Test bed for rowing machine

I guess I'm going to need some additional details, like I don't see any elastic cords here...

https://www.concept2.com/indoor-rowers/dynamic

In a sentence, how does your concept improve on the linked machine?

All commercial rowing ergometers use an elastic cord (usually) or a spring to return the handle on the recovery portion of the stroke. The one in your photo has a long elastic cord, together with the chain drive, back and forth between pulleys and concealed in the horizontal black box at the bottom of the machine.

Gravity, unlike an elastic cord, never wears out, never goes out of adjustment, and is unaffected by ambient temperature. It is a more elegant mechanical solution to the problem of chain take-up and handle return. Everything in my machine is open to view. The mechanics are pleasing to the eye and to reason. The machine in your photo has a labyrinth of chain and elastic cord looped between myriad pulleys. It is concealed in a horizontal box because it is ugly.

....but isn't it true that the rowing force needed also varies according to speed, I mean starting out the rowing is slow and difficult, and as the speed increases the dynamic changes from force to speed, and at speed the resistance is less....so looking at it this way seems that the flywheel properly damped would seem to mimic the actual conditions....and perhaps water currents would also play a role, either with the current or against it...

Using the motor to mimic a rower could also be a problem if you are measuring the forces by amp draw or total wattage, which is more dependent on a properly sized motor, with a varying load you are changing the efficiency....it would work better I think with a DC motor...

Could you provide a simple drawing of the device...?

So your concern is about consistency of measurement, force in, work out....and of course, simplicity of design, if I understand you correctly...

I don't have a drawing of the device. I have some ideas.

A 300 watt average stroke by stroke power input by a rower translates to about a 7 minute 2k distance on my machine and on the one you linked. That is not elite athlete time, but it would suffice for the test I describe. I would like to build a simple test device that would pull the handle with a constant 300 watts of power (or thereabouts), 30 strokes/minute, approx. 5' stroke length. Then I would adjust the handle return force (not the flywheel damper setting) and observe how this affects the PM readouts. You are right, adjusting the handle return force will change the load and affect delivered watts by the motor, so I would need to able to control motor power so that it remains a constant, otherwise the test is meaningless. Maybe I should see if I could salvage motor and controls from an old treadmill exercise machine.

..."What is the relationship between heart rate and workload?

When the degree of an activity increases, the heart rate and workload (the amount of work placed upon an individual expressed in kpm, watts, or oxygen uptake) will also increase.

Couldn't you measure the heart rate of the same person working on different machines several times to see if they are the same...

The elastic cord used to take up the chain and return the handle in rowing ergometers usually require 5-10 pounds of force to stretch. The variability of the cord strength from machine to machine, I argue, will affect user's times. I don't expect the time difference to be much, but 5 seconds over a 2k distance can mean a win or a loss. It's too small to pick up with a heart rate monitor. The other problem is the variability of human performance. A human cannot hold a steady watts output during an activity. A motorized test rig can ensure that constant power is applied every stroke.

A DIY rowing ergometer test rig is from my half-baked ideas file anyway. I read about a motorized test rig for rowing ergometers that has recently been developed at the University of Ulm, Germany. Very sophisticated. Can even shape the force curve of the stroke to approximate that of an elite rower. It uses precise sensors to measure force and speed at the handle. Derives the delivered power from Power=ForceXVelocity. This can then be compared to the PM readouts on the rowing machine. I thought maybe I could come up with something not quite as sophisticated, but with sufficient accuracy to still answer my question about the effect of differences in elastic cord strength. The Ulm team could answer the question if they ever choose to do the test, but they are moving at a glacial pace (all experiments, data, conclusions, peer reviewed and so on) and are not forthcoming about future tests they plan to do. Thought someone here might have related experience and suggestions about how to build a go-cart version of the Ulm Ferrari.

If the un-elegant tensioner is truly elastic, then there would be no work done, is that not true?

So if you could measure the efficiency of a typical spring or elastic, you would get an idea of the magnitude of the work.

If there are many sheaves or pulleys, then there is work created by the bearings and friction. If you put the machines in an enclosed controlled environment and measured the temperature rise of the air in the enclosure over a specific time frame, you might get an idea of the differences between the two methods. Or it might be easier to test the components that show friction, and then add the temperature rise of the air caused by motion through it. Or maybe calculate the losses of each component?

Re: "If the un-elegant tensioner is truly elastic, then there would be no work done, is that not true?"

That is true in a machine, but as I wrote in my reply to Randall, this is taking the human/machine analogy too far.

Re: My comment that an elastic cord's elasticity is affected by ambient temperature.

In indoor rowing competitions this is not a problem. They are held in large gyms. The problem can occur for the home user of these devices. Where I live, if the machine were set up in an unheated space, it would be unusable for several months of the year, because below a certain temperature the elastic cord will not take up the chain and return the handle with sufficient briskness.

Re: Friction of various components

If the machine is well maintained it is reasonable to assume that the work required to overcome friction of components will be a constant. So although that work is not measured by the performance monitor, it will not affect parity between machines. The work done to stretch the elastic cord is also not measured by the PM, but the strength of the elastic cord, is not a constant. It degrades with use and time. Does this affect parity between machines? I contend that it does (see my reply to Randall). Many users of these machines report that one machine feels "heavier" than another, that they record better times on one machine than another. I say that differences in elastic cord strength is the probable cause. Others dismiss my argument and say that the sense of any difference between machine is all psychological, or that the complainers probably just didn't eat their Popeye spinach that day.

To measure power,you need a torque transducer and a speed transducer on both input and output shafts.

This will allow you to obtain the data you need for the difference in power requirements.

The transducers are available with digital outputs which will allow computer monitoring and measurement.

Good luck.

Torque and speed transducer. Interesting. Thankyou.

You also may need an absolute encoder for speed which can detect the motion in forward or reverse rotation,with the difference displayed by software program when measuring speed,to capture any backlash from the stretching of the cords.

If you wish to measure only watts with each configuration,there are plenty of watt meters available at reasonable prices on line.

All this is useful information. Thanks.

There are many programs that will analyze and compile your data in any format that you desire.You will be able to present proof of your concept much better than a simple explanation or demonstration.

If I were to be convinced to change from a currently accepted and verified model,I would need the data from many runs,vs the competition.

There would have to be some unique features to justify the change.It is hard to break into a large base of an established market.

Your potential customers will likely require the same.

Inventing and marketing are very different,and specialists in each field are required to get the best results.

I know.Been there,Done that.

Re: "Inventing and marketing are very different, and specialists in each field are required to get the best results."

I know this. I have never had any intention or interest in manufacturing and marketing this device myself. An attempt to do that would be a good way to go broke. I applied for patent protection because I believe my innovation significantly advances the art (to use patent language), and I could afford it. A couple of years ago a U.S. fitness company contacted me, interested in buying my intellectual property, but we couldn't meet in the middle, and the negotiations came to naught. That's okay. If it never goes anywhere in my lifetime I will not be angry and bitter. I am glad I did what I did. It was an enjoyable project.

Things move slowly in the rowing world. Rowers and cyclists are a different breed. Cyclists are always hungry for innovation. Rowers are suspicious of innovation. Example: I first tried a flywheel-type rowing machine in 2006. It was a C2 machine. I had heard nothing but praise for the perfection and wonderfulness of that model. Within two weeks I had a list of its design deficiencies. I couldn't believe, for example, that for the past thirty years users had been uncomplainingly pulling on a straight stick of a handle (and I later learned, denying that the handle was the cause of the chronic wrist pain and injuries suffered by many of them). The handle is a human/machine interface, and at any such interface the machine should adapt to the natural movement of the user, not the user to the machine, as is the case with the rigid, single-piece handle. In a rowing magazine I came across a photo of a U.S. professor of biomechanics on a rowing ergometer. The photo had captured him at the end of the stroke, his wrists cocked at an awkward angle, his wrist position in violation of everything he knows and professes - but he had a big grin on his face.

It is not just the conservatism of rowers that is holding things back. For decades the rowing ergometer market has been dominated by one company, and that has also stifled innovation in the field. Static machines, as opposed to dynamically balanced machines, are forty-year-old technology, but it is static machines that are used worldwide in all the indoor rowing competitions, and those machines are made by that one company. So that company has a vested interest in perpetuating the use of a technology that should have long ago gone the way of the dinosaurs.

Wouldn't it be easier to just calculate the work needed to stretch two springs/elastic cords with different characteristics?

Sorry to throw in a diversion here, but, I've been thinking about this a bit recently. How about replacing the chain/cord with a rod, and arranging for the rower to drive the flywheel on both the pull and push strokes. By increasing the active muscle groups, it might be possible for an individual to do more total work.

You might be able to extend this to to a real boat by simply attaching springs between the oars and a point behind the rower. This way the rower does more work on the push/inactive stoke, and, the springs then help the rower in the pull/active stroke.

Both ideas might need better anchoring for the feet.

In the 1970's a company in Norway (Gjessing-Nilson) manufactured a flywheel-type rowing machine that used a rod (partially seen in the photo) to drive the flywheel via a one-way clutch, on the pull stroke. It free-wheeled on the recovery portion of the stroke. Although true that more work would be done if it drove the flywheel during both pull and push, and possibly have exercise benefits, that does not replicate the work done during actual rowing. These devices are used by many for general fitness, but they are also off-water training tools for rowers.

The pictured machine gave a good replication of the feel of rowing but required a lot of floor space to accommodate the drive rod. In 1981 Dick and Peter Dreissigacker, and Jon Williams, patented a rowing ergometer that employed a chain drive with an elastic cord for chain take-up and handle return. The prototype was built in their barn using salvaged bicycle parts. That was the first Concept 2 machine, and the mechanical strategy it employed became the pattern for most subsequent designs by others.

Re: "Wouldn't it be easier to just calculate the work needed to stretch two springs/elastic cords with different characteristics?"

That calculation can be done, but it doesn't resolve the argument among rowing ergometer users. Most now accept that since the performance monitor (PM) does not measure the work done to stretch the elastic cord but only the work done to spin the flywheel, then a stronger elastic cord will result in a lesser portion of the user's total power going to the flywheel than if the machine were fitted with a weaker elastic cord. They accept this but argue that it doesn't matter, because although the stronger elastic cord will require more effort by the user on the drive portion of the stroke, it will make it easier on the recovery portion of the stroke, so it all balances out, and a level playing field remains. I disagree. I say that this is taking the human/machine analogy too far. Everyone knows that pulling on a strong spring is more fatiguing than pulling on a weak spring. On the return, the stored energy in the spring/elastic cord assists in the human movement, but in no way does any of that energy flow back into the muscles, re-strengthening and re-invigorating them.

The question of whether a truly level playing field exists is important. Indoor Rowing competitions have become a big deal. It is now an event in the World Games.

I see what you are saying about the work put into the elastic component, how it cannot be measured by the flywheel motion. So that work is solely put into the drive line equipment, and could represent a significant energy input, due to the basic inefficiency if the design, as you have been saying.

So a simple force transducer (load cell) in line with the elastic element would record the work, (maybe start at Omega Engineering) integrated over the session, and give you the differential work input. They come with a bridge, and can be connected to a laptop with a simple I/O module, like from Acromag, Weedtech, or Beckhoff.

I appreciate your suggestion. I knew such components are available but have no experience with them. I made a similar suggestion on a rowing forum: Connect a force transducer between the elastic element and the frame, and feed the force data to the performance monitor. The PM measures the user's stroke rate and length, so the rate of doing the work (the power) to stretch the elastic cord could be calculated by the PM and added to the measured user power to drive the flywheel. That would give the total user power input, which would be used to calculate and display the pace (what distance in what time). Differences in elastic cord strength would then be irrelevant. Do you agree, and in your view is what I describe a workable strategy?

My suggestion was pointedly ignored on the rowing forum on which I posted it. There is yet to be an acknowledgement that there is any problem that needs to be corrected. All of the machines used in sanctioned indoor rowing competitions are built by Concept 2. Concept 2 has a huge market influence, with a cult-like loyalty among C2 enthusiasts. To them, the C2 rowing ergometer is perfect, and the founders of the company are omniscient. Users have an unshakable belief that parity exists between machines, and any attempt to point out that what they believe might not be true, is heresy. Even the governing body of the World Games seems infected with the belief in Concept 2's omniscience. I wrote to them and described the possible problem of parity between machines. They thanked me for my "opinion".

Yes, I think that would work. Not sure how you could test a population of machines, but if there was a fairly common linkage design, you should be able to fashion a transducer connection that would be fairly short, and insert it on a group of machines, and get some hard data to make your point. Maybe 25 repetitions on the same machine by different people, get the total force over time, and then same test on an adjacent machine? Then test for temperature dependence, while you are set up.

There are several manufacturers of rowing machines, but it is the 'static' Concept 2 machine that is used in indoor rowing competitions, and at the World Games, so that is the machine that needs to be tested. Mechanically, all the C2 units are the same. Parity between machines of other brands is important, but not as important. Many athletes prefer other brands for training, but none of those units are used in international indoor rowing competitions.

"Then a stronger elastic cord will result in a lesser portion of the user's total power going to the flywheel than if the machine were fitted with a weaker elastic cord. They accept this but argue that it doesn't matter, because although the stronger elastic cord will require more effort by the user on the drive portion of the stroke, it will make it easier on the recovery portion of the stroke, so it all balances out, and a level playing field remains."

They accept that a difference exists, but argue that it doesn't matter, so measuring the difference won't help.

I agree with you that it does matter, but I'm not sure that building the test bed will further your cause.

Re: "...I'm not sure that building the test bed will further your cause."

I'm not sure either, but I'm interested in different ideas about how it could be done. As I wrote elsewhere in this thread, Concept 2 (manufacturer of the units used in competitions) enjoys a cult-like following and loyalty. It is almost impossible to convince C2 enthusiasts that even the slightest design flaw could exist in their beloved machine. For example: Many, many users of the machine suffer from chronic wrist pain. For some, it becomes so severe that they are no longer able to row. But no one ever blames the rigid, single-piece handle design. They always blame the injured person. They say that the injury was caused by poor technique. I reply that technique cannot overcome deficiencies in equipment design. I equipped my "SlideWinder" machine with an articulated handle of my design (you can barely see it in the photo in my original post). It maintains a bio-mechanically correct alignment of the hands, wrists, and forearms, throughout the stroke, eliminating the cause of wrist injury. Generally, in the Concept 2 community it has been greeted with derision and mirth. The attitude being that if Concept 2 didn't think of it, or implement it, then the idea whatever it is must be without merit.

Here's how I understand it...

The rowing machine simulates a boat. Your machine will simulate the rower.

The spinning flywheel simulates the momentum of the boat which the rowing action adds to. The springs are just there to pull back the cables of the linkage on the return stroke, but they add to the work the rower must do.

Forget about the flywheel and just consider the energy required moving back and forth against a spring force. If friction is negligible, the mathematical calculations say that the energy required equals zero over a complete cycle. Obviously, this is not realistic for a rower as you cannot put energy back into the muscles. I would think that the energy provided by the rower (and your rower simulator) would only be the energy pulling against the springs and ignore the return stroke.

So, maybe drive a wheel by your motor through a gearbox, and record the current only through the half turn of the wheel on the pulling stoke. The current is proportional to the motor torque. The wheel would be part of the mechanism shown below that would be connected to the oars.

Thankyou Rixter. I'll give that some thought. I like the simplicity.

I appreciate that you accept my argument that you cannot put energy back into the muscles. It is not a difficult concept, but users of rowing ergometers refuse to accept it (see my post #15 reply to rwilliams). They can't get past the machine analogy that the energy required to stretch the elastic cord goes back into the system upon release, and therefore insist that over the complete rowing cycle of drive and recovery the energy required equals zero, regardless of the strength of the elastic cord. They so fervently believe that parity exists between machines, and want to continue to believe this, that they pretend not to know what everyone knows - that pulling on a strong elastic cord requires more effort and will tire you more quickly than pulling on a weak one.

This is what I have come up with (side and top view) based on your suggestion. Thankyou. I worked through the force vector geometry from 0 to 180 degrees in 10 degree increments and from that it seems this device will generate a decent replication of the force curve of a human rower. Initially I drew the pull cable/strap (extending to the right) passing between two sheaves and then running horizontally, but I found it gives a better shaped force curve without the top sheave. Not a difficult construction. I'm thinking a salvaged variable speed DC treadmill motor would work fine.

The test rig in Ulm, Germany utilizes a stepper motor and linear actuator. The design team put considerable thought into the control system to produce force curves similar to rowing athletes. I like that from what you suggested and from what I have drawn, it appears that a rowing shaped force curve would be generated, not from sophisticated electronic control, but simply from the structure and mechanics of the device. The purpose, of course, is to eliminate human variability. A human cannot deliver a stroke over stroke constant power. I am optimistic that this device if built, could. It would generate stroke over stroke identically profiled force curves - displayed on a monitor specific to the test rig, along with strokes/minute, power in watts, and so on. These would be compared to the readouts of the monitor of the rowing ergometer under test.

The Ulm team in Germany had institutional funding. For me this is a personal interest project. I still haven't decided if I will actually build it. It's encouraging though to have what seems to be a viable concept. I have found someone with a similar interest who has the knowledge and experience to put together the electronics, data acquisition and processing side of it, so maybe...

I welcome critical comments from anyone here.

...."Rowing Boats do not move at constant velocity. Velocity is constantly changing over the course of a stroke in a rhythmic pattern. The change in velocity is called acceleration.

In that variation lays the answer to rowing efficiently and increasing the average speed at which the boat moves towards the finish line. The average speed for one stroke-cycle follows a simple rule:

any variation in velocity increases the work necessary to overcome fluid drag in the water. Minimize these fluctuations and you will be able to move the shell faster with the same energy expenditure. This is, in essence, our optimization goal in rowing."...

https://www.rowinginmotion.com/optimize-your-rowing-stroke-with-rowing-in-motion/

https://worldrowing.com/2019/01/16/the-avatar-how-video-for-rowing-coaches/

Re: "Rowing boats do not move at constant velocity."

The velocity fluctuation cannot be eliminated. In addition to the intermittent thrust of the oars, the rower pushes against the footrests during the drive portion of the stroke, and pulls against them during the recovery portion of the stroke. The mass of the rower and the mass of the boat oscillate in dynamic equilibrium.

In 1995, Casper Rekers, a Dutch engineer, patented the first "Dynamically Balanced Rowing Simulator". It better replicated on-the-water dynamics than any prior art. The flywheel and footrests were mounted on a sliding carriage, and the rower sat on a sliding seat. During use, the rower and the carriage moved together and apart in dynamic equilibrium. Typically, the flywheel carriage is less mass than the rower, so relative to the floor, the rower moves very little. The rower is no longer flinging his body mass up and down the seat rail. Anyone who tries one of these devices will attest to their superiority over "static" rowing ergometers.

I incorporated the insights of Casper Rekers, together with my gravity-return innovation in the embodiment shown in my original post. Calvin Coffey, a former U.S. rowing Olympian, has been quietly building and selling his sculling replicator for the past thirty years. He has also added dynamically balanced functionality. Users say that it is the closest sensation to being in a boat while still on dry land. Another company is now building something similar, but Coffey was the first to mount the flywheel flat at the rear of the machine, and develop the drive linkage, and to patent this innovation. (Coffey machine shown in the photo)

It seems like you have two options: 1) Win with a logical analysis. So far that doesn't seem to be working. Option 2) win by test & demonstration.

Trying to set up a good test often starts with trying to set-up and test a good model. To do this there are things we need to understand a little better:

1) How does this gravity return work? You obviously don't have an overgrown version of an old window sash weight.

2) To spin-up the flywheel you will work against rotational inertia. Once you are spun-up an ideal flywheel with ideal bearings and no air friction would spin forever. You must have something dragging down the flywheel. Is there a friction band? Is there something like the fins of a squirrel cage blower translating air resistance into rotational drag?

3) How does the rower become engaged with the flywheel during the "pull" part of the cycle and disengaged from the flywheel during the "return" part of the cycle? Are the peak-to-peak distances always the same or is the rower free to adjust his/her stroke length on a cycle-by-cycle basis?

4) Are the machines always level when used?

5) Can anything be adjusted on the machines?

6) You stated that the Performance Monitor (PM) takes all its data from the spinning flywheel. What data is it collecting beside speed?

7) How does someone win? First to cross a PM calculated distance? Most watts burned in x amount of time? Do you have "sprints" and "marathons" or are all indoor races the same?

8) Is there anything someone can do to cheat other than rigging their machine?

9) Could you bracket some of the parameters of a rower? Min/max strokes per minute? Min/max pull distance? Min/max energy per stroke?

10) What is the weight of the flywheel? Does this vary with child vs. adult or male vs. female?

1) "How does this gravity return work?"

On a rowing ergometer, the force to take up the chain and return the handle needs to be strong enough to do this briskly, but you want the main resistance to be from the flywheel, because the inertial resistance of a (vaned) flywheel closely replicates the feel of actual rowing, whereas the linear resistance employed to return the handle does not. Typically, on machines that use an elastic or spring to perform this function the handle return force is between 5-10 pounds. That is the range I aimed for in my gravity return device.

If a "sash weight" strategy were used, with a 10 pound weight, for example, the user would have to pull the weight through a vertical distance equal to the user's stroke length, say about 5 feet. This would be an awkward arrangement. Also, there would be the problem of inertial overrun - whereby the weight continues to move upward for a short distance after the user's stroke is complete, and then drops back, jerking the handle chain. You could use a 50 pound weight and through gearing lift it only one foot, but that is a lot of weight, and there is still the problem of inertial overrun.

My solution was to use the weight of the person on the machine. As seen from the photo in my original post, the upper carriage is comprised of a flywheel, footrests, and seat rails on which the seat can slide fore and aft. The carriage can also slide fore and aft. On the drive portion of the stroke, when the handle is pulled rearward, extending the chain, this spins the flywheel via a chain sprocket on the flywheel shaft. The chain end opposite the handle, via a force multiplier, pulls the described carriage forward up a front and rear slope. On the recovery portion of the stroke, the carriage, by the force of gravity slides (rolls) back down the slope, thereby taking up the chain and returning the handle. Inertial overrun does not occur because some of the force of the pull chain is working against the forward movement of the carriage as it rolls up the slope (since the chain takes a 180 degree turn around the flywheel sprocket).

To keep the handle return force within the 5-10 pound range mentioned earlier, the front and rear slopes are gentle. The sliding carriage moves through a vertical distance of 3 to 4 inches. The gradients are also adjustable. The device in the photo has gradient adjustment of the rear slope only. A more recent embodiment (not shown) has front and rear gradient adjustment. This enables the user to adjust the handle return force to preference. I outweigh my daughter by about 80 pounds, but with either of us on the machine the chain take-up is brisk.

The entire described assembly (sliding carriage, and front and rear slopes) is free to roll fore and aft on the tracked base. This adds a dynamically balanced functionality to the device - borrowing the insights of Casper Rekers' 1995 patent. The base includes levelling means, and the tracks on the base slope upward slightly at both ends. During use, the mass of the user and the mass of the flywheel carriage move together and apart in dynamic equilibrium. The slight upward slopes at the ends of the base tracks keep this linked movement of the two masses centred on the base tracks. Since the mass of the user is greater than the mass of the carriage, the sliding seat of the user moves only a short distance relative to ground. The effect is a floating sensation, very much like the sensation of being on the water in an actual boat.

So that is how it works. I have to be somewhere today so I can't answer your other questions now. This post is long enough anyway.

Very clever.

My daughter has a rowing machine, but I haven't gotten a close look at it. I do know that it has a water tank to do the flywheel function, so you get sound effects as well.

Yes. The WaterRower.

I actually don't use my built unit much. I tell people I have spent more time working on the machine than working out on the machine. It was the mechanical problem that interested me. Reminds me of an ad for an exercise machine: "Designed by a former Navy Seal", the ad proclaimed. Well, undoubtably a Navy Seal would be good at using an exercise machine. Not so sure he would be good at designing one.

I always assumed they used an aluminium disk rotating close to permanent magnets. Just adjust how close the magnets are to adjust the resistance. Maybe it's too difficult to get the accuracy that way.

There are rowing machines that use what you describe - an aluminium disk and magnet strategy. The advantage is that it is quiet. The disadvantage is that the resistance is more linear and less accurately replicates the feel of actual rowing. The vaned flywheel approach was the 1981 Concept 2 innovation, a big improvement over the prior art flywheel-with-resistance-band strategy. The flywheel is then an air pump, and adjustment of a vent controls the volume of air pumped, and thus, the resistance. Noise is their disadvantage. But the combination of inertial and air resistance closely replicates actual on-the-water non-linear resistance. Some recent designs combine magnetic and air resistance. Then there is the liquid flywheel design mentioned by Rixter. Some people will use nothing else. Regardless of the resistance source, most rowing athletes usually work out on a fairly low resistance setting, as this provides the responsive feel of a light racing shell.

Last weekend, an art show was held in a nearby de-sanctified chapel. A friend had organized the show and he let me put my most recent iteration of the gravity-return rowing machine on display as a piece of kinetic art. We set it up on the altar of the former chapel. The chapel is part of a collection of buildings that for over one hundred years was a Sisters of St. Joseph monastery. It is now a community centre. It includes a progressive school for kids 5-8 years old.

I left my rowing machine in place until Monday. Before I moved it out, two teachers brought their students over to see it, and to watch a demonstration. It was great fun! A couple of hours later they returned with artistic renderings of their experience. I include one here. That is me on the machine.

Another technical innovation of rowing ergometers that might interest the electronics people here, is the 1989 patent for a "Self-Calibrating Monitor with Energy Absorbing Means" (US Patent 4875674A), granted to the Dreissigacker brothers, the founders of Concept 2. Together with their 1983 patent for a rowing machine with an air-braked flywheel, these are the two big advances in rowing exercise technology by that company.

The self-calibrating monitor was a big deal. As a consequence of that development, competitors in indoor rowing competitions were allowed to set the drag of the flywheel to any resistance they wished. That is, they could now set the vent opening on the flywheel housing so that the vaned flywheel pumped a greater or lesser volume of air, thereby choosing the resistance level they preferred, and the monitor would automatically compensate for the drag setting and maintain a level playing field. For example, if a rower maintained a constant average watts output, the drag setting chosen wouldn't matter, the time for the 2K distance would remain the same. No advantage could be obtained by choosing a lower resistance setting.

But is this automatic compensation of drag 100% accurate? Does it ensure that the playing field is absolutely level? The patent document describes the means of the self-calibration. It makes no claims as to the accuracy of that self-calibration. For three decades users of these machines have trusted the manufacturer that there is parity between machines, that the monitor self-calibration function is 100% accurate. To my knowledge the accuracy has never been independently tested. Now finally (why has it taken more than 30 years?), with the development of the test rig in Ulm, Germany, it could be tested. I write 'could be tested' because at the rate the Ulm team is moving it could take another thirty years before they get around to doing the necessary tests. So, as per my original post, if I could build a test rig of sufficient accuracy, the following is my conception of how the test of the monitor self-calibrating function could be done (this is a different test than the one described in my original post, which was to do with variations in elastic cord strength). Critical comments are welcome.

1) Via a motor driven rig, deliver a constant stroke over stroke average power to the rowing machine.

2) Set the drag setting of the flywheel to the least resistance.

3) At that drag setting record how long it takes to go a certain distance, as displayed on the rowing machine monitor (typical distances are 500M and 2K).

4) Increase the drag setting by an increment and repeat 3).

5) Repeat 4) until the entire range of drag settings has been covered.

6) Compare the time at each drag setting for the given distance.

7) Repeat 1 - 6 with different power inputs to the rowing machine.

I know that the accuracy of the test rig must be taken into account, but that aside, if the self-calibrating function of the monitor is 100% accurate, all of the displayed times for a given distance, at a constant average power input, should be the same. If they are not the same, the times displayed will indicate the magnitude of inaccuracy. Have I described a viable test?