Kensington Orbit Fusion 72362

Archie
Site Admin
Posts: 144
Joined: 31 Jan 2019 15:27

Kensington Orbit Fusion 72362

Post by Archie » 08 Apr 2021 08:35

https://www.kensington.com/p/products/e ... trackball/

Ball size: 1 37/64" (40mm)
Interface: USB wireless
USB ID: VID_047D&PID_807B
Product Name: Orbit Fusion Wireless Trackball
M/N: M01490-M
P/N: K72362/K72363
FCC ID: GV3M01490-M
Year: 2020

Image


Currently, most recent Kensington trackball. Finally manufacturer decided to include feature-rich model in Orbit product line, traditionally limited to just two buttons for decades. Also it's a first non-symmetrical Orbit. Looks almost like a copy of Elecom Deft Pro, except for the scrolling control implementation and lack of USB cable and Bluetooth connection versatility. Nice carrying case is available separately. Oddly enough, dual P/N numbering just reflects package type: "Retail Packaging (K72362WW), B2B Packaging (K72363WW)".


Image Image

Image Image

Image Image

Image Image

Image Image


The sensor is a PixArt PMW3610DM, all the switches are T-MEC (cheap Omron D2L-D copy). Controller is labeled 24AT01 DS0342F (can't find datasheet at the moment, but the same chip is used in Elecom products and I've seen information on Japanese sites that it should be similar to Nordic Semiconductor nRF24LE1E). Power is provided by iDesyn iD8603 voltage converter. Resolution range is specified in documentation as 400-1500 DPI.

Image Image


The ball is supported by synthetic ruby beads of approximately ~2.5mm size. Unlike other Kensington models they're not easily replaceable, being permanently pressed into cavities - so, drilling from the outer side is required in order to extract them, like in many old trackballs from Logitech, Microsoft, etc.

USB dongle labeled M01480-D is the same as in Pro Fit #75326 thumb trackball.

Image Image


This model is not supported by former TrackballWorks software, but installation of extremely buggy KensingtonWorks is not necessary: all five buttons generate standard HID codes, so any third-party mouse utility (e.g. XMBC) could be used to assign desired functions.

Product information PDF:
https://trackballs.eu/media/Kensington/ ... 363WW).pdf

Design story:
Spoiler
Show
Behind the Scenes: How Kensington™ Created the Orbit® Fusion™ Wireless Trackball
https://www.kensington.com/news/ergonom ... trackball/

Interview with designer Erik Campbell about 2.5-years development story. The article is full of incorrect statements, but worth reading anyway. Among other things, there's an evidence that initially even Elecom's angled right button and additional Fn1 - Fn3 buttons were planned to be implemented in Fusion:

Image
Installation Guide
https://trackballs.eu/media/Kensington/ ... K72363.pdf

Undocumented pairing procedure:
https://old.reddit.com/r/Trackballs/com ... s_erratic/
Spoiler
Show
Few weeks after I bought it, got similar behavior, then it suddenly stopped working. After contacting support I was told to re-pairing the device and the usb dongle by following this undocumented process:

1. disconnect usb dongle and turn off device
2. plug in usb dongle with the device turned off
3. keep pressed the middle button (the gray one) and turn on the device
4. the led will blink 6 times
5. you should be good to go

FYI I had to repeat this process a couple of times in the next few days, but since then it has been working without issues. My best guess it that there was some issue with the wireless signal between the dongle and the receiver.
Original package box and leaflet:
Spoiler
Show
Image Image

Image Image

Image

Image Image

Sad to say, actual impression is tremendously disappointing. If Elecom Deft Pro or Logitech Cordless TrackMan looks and feels like premium product, the Fusion is cheap and poorly made. The scrollring is just plain inferior: wobbling and catching randomly when operated, with highly inconsistent rotation resistance and scratching feel. It was old problem in Expert models, but this one is much worse. The issue is very complex, and caused by more than single defect. In general, it's an incorrect dimensions of matching parts: the ring is rubbing almost everything. Both inside and outside diameters of cylindrical slotted part are too large - so it's unable to stay centered, rubbing against curved walls of ball cup.

Image Image

Image Image


There are 4 low-friction plastic squares glued to the cup, obviously intended to serve as a support and centering aid - but they're too thin, so there's a gap that allows the ring to be shifted and skewed. Upper edge of cage is flat, but corresponding inner surface of ring is conical - so it's physically impossible for it to keep the same position.

Image Image


As a combined result, rubberized bezel touches upper case cavity, while slotted part touches everything inside: the cage, mounting post, and even switch assembly's PCB edges. Extremely difficult to fix, unfortunately.

Primary button makes very annoying tactile "double-click" which is quite distracting. Electrically it's normal single click, but feels and sounds like doubled - making an impression you've accidentally double-clicked some item by mistake. Turned out Kensington actually used two separate microswitches for left button, connected in parallel - so, depending on particular zone of button and click force applied, either the first, second or (most frequently) both switches are engaged within a fraction of second interval. Unbelievable! Popular fix for this issue is a cutting out the actuating pin for one of them. In my case, I've simply desoldered "extra" switch.

The right button also produce highly unpleasant feel, like if it's scratching something on press and on release. The problem caused by unfinished edges with protruding pieces of plastic left after molding: they act as a kind of "latch" preventing free travel of button.

DPI indicator / battery discharge LED is very dim, barely visible even in dark room: in normally illuminated office it's totally useless. And finally, rubber anti-skid strips are slightly longer than case cavities for them, so the ends of strips are not seated fully - resulting "bumps" cause slight rocking on the desk.

It's impossible to remove the ball without some tool, e.g. pen, as the hole is ridiculously small (about 10mm) and deep, implemented as a long tunnel. Ironically, in the product sheet this pinhole is labeled "Trackball Access (easily remove trackball for cleaning)".

Scrolling seems to be flawed a bit: some of my programs don't respond to Fusion scroll messages. Maybe it's just rare coincidence or compatibility problem between particular OS or software versions, but scrolling issues were reported in many reviews.

Detailed steps to repair and make it usable, in order of importance:
Spoiler
Show
Completely disassemble the device. 7 screws are located under adhesive rubber strips. Tri-point (a.k.a. Y-type) screwdriver is needed. The tricky part is a scrollwheel: you'll need to remove sensor PCB and block of buttons first, then carefully disengage 4 plastic tabs and push the inner retaining ring out. After that, ball cup (held by 2 screws) could be separated from the case.

1. Phantom double-click.
Either cut the pin on inner side of button, or remove the second microswitch (I'd prefer the latter method). Simple and easy.

Image Image


2. Jamming scrollring.
Most difficult job. Trim and file all the sharp edges involved in contact surfaces. It mostly relates to the ball cup, but needed for the ring and its retainer as well. Special care is needed for outer edge of slotted part: after processing, carefully inspect all the slots and remove debris - otherwise it will block the infrared beam to the rotation sensor and interfere with normal scroll operation.

Image Image


I've removed 4 teflon-like parts, drilled 2mm blind holes slightly more than 1mm deep, and glued 4 ceramic 2mm balls into them. Depth of holes should be carefully adjusted (several iterations of checking needed here) to provide free motion of scroll wheel with as small gap as possible.

Image Image

Image Image


Curved edge of PCB with switches needs to be slightly rounded to provide a clearance: pay attention to not damage copper trace.

Image Image


On the surface of ring under rubberized crown, molded digit "1" slightly protrude, touching textured surface of upper cover when rotated. It needs to be filed off.

Image


Before final assembling, I've thoroughly lubricated all the contacting surfaces with good amount of PTFE consistent grease.

3. Scratchy right button.
Using sharp knife, cut plastic residues everywhere along the slit separating the button and the rest of case, especially near the end of button.

Image Image


4. Uneven base.
Cut the edges of rubber strips to make them fit the cavities properly.

Image


5. To make the LED visible, additional transistor is needed - as it's driven directly by microcontroller, and common method of increasing the current by changing resistor value is not safe here. I've decided not to do that currently.

6. For ball extraction, some mechanism should be added, e.g. push button like on Pro Fit #75326 model. Will do that later.

7. Scrolling issues. Looks like generated scroll messages are not completely standard, or something like that. In my practice, it could be fixed on per-program basis with XMBC utility. For example, to restore scrolling in FileZilla FTP client "Method 1 (SCROLL Msg)" works, while for FastStone Photo Resizer utility "Method 5 (WPF Scroll Msg)" should be used.

8. To eliminate free play, I've added small pieces of electric tape to the ends of button pins actuating the switches. Fairly minor thing (not specific to this particular trackball), but it significantly improves usage feel.

At the end of this story, the repaired unit with major issues fixed works just fine. Can't say the scroll ring became absolutely perfect, but it's good enough: no more sticking, scratching and seizing; rotation is very smooth, and even have some kind of inertia. Button clicks are soft and nice, with no any tactile problems. In practical use, device is really convenient. The shape is excellent, palm position and location of controls is great: easily accessible and handy. Left-side group of buttons is well made: each button have some minor perceptual difference (shape, height, protruding dash etc) for easy tactile recognition. Ball movement is butter-smooth. Of course it was used to prepare this article.


Overall, in my opinion this model is total failure, and requires heavy repair/modding right out of the box. Giving third of a century Kensington reputation as a vendor of great trackballs, such junk-grade product was truly unexpected. Definitely it was a great idea behind Fusion, and really hard work was done by designers and ergonomists: it has immense potential to be an excellent device - but lack of proper engineering, manufacturing and testing can easily bring any good plans into discredit. In terms of quality, the best part of this trackball is probably its carrying case. For those who like such ergonomic shape, and don't have necessary equipment/skill/patience to make it usable, I'd recommend to get Elecom Deft Pro instead: the difference in quality is day and night. After all, it fits Fusion's storage case just fine.

Image