn't it? Again, it doesn't need to accurately measure the actual force, simply get a reading that can be used as a relative comparison from punch to punch.

http://www.parallax.com/detail.asp?product_id=30056

steve_b

Wha'ts the maximum compressions those flexiforce sensors can take? I'd think that if a big gorilla hammered the bag, he'd crush the sensor. I just got a 'pressure' sensor yesterday (pressure being weight, not atmohspheric)...they wouldn't be good for this idea as they can't handle a lot of weight and they only have a mechanical movement of about half an inch....you could use a levered system and have a shock-absorbed plunger making contact with this sensor....either sensor really!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·

Steve
http://ca.geocities.com/steve.brady@rogers.com/index.html
"Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."

bannor32

I was checking our the Tekscan website (they manufacture the Flexiforce) and according to their spec sheet their high force range sensor can handle up to a 100 lbs. There is an asterix beside it with a note that it can measure forces up to 1000 lbs by reducing the resistance of the feedback resistor.

http://www.tekscan.com/flexiforce/specs_flexiforce.html

I would want to install this under the surface of the bag, so that would absorb some of the impact to the sensor, and because there is nothing behind the bag, and it moves freely with the punch that should help absorb any potential damage to the sensor.

Beau Schwabe

From experience of using these type of sensors they should remain perfectly flat. Make sure that they don't get creased or distorted.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

---

Beau Schwabe - Mask Designer III

National Semiconductor Corporation
Latest Company News
(Communication Interface Division)
500 Pinnacle Court, Suite 525
Mail Stop GA1
Norcross,GA 30071

---

the other one

Steve said...
I'd think that if a big gorilla hammered the bag, he'd crush the sensor.

Steve, you hit the nail on the... eh... you hit the bag on the sweet spot. I have worked for an arcade company for almost 20 years and "gorilla" is my usual terminology for the mighty force that is the idle young male who has so much energy he should be out ploughing a field with a collar around his neck or something.

Keep that in mind, bannor32, as you build. Our customers shake bolts loose (use Loc-Tite). They break welds (use an experienced welder). They kick cracks in particle board, tear tubular forms in half, bend steel reinforced steering wheels, unscrew knobs and take them for souvenirs. Do not underestimate the player... especially if he is half drunk, feels he has something to prove, and wishes to use your machine to do it.

Back to your engineering... the problem with a sensor in the bag is not the sensor itself but preventing the connections to it from being broken or becoming intermittent. For any mechanical game, shock and vibration will be your nemesis. Whatever you put in, make it as simple as possible. Do not allow wires to be stretched; make them long enough. Prevent wires from moving relative to the sensor, as they will simply snap off after a few flexes. Use finely stranded wire, which is less likely to break.

NOT to discourage you from your project... but the Boxer game is pretty well made as boxing games go. If you do buy a Boxer game, stroll out to the hardware store and buy a bottle of Loc-Tite(tm). You'll need it.

(Hint: ask for a local dealer at the manufacturer's site, www.jakar.pl)