• If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • You already know Dokkio is an AI-powered assistant to organize & manage your digital files & messages. Very soon, Dokkio will support Outlook as well as One Drive. Check it out today!


Richard Johnston Dumbbell

Page history last edited by R H Johnston 4 months, 3 weeks ago

Richard Johnston's "Tyes" mini-Dumbbell

at Plymtree Devon


This dumbbell has developed over time and has been adapted in stages to be suitable for distanced ringing, so if you are looking for a pattern read the whole thing and understand how to make the best version of this - or indeed a better - dumbbell.  There are more thoughts on dumbbell design here


The Dumbbell in context in the down position as it was in early June 2020


Having used John Horrocks' Saxilby dumbbell (his brother makes them) quite a lot when he lived in the neighbouring village of Payhembury, I would have very much liked to have had one of those - they work very well, and I rang a number of quarter peals on that, did a lot of experimenting  with "pullometer" possibilities on it.  I have also rung two peals of surprise royal (Link to the first one, which was the first simulator peal recognised by the CCCBR.) on the Denmisch ring of 10 Saxilby simulators at Exeter Cathedral school.


However I live in a bungalow with a shallow sloping roof, and the maximum headroom at the ridge is just 3 foot high.  The garage has a flat roof, so there is not enough height for a conventional dumbbell. The only realistic option for anything was the light well for the Velux window that lights the main entrance hall.  That meant the installation had to be very small - the width of the lightwell is only 550mm wide (about 21.5 inches).  So the options were limited.  Also the only place the unit could be fixed in place was at the bottom,. so the unit also needed to be light, structurally self contained, and it had to be possible for one person to erect it as there is no room for ore than one person to work.


This dumbbell was mainly made from the contents of the scrap bin, and designed evolved around what was available.  Originally I attached the wheel to a beam in the garage (see the Annex below) - and this worked fine, but the garage is not hospitable, too full of vehicles and the height was not enough.  But it proved some of the principles that are incorporated in the final result.



The wheel is a "16x 1 3/8" front wheel from the original Moulton bicycle - the rim was damaged but it works fine for this job..  The wheel is about 14.5 inches in diameter.  Very small.  The main "bell weight" is a piece of square steel tube, of the kind desk legs are often made of, but thicker.  This is attached with cable ties. 


Sensing points

Until 2020  the sensing was done by an ActionXL controller, of the sort widely used as dummy handbells, so I did not need any.


Now, the attachment to the wheel at the bottom of the picture is the main reflecting surface for the light based sensor, which is the black box tied in place with a shoelace. Two further optional sensor points are at 11 o'clock and half past two - these are detachable and held in place by Blutack.


Stay and slider

At the top, on the wheel is a bracket which carries a bolt which has rubber tube around the thread on the far side.  This is the stay.  It engages with a nylon cord which runs vertically down through a hole in a block of wood down to the weight at the bottom.  Note the knot just above the weight.  That is the slider stop.  The system is adjustable as the cord goes round the axle spindle and down the right hand side of the block of wood, underneath up the other side, across the top and down the right hand side again.  The second pull down on the right is on the front side, and when pulled tight this locks securely into place. The weight has to be chosen carefully, too light and the stay doesn't work well, and too heavy and it drags the bell back out.


This provides a very simple and very effective stay and slider system suitable for any mini-bell installation.



Here the bell is set at handstroke .- and backstroke





Rope system

The main rope is the 8mm hempex one on the right.  Because the bell is light formal pulleys were dispensed with and 12mm bolts used instead.  Two nails serve to keep the rope aligned fore and aft.  The garter hole is the hole for the tyre valve (the wheel has to be orientated to get that in the right place ). 


The rope is one of the heavier elements of the design, so a balance rope is fitted to the system - this is a white soft cotton 6mm cord, which shares the valve hole with the main rope but turns the other way.  This cord also has its own guides.  It simply hangs down and is long enough to reach the floor at all times, and some, so it stays in place and under control.  A small H4 knot in the rope stops is going through the eye, and adds to the rope weight a bit. 


There is no conventional sally - instead an H4 knot is made to provide a sally.  The H4 is a style of rope shortening knot used by method ringers in Devon when they ring at towers rung by call change ringers (who often like their ropes very long): The rope from the bell is turned back up a short way, and the rope wound round back towards the loop end that was made and the end of the rope put through the loop and made tight.  Done carefully it is very neat, and is kinder to ropes than figure of eight knots.  The idea of using this approach is to keep the weight down, and by adjusting the number of turns the rope weight is adjustable.  The tail end is spliced back to make, in effect, a Devon tail (a real one a small loop with the rope spliced back in).




H4 knot sally




The sensor system


The sensor point detection system consists of a low power laser (with light level controlled by a 100 ohm variable resistance in series) shining onto reflective tape detected by a phototransistor. The circuit is trivial.  The USB connection is provided by a 3.3volt FTDI TTL-232R, TTL to USB Serial Converter Cable -there are different cable connectors available but any will do as only the bare wires are needed.  This unit has 6 wires and formally provides the power needed by the laser.  Ideal, as the electronics is built into the USB plug..




The frame is just a box built from a 12mm MDF offcut I had in stock.  A hardboard cover hangs across the front to keep light away from the sensor, and to make it less obtrusive.




It is like ringing a good mini-ring bell.  But with a stay.


The installation as illustrated rings very smoothly at between 2h30 and 3h10m speed, and works very well with Abel and Beltower.


Distanced ringing is often much slower, and I made some alterations to make it possible to ring slower easily - I added extra weight mainly to the side away from the bell to make it turn more slowly, and also a little weight to increase the moment of inertia generally. I had to make some changes to the roping and to the weight on the slider system.  This has had the result of improving the control generally, with the option of a very wide range of ringing speeds.


Here it is set at backstroke:



The stay system was now starting to pull the bell out sometimes making setting difficult.  Nevertheless the longer pull and effectively larger wheel was an advantage. I did some mathematical calculations that showed that if the bell were set deeper it would better stay set, in circumstances where the bell bounced on the stay.    So a further change was made to the stay system so the bell set more deeply (27 June 2020):




but the friction in the stay system became more apparent. 


While attempting a further adjustment the stay cord fell out, and revealed that it was not really necessary, as the bell adopted a natural set position that was short of the bell being at BDC.  The main rope "sally" acted as stop at backstroke, while a washer and knot in the balance cord at handstroke prevent BDC being reached and a false "bell strike" occurring.  The resulting performance is excellent and the bell can be rung at between 1h 45m and 3h 30m peal speeds.  The fast speeds are achieved with more energy and the bell lower, while the slow speeds are like ringing the trebles on 12.  Ringing it is now enjoyable, without any fear of the bell dropping.  With this setup, it is mainly a moment of inertia system, rather than mainly driven by the weight of the bell, as would normally be the case. 


It gives a long pull at both strokes, and effectively the wheel is no longer 14.5 inches across but about 24 inches, at no extra cost!


What is interesting however, is that the sense of pulling off the stay is stronger, because the early pulling needs more  effort (because the torque from the centre of gravity offset is a maximum when the "bell" is horizontal, and then progressively falls away as the bell approaches the vertical before then  falling in the usual way.  The "feel" of the bell is therefore improved compared with any of the earlier arrangements. 


The important thing about a dumbbell is the rope feel.  How that is achieved is not important.  The dumbbell is as hidden from sight and attention as any real bell in a tower.  


Here is the current (July 2020) configuration, set at hand and back strokes.



Handstroke - showing the balance rope washer and knot, not engaged

with the stop


Backstroke - the main rope visible as the rope is over itself and the balance

rope, but there is no tendency to slip wheel.



August 2020: This not the end of the story as I have now added an Arduino operated 6-axis gyro unit to stream continuous bell position data - allowing the possibility of displaying an accurate rope movement image - of which more in due course:



Update Jan 2024:


The arduino sensor

I never got around to completing the arduino system: the difficult was with the wiring needed to get the data from the gyro unit.  and there is not much space between the wheel and the front support bar.  It would hbe much easier if the wheel could be supported solely from the other side, but that isn't realistic with a bicycle wheel.  I'd juist about got everything to work reasonably well, and then the cable connector to the arduino came away from the unit.  This wrote off the arduino, as although I could get power to the unit by alternative means, the critical serial data connections are too tiny to resolder and there is no alternative pins on the unit.  So I took it off.  I don't need it for Ringing Room or Abel or Beltower.  To bhe useful for distanced ringing, which was the original intention, we would need a new ringing platform, and all users would need a unit like the one I made.  Shall we say this is a "hoped for sometime" project.  The arrangement used the same eleectronics and the "ebells" now available toi use with Handbell Stadium.  All we need is someone tp develop the towerbell equivalent. :-) 


The central disk and leather wear protector are currently redundant, joining a number of other features from the dumbell's earlier development..


Performance upgrade

During 2023 I became not entirely happy with the dumbbell and I felt it needed more moment of inertia in particular so I added extra weight around the wheel to accomplish this, and also added a little more off-balance weight.  This has made it much easier to use, and strike well.


Below are photos of the revised arrangements, set at handstroke,



and backstroke:


And the name?  I live in Tyes Orchard, Tyes being one of the ancient landholdings in the village, names after the Tyes family which once owned it.


Annex Early Development of the "Tyes" dumbbell


18 January 2019 Garage prototype


The prototype never had a sensor, but mechanically worked well despite the old chair rests that served as a frame and rope guides!  The axle was bolted through the steel beam that serves as support for the engine hoist



Some of the detail of the rope guidance:


18 January 2019 The dumbbell as first installed.


As may be seen, it then used an ActionXL controller to determine the position, and a modified version of Handbell manager to collect the data.


These are the outputs of the sensors for the 3 axes at Handstroke pull...


and backstroke pull



This system worked well once I had sorted out a suitable algorithm.  The main drawback I foresaw long term was that the cable would fail from the repeated bending.  Reluctantly I eventuallt changed it for the current light based system.


Development during 2019 and 2020 led to a dumbbell that I was happy with with Beltower, which I find better for me than Abel.


By early June 2020 the simulator looked like this:



As mentioned above distancing resulted in significant changes in mid-June 2020:


Subsequent developments are described in the main article above.


Richard Johnston 17.6.2020, 2.7.2020, *.8.2020, 21.10.2020, 5.1.2024


Comments (0)

You don't have permission to comment on this page.