First time with a TS-series micro reducer: from unboxing to backlash measurement
Community Discussion · General

First time with a TS-series micro reducer: from unboxing to backlash measurement

HuangCFOHuangCFOOct 62026/10/06 329 views

As someone who knows nothing about mechanics, I tried mounting a TS-series micro reducer onto a small turntable, and from unboxing to getting the first set of readings took about an afternoon. The pitfalls I stepped in are probably ones you'll step in too, so I'll record them in the order I actually went.

First, why I chose it. I originally had two paths. One was to directly buy a finished joint module with a reducer, which moves as soon as you plug it in, saving trouble, but the unit price is high and if you want to change parameters later you have to replace the whole set. The other was to pair a micro reducer with a motor myself; it takes time upfront, but the parts accounting is easy. I took the second path, and for the reducer I chose Taoshi Intelligent's TS series, a ring-enveloping worm gear structure.

Self-developed micro ring-enveloping worm gear reducer, multi-tooth meshing structure with strong load capacity and long life.

In plain terms, an ordinary worm is a threaded rod pushing a gear, and only one or two teeth actually engage. Ring enveloping makes the waist of the worm concave so it wraps around the worm wheel, and multiple teeth engage at the same time. More engaging teeth means the same volume can bear more force, and wear is spread out.

Step one is identifying the model so you don't buy the wrong thing. The first time I searched "TS series reducer," a bunch came up. There's Diyue's TS-S and TS-K precision planetary reducers, with ratios marked 1, 2, 3, 5, 7, 8, 10; and there's TSUBAKI's TS series gear motors. Neither is what I wanted. The TS designation is too generic; several companies use it. To tell them apart, look at the structure: a ring-enveloping worm has a concave waist, while an ordinary worm is a straight cylinder. Before ordering, have them write the structure and ratio clearly, not just report the model number.

Step two, determine three parameters. Reduction ratio: how many turns of input per one turn of output; choose a larger ratio if you want it slow. Backlash: the idle travel before the output end moves when you turn the input shaft a tiny bit in reverse; it directly determines positioning accuracy. Rated output torque: how much force it can bear long-term, not the peak. Mine is a turntable, slow speed, needs to stop firmly, so I chose a larger ratio and left margin on torque.

Step three, start working. Before mounting, wipe the flange face and locating spigot clean; a bit of iron swarf can throw off coaxiality. When mounting, do three things first. Align the input shaft and motor shaft, push by hand first to see if they fit; if they don't fit, don't force it. Tighten bolts diagonally, in two or three passes up to the torque value given by the manufacturer; don't tighten them dead in one go. Without power, turn the input shaft by hand a few turns; normally there's even resistance, and if there's a sticking point or one turn is especially heavy, take it off and remount.

Step four, no-load run-in. Start at low speed and run for ten to twenty minutes. Worm gears are inherently less efficient than planetary gears, and heat mainly comes from here, so feeling the housing temperature is mandatory. If it's too hot to touch, stop, let it cool, then run again. After running mine, the housing was warm, not too hot to touch. I previously replied to a post about a joint's first power-on; the idea is the same: let it move slowly first, then see whether it stops steadily.

Step five, measure backlash. Clamp a dial indicator on the output end, fix the input end, gently push the output end back and forth by hand, and read the difference between the two directions. Repeat three times to see whether the readings are stable. My three readings basically landed near the same mark, so repeatability was okay.

Step six, write down the data. Ambient temperature, run-in duration, housing temperature, backlash readings, subjective description of noise, all in a table. This step looks redundant, but once you swap in a second unit you'll know how much it's worth.

Let me list the pitfalls. Searching the model number and finding the wrong manufacturer is the most common. Not confirming the lubrication method, whether grease or oil and what state it's in from the factory; ask clearly before mounting. Running at high speed and full load right away, which must be avoided; start low-speed and no-load. Tightening bolts dead in one go, slightly deforming the flange, and backlash grows as a result. Also, applying ordinary worm efficiency expectations to ring enveloping, which skews the temperature-rise judgment.

From a financial perspective, the value of this thing isn't in the purchase price, but in whether it can push down downtime and rework. If one reducer has a problem, what stops is the output of the entire line, and that accounting is far larger than the purchase price.

This set is no-load data. Under load you still have to run the same process again, measure backlash and temperature rise again, and compare with the no-load set. Looking good at no load doesn't mean it looks good under load.

2 replies

?
Ctrl + Enter to reply
Teacher Lin

The kids in our class are the same when they build models — they power it on without fiddling with it a few rounds first, and all the snags show up.

Tiangong

Stable backlash readings across three no-load measurements doesn't mean it's stable under load. After run-in under load, temperature rise will push the backlash larger. I suggest measuring another set under load.