Metalmagpie:
Idlers to tension flat belting were used in a number of applications. A very common one was a weighted idler to tension the flat belt driving steam engine governors. This idler had a dual purpose. Aside from tensioning the governor drive belt, the 'hub' of the idler had a cam called a 'knock off cam'. If the governor belt broke, the engine could run away (dangerous, if not destructive overspeeding). The knockoff cam rotated with the idler arm and worked the throttle valve in the governor to shut off steam.
Ingersoll Rand described what they called a 'short center' flat belt drive for their bigger recip compressors. This allowed the motor to be mounted close to the compressor, and the idler was used to increase angle of wrap of the belt on the motor (relatively small pulley) and on the compressor flywheel.
My first exposure to idlers on flat belt drives was at Brooklyn Technical HS in 1964. The first shop course we took was wood patternmaking. The shop class rooms dated to the 1920's. In our pattern shop classroom was a Tannewitz band saw with flat belt drive. There was a weighted idler pulley to maintain belt tension, as well as get more angle of wrap since the motor sat at floor level, quite close to the bandsaw. When the motor was started, the bandsaw wound up to speed, and that idler hopped up and down on the belting until load was put on the saw.
At Hanford Mills Museum, there are numerous flat belts and line shafts as well as woodworking machinery driven off the line shafts. For the most part, the belting is tensioned by taking pieces out as needed and relacing (installing new hooks). There is one really wide belt that transmits most of the power in the mill. This belt is on fairly close centers and was always a problem to maintain tension. One of the mechanics who worked at the Mill made an idler pulley, and used a scissors jack from a car to adjust the idler and tension the belt. Backwoods fix, but it does work.
In shops and mills with a lot of flat belting coming off line shafts and counter shafts, maintaining belt tension was always an issue. In the shops and mills with a lot of belting, a cart with the belt cutter (shear) and lacing machine and assortment of lacing hooks and pin material was often used. The cart was knocked together out of wood, and was rolled to wherever a belt had to be relaced.
Enginebill is correct a bout running the belt off by hand to relieve the tension. When I was a student at Brooklyn Tech, we still had many lineshaft driven machine tools. Similarly, as I began working part time and summers in a machine shop, there were still some line shaft driven portions of the shop. At 'Tech, the belts were taken down over holidays and vacation time. This was to prevent the belts taking a permanent stretch. In the machine shop, over a summer weekend with humidity, the fear was the belts would permanently stretch, or at least be too slack to transmit power without slippage come Monday mornings. We used to run the belts off the pulleys each Friday when work was done. The belts were hung adjacent to the line shaft pulleys on heavy wire hooks. This was to keep the belts from soaking up oil that might have been in the grunge on the line shafting. Coffee cans or similar hung on wire under some of the lineshaft bearings which did not have the cast iron troughs below them. These cans of dripped oil had to be handed down and emptied into a bucket which went into a used oil drum. Come Monday morning, we had the job of putting the belts back up on the pulleys. Several of us worked at it. With no formal 'lock out/tag out'. the foreman would place himself bodily in front of the motor starters for the line shaft drive motors. One guy would be heaving on the spokes of the big lineshaft pulley nearest the drive motor. A couple more of us would climb like a squirrels and work the belts onto the pulleys as the lineshaft was heaved around. We came in a half hour early for this detail, and oiling the lineshafting was also a part of it.
With all of this, we still would start up the shop and as someone hogged into a cut, a slipping belt was occasionally encountered. Out came the lacing cart. The cart was ancient (or so it appeared to me as a kid). It was battered and covered with belt cement from cemented splices, and had a lever operated shear for cutting and squaring the ends of the belting, as well as the lacing machine. There were also some clamps and a drawknife for 'skiving' belting to be splced by gluing. I never saw a glued belt splice made up. Lacing hooks were about all I ever saw in use. The slipping belt was shortened and relaced, and put back up without stopping the line shafting. It was run onto the pulleys. At Brooklyn Tech HS, our shop teachers showed us how to shift a belt on step cone pulleys on running machine tools. They used the sides of their hands, and had the belts on cone drive lathes seeming walking from step to step of the pulleys with no problem. We kids were warned not to use our hands, but to use the smooth side of a wrench, or a stick. An 'engineer pattern' wrench with an oval cross section to its handle, laid with the flat side against the edge of the belt made an ideal belt shifter.
It's a long time ago that all of this was happening. I was fortunate to get a belt cutter (shear) in a lot of junk my Dad bought at an auction in 1958. I got a belt lacing machine and a mess of lacing hooks on paper carding when we cleaned out an old hydroelectric plant. In 1921 when the plant was built, the governors were flat belt driven from the turbine shafts. In 1959, the plant was modernized with Woodward ballhead hydraulic governors using a synchronous motor drive and permanent magnet generator. This eliminated the need for flat belting. The flat belting stuff was stored on a shelf and forgotten about, along with glass globes for vapor proof lighting that had been painted for WWII 'blackout' lighting. I grabbed the belting stuff when the contents of that store-room went into the dumpster. I had to explain what 'blackout lighting' was to people who wondered why there were so many light fixture globes painted black with only a small slit left unpainted. There was a lot of old stuff in that store-room which I had to explain the usage of and determine if it were to be kept or junked. I got a lifetime supply of belt lacing hooks, aside from the lacing machine, so made out handily. I also got a Pexto drawknife, which I suspect may have been used to skive the belts for glued splices. Never found any clamps for glued splices, so am guessing about the purpose of the drawknife.