This review has also been posted on Deskthority (DT):
http://deskthority.net/keyboards-f2/kul-es-87-with-cherry-mx-clears-comparisons-with-topre-and-ibm-t8833.html#p185719Recently, I bought a new Kul ES-87 with Cherry mx clear switches. This board has been available for some time, and a number of reviews are already available. Therefore, this post is not intended to be a thorough review with keyboard pictures, but rather a collection of my impressions about the keyboard and Cherry mx clear switches, including some comparisons with other Cherry mx switches as well as Topre and IBM buckling springs (complete with an inspection of force-displacement curves).
Thus far, the Kul has been available only with mx linear switches (red or black) or tactile switches (brown or clear), but not tactile/clicky switches (blue or green). I had previously tried other boards with black, brown, blue, and green switches, and I had not really liked any of them. Accordingly, I ordered the Kul with clears to try out this new board with a switch that was also new to me.
Kul keyboards are a available through several sources, including Elitekeyboards.com (EK), Mechanicalkeyboards.com (MK), and Amazon. The price is reasonable at 129 USD, and MK and Amazon (Amazon Prime) offer free shipping.
The keyboard is attractively packaged in a sturdy box that includes the keyboard, detachable mini-USB cable, keycap puller, extra keycaps for swapping CapsLock/Control, Escape/Backtick, and Backspace/Backslash, a plastic keyboard cover molded to fit over the keycaps, and a manual written in English.
The keyboard itself is a standard TKL design housed in a sturdy matte black case with no flex or creaking even when pressing down hard anywhere on the top panel. It is equipped with non-skid rubber pads on the bottom front and rear of the case, rubberized feet that lock into position with a reassuring solid snap, a mini-USB connector in a recessed area underneath the keyboard, cable management channels for directing the cable to the left, right, or center, and 8 DIP switches. It has a CapsLock feature with a CapsLock LED, and an LED under the Escape key to indicate Windows Power Mode, but it lacks an embedded numeric keypad and so it has no NumLock LED. My keyboard was equipped with Cherry mx clear switches, and Kul keyboards use Cherry stabilizers. A nice touch that I didn't notice right away is that the bottom of the case has a beveled front edge with its own set of rubberized grip pads, presumably to provide better traction when the feet are extended -- the only other keyboard I know about that has this feature is the RF 87u series.
Although many people like detachable cable connectors in a recessed area with cable management channels, I find it difficult to connect and disconnect the cable from such an arrangement. I prefer to have a connector on the back of the case, as it is often found on sub-TKL keyboards. In fact, I like to have the connector on the left rear of the keyboard (near the Escape key) or the center rear, and rather than being disappointed if the manufacturer decided to put the connector on the right rear of the case, I suppose it is good to have the cable routing channels.
One reason I bought this keyboard is that the abundance of DIP switches (8 of them) provides some handy hardware customizations that I would otherwise have to remap using software. These include the following: 1. Windows/OS X mode; 2. Left Control/CapsLock; 3. Escape/Backtick; 4. Backspace/Backslash; 5. Shift+Fn+Menu/Fn=Menu; Shift+Fn=Fn; 6. 125Hz polling/1000Hz polling; 7. 6-key rollover/N-key rollover; 8. Future expansion. Extra keycaps are provided to correspond with various key swaps.
The DIP switches get high marks from me. I used the DIP switch settings to give me Mac OS X mode, Control to the left of the "A" key, HHKB-like Backspace/Backslash swap, and a remappable Fn key -- with this arrangement, the Fn key sends a scan code to the computer and I do not have any dead keys on the keyboard, as can be the case with keyboards that choose to have a hardwired Fn key. Excellent!
The stock keycaps on the Kul are thin ABS with laser-engraved infilled legends. Because most people would want to replace these, at least eventually, the advantage of the included keycaps for the various possible key swaps would then be lost. However, the aesthetic (and even functional) benefit of dressing up the board with some handsome doubleshots or dye-sub PBT caps would tend to outweigh the specificity of keycap legends (although this is especially nice to have for those who swap Backspace and Backslash).
Everything was fine with the keyboard itself. However, when I started typing on it, my heart sank, as I discovered that I hated typing on Cherry mx clear switches. This surprised me, as I had read glowing accounts from others for whom this switch was a favorite. The tactile bump that has been added to this switch felt scratchy or gritty in the extreme, and rather than representing the collapse of a dome or spring followed by a decrease in force, as in a Topre or IBM buckling spring switch, there was a markedly rising force all the way to bottoming out. Because I normally bottom out when typing on any other type of switch, the force curve of the mx clear switch made me feel as if I was fighting the switch rather than having it do my bidding. Moreover, I found that some of the keys would actually stick at around the actuation point, requiring a slight nudge to get them to spring back to the starting position.
My discomfort with the Cherry mx clear made sense when I actually looked at its force-displacement curve and compared it with the force-displacement curves of two of my favorite switch types, the Topre and the IBM buckling spring.
Cherry mx ClearTopreIBM Buckling SpringWith the mx clear switch, the peak force of the tactile bump is 65 cN. The force decreases after the bump to a local minimum of about 50 cN, and the actuation point (operating point) occurs at around 55 cN while the force is increasing with a steep slope. The force then continues to rise steadily to about 90 cN before reaching the bottom of the stroke. It is easy to see why this switch gives a feeling of fighting against it after the actuation point if you try to bottom out the stroke. In fact, the force-displacement curves of all Cherry mx switches share the characteristic of increasing force immediately after the actuation point, regardless of whether the switch is linear, tactile, or tactile/clicky.
In contrast, the force-displacement curve for a 45g Topre switch rises to a tactile bump at 45 cN, but actuation (operationg point) occurs while the force is decreasing. The force continues to decrease, enabling the user to allow the key easily to finish the stroke until reaching bottom.
In like manner to the Topre switch, the force-displacement curve for an IBM buckling spring switch, shown in the bottom panel, increases steadily to a maximum of about 70 cN, whereupon the spring undergoes a catastrophic collapse (it buckles) accompanied by a precipitous drop in force from around 70 cN to about 55 cN, corresponding to actuation ("make") of the switch. After the sudden collapse of the spring, the force then increases again to the bottom of the stroke.
The force-displacement curves for all three switch types exhibit hysteresis -- the return or recovery (upward) stroke does not retrace the downward stroke; instead, the upward stroke follows a different path as the switch resets itself to permit a new activation and recovery cycle.
This analysis of force-displacement curves made it clear to me why I did not like any type of Cherry mx switch and why I liked both Topre and IBM buckling spring switches, which might seem worlds apart in their mechanisms until the force-displacement curves are compared. In general, all Cherry mx switches display increasing force immediately after the actuation point, whereas Topre and IBM buckling spring switches have decreasing force directly after the actuation point.
All was not lost, however. I decided to try replacing the stock keycaps on the Kul with Imsto dye-sub thick PBT and 40A-L O-rings. This resulted in a shortened key travel, thus reducing the interval of the increasing force, bottoming out before the force could go as high as it would with a longer stroke. This gave the feeling of actuation almost coinciding with the bottom of the stroke -- a similar feel in this regard to that of Topre or IBM buckling spring switches.
Indeed, after replacing the keycaps and adding the O-rings, I found that although I still preferred the feel of Topre or IBM buckling spring switches, I enjoyed the Cherry mx clears much more than before, and I liked the clears more than any other Cherry mx switch I had tried to date (black, brown, red, blue, and green). Moreover, objective tests carried out on Typeracer showed that my typing speed and accuracy were markedly better with Cherry mx clears than with any other Cherry switch and even better than my performance on Topre or IBM buckling spring switches. I attribute this, at least in part, to the fact that I make fewer accidental key presses with Cherry mx clears.
Despite my improved typing performance on mx clears, I find prolonged typing sessions with these switches tiring. I can type comfortably much longer using Topre (45g or 55g) or IBM buckling springs (Model M or F). In addition, I still find the feel of the mx clears raspy or gritty, although this sensation has been diminished with the change in keycaps and the installation of O-rings. Nevetheless, I am glad that I did the experiment with changing keycaps and installing O-rings, as this led me to discover that I could tolerate mx clears after all, and rather than putting the Kul ES-87 up for sale for want of liking the switches, it can now be a keeper.
Indeed, all things considered, I like the Kul ES-87 better than any other Cherry mx TKL keyboard I have tried, including the Filco Majestouch 2 TKL Ninja with mx blues, Ducky Shine II with mx browns, and CM Storm QFR Stealth with mx greens. The Kul ES-87 with mx clears even compares favorably with my non-Cherry TKL keyboards, such as the CM Novatouch, but I would still give a higher rating to the IBM Model M SSK and the RF 87u, both the 45g and 55g versions.