- Published 25 February 2021
- Revised 10 April 2021
- Est. Reading Time: ~10-14 minutes
Around the end of January, I was contacted by Unicomp's Vice President of Development, Don Bowman, with a proposal for me to conduct a critical review of a prototype of their upcoming Mini Model M tenkeyless keyboard. Thrilled with the prospect of helping bring the Mini M to fruition in the best quality possible, I gladly accepted, and after a few weeks with the keyboard, here are my findings, things I reported back to them, and what I think about this upcoming product!
Please note four things: firstly, this is a prototype keyboard, improvements will be made based on comments gathered, thus not fully representative of the final product coming next month. Rest assured, they took my comments seriously and the final product will be even better than this one. Secondly, these photos were taken after a week or so of testing, thus fingerprints and dust in some of the photos may be unavoidable. Thirdly, I reference the New Model M several times in this article - for context, you can check out my review of that keyboard here. Fourthly, I did not receive any money from Unicomp to do this nor am I monetising this endeavour - everything I've written is my honest thoughts and is unfiltered by external parties.
Around the outside
A lot of the styling update comments I made for the New Model M in that recent review are apt here - this is thing is a looker from the front. Blemish-less, dimple-less, and slick. Immediately obvious, there are a few minor tweaks since Unicomp's gift certificate photo from around a year ago that I'll explore below.
The first major change from 2020 is that the Alt keys now have green legends, echoing PC-compatible Model Ms from the '80s and '90s. Quite a nice callback and attention to detail.
Another classic callback is the green sublegend for SysRq. NumLk is also now blue, which is something the original SSK didn't have but I believe found on some terminal Ms.
Completely absent from last year's photo is the overlayed numeric keypad legends on the alphanumeric block. Just like the original SSK, this numpad is toggled by pressing Shift and the NumLk/ScrLk key. Unlike the dull grey or black from SSKs of the past, the numpad's legends are now blue. I think this is a definite visual improvement over the old SSK. We also get a glimpse of the rearranged modifiers at the bottom - Alt and Windows key have been swapped, matching post-Windows PC keyboards in arrangement but not in unit size. My guess, the unit sizes remain the same due to the need to have a universal barrel plate across their designs.
Also of note is the misalignment of the Windows key. I raised this with Mr Bowman and it will be sorted before release, as explained in the feedback section of this article. In fact, Unicomp has since shown photos of the Mini M with this issue corrected!
Despite the fault with the Windows key, this prototype has the best legend alignment I've seen from Unicomp in many years. Shown below is a side-by-side comparison between my late 2020 New M legends and this prototype, and you can see that the return/enter, backslash, quotes, GUI menu keys are all improved. As pointed out in my New M review, the return/enter key was the biggest current offender in their alignment and now it's fixed!
The lock-light LED overlay has changed since the earlier photo. Gone are the symbols-based overlay, and in is the clean text-based overlay.
And finally, we have the new modular Type-A USB port. It is a sort of 'self-locking' design and it works fairly well. You can still pull a cable out with considerable force, but this will for sure cut down on many accidental removals. In practice, you should use something flat like a screwdriver to release a cable.
On the surface, I found two issues with fit and finish and one hold-over from the New M observations. The first issue with this prototype was that I could move the port and the board inside by applying force to the release clip. I raised this with them and it will be fixed prior to release as described in the feedback section.
The second issue that required my soft-box light to notice a few days after I started using the Mini M was a scratch in the top-left case. I inquired about this, and Mr Bowman said the scratch was likely due to the keyboard being banged around his office before it was shipped to me. The final product should not have this sort of issue.
And finally, the other thing worth mentioning is that the rings found on the back case of the New M in my previous article are present here. Not surprising given the Mini M is a 'sawn-off' version of the New M in terms of aesthetic, but I still don't believe it's not a big issue given it's on the back. But, it's worth mentioning though.
Using the Mini M, its sound & new features
In terms of the typing feel, it's simply a Model M. What is interesting for me is that I have a hard time choosing a favourite between this and my old SSK. For context, I have a 1987 IBM Space Saving Keyboard that I received in essentially unused condition, and I was seriously impressed by how well this keyboard holds up. In terms of sound, there's an obvious difference but I would argue the difference isn't as wide apart as usual (for example, the difference between the New M and a 1995 Lexmark terminal Model M in the previous article). Most notably and as evidenced in the typing demos below, the spacebar sounds different, largely because Unicomp applies grease to the stabiliser from the factory.
Actual day-to-day usage was pleasant. If you've tried a Model M, this shouldn't surprise you that it's the 'same-old' experience. But, Unicomp has made some changes to the Model M design with the Mini M that are very interesting. Specifically, there are three new features.
The newly designed membrane
Whilst I typically question whether two-key rollover is as big of an issue as people claim, especially given 2KRO doesn't mean a maximum of 2 keystrokes together (in fact, Model Ms can handle 3-4 keystrokes together in some cases). But no doubt, an improvement in this regard will be welcomed by everyone. Unicomp is currently calling this improved gamer support, so you may be curious what the "KRO" rating is. Since the technology is still fundamentally membrane-driven and doesn't use diodes, 2KRO is technically still the rating. But, as explained below, they've been able to 'hide' this limitation. As it stands, I would rate this a comfortable 6-key rollover-like experience. Ten keys together are indeed supported, but due to some issues with the pre-production code this keyboard runs, combinations higher than 6 are prone to some phantom key instances. In case you're curious, Unicomp was up-front about the code my prototype runs, and the problematic combinations I found were quite artificial use cases. The good news is that Unicomp tried to reproduce the two most problematic combinations I found on a keyboard with production code and found the issues did not affect it. More info in the feedback section.
As for the technical details behind how they achieved what seems like 6KRO+; as explained to me by Mr Bowman, the membrane's matrix is 12x16 (ie, 192 members) for just the 87 keys, contrasting the 8x16 matrix their present 1xx-key keyboards use. This member increase allows them to pick what members to use as actual keys based on their unlikeliness to block with others, allowing them to effectively hide the fundamental 2KRO limitation of the technology. This, along with their latest code, results in what seems like an improved rollover experience and a greatly reduced chance of phantom keys being outputted.
The addition of a pressure-locking USB Type-A port
To be clear, this isn't a complete lock solution - with enough force, you can pull a cable out. But, this will greatly cut down on the number of accidental disconnects. Whilst my initial train of thought would have preferred a Type-C port due to the wide support of artisan cables, I'm convinced Type-A was the right call due to the lack of non-screwing locks for Type-C (as fas as I could tell) and the fact Type-A is a robust and proven reliable connector. And of course, everyone should be glad they didn't go with a Micro-B connection!
The addition of lock-light LEDs
I am aware already opinion on the inclusion of lock-lights is rather split. Given the fact the Mini M has an overlayed numeric keypad feature, the inclusion of lock-lights may be very useful since I can imagine accidental toggles leading to frustration as letters seemingly output numbers and operators! I know from experience, I was frustrated to experience this exact issue when I first got my 1987 SSK that, of course, gives you no clues if num lock is engaged.
My feedback to Unicomp
Here is a summary of the biggest things I raised to Unicomp regarding my findings and experience with the Mini M. Most of these I've touched base with already, but just so it's clear.
|Issue found||Their response|
|Scratch on the top-left||Before they shipped this to me, it spent a few days being knocked around Mr Bowman's office and thus not representative of the condition they ship with. For the final product, don't expect scratches like this.|
|Misaligned right Windows key||This will be resolved by first ship. In fact, the photos on Unicomp's website and Facebook posts reveal they have indeed sorted this!|
|Moving USB port and board||This will be resolved by first ship.|
|Weird phantom key behaviour||This is due to the non-production code this prototype runs. The final release code they recently compiled should not exhibit this behaviour with these scenarios I found. The scenarios were pressing Q+W+E+[ results in ]+\ being returned with the rest, and A+S+D+F+H+J+K+L+Space results in down arrow being returned with the rest. Both very artificial combinations.|
|Scratch artefacts on the bottom interior around rivet support recesses (photos pending, see New M review for context)||The scratches in the Mini M (and New M) interiors are due to a change in height of those bosses. Their tool maker shaved away some material, and they didn't specify that specific surface to be a finished (ie, polished) surface to suppress time and monetary costs on something that only affects the interior.|
My general feelings toward this keyboard are a mix of the same-old from the New M review and general relief that the Mini M isn't a simple SSK rehash. All my comments for the slickness of the New M apply here - this is clean and spotless on the top and I'm really happy with the finish. Coupled with the innovations Unicomp has done (reduced the side bezel sides, improved the contrast of the numpad sublegends, the addition of GUI keys, and enhanced membrane), I genuinely believe I'll be using this more than my 1987 SSK. Over the last three weeks with this keyboard, I have indeed come to love this device!
To address the elephant in the room for some - I know a number of enthusiasts were hoping that the Mini M would be a simple SSK reproduction. But, of course, that was never going to happen since it wouldn't make business sense. Where it matters, the profile, switches and utilitarianism, the Mini M remains a Model M. But, it has features that give it mass-market appeal (especially with the inclusion of Windows keys and black colour). That said, I would welcome the possibility of a beige case and perhaps a blank overlay for the lock-lights, but I am still well happy with this Mini M and its style. And the fact is, I'm happy to have the SSK and the Mini M as fairly different animals since it keeps both unique.
I want to thank Unicomp and Mr Bowman for giving me this fantastic opportunity to try this keyboard before launch. I'm thrilled to be a part of and hopefully contributed to the development of the Mini Model M. Anyway, before heading out, see below links to Unicomp's latest update for the Mini M availability and features, where to buy one, and my discussion threads for Unicomp's latest update. And, some more photos for your viewing pleasure.
Corrections & changes
- Due to some confusion, I've altered some of the wording under the "The newly designed membrane" heading. It previously seemed like I claimed the new membrane design is an outright 6KRO+ technology. Technically, claiming such wouldn't be true since it is still a membrane circuit and doesn't make use of diodes in the design. What I meant is that they achieved a 6KRO+ like experience by planning their matrix design to effectively hide any combinations that block at just 2 keys. Anyway, this doesn't alter my recommendation or the fact that Unicomp has greatly improved the technology's experience for the end user. It's good enough that you'll never notice the fundamental limitation hidden underneath. But still sorry for any confusion, and for the sake of accuracy and integrity, I've tried to make things clearer.