Keycaps - PBT vs ABS

User avatar
ShivaYash

02 Oct 2017, 15:36

Dear all,

Why do most, if not all KB manufactures, high-end stuff, TOPRE et al, provide PBT spacebars? Is this particular key tricky to produce in PBT versus ABS?
I have purchased some PBT spacebars for my various KBs from MassDrop but until now did not question why they don't just come with.

With thanks,
SY

Engicoder

02 Oct 2017, 15:44

From the wiki:
"Because of shrinkage during demoulding, PBT is also rarely used for the largest key, the space bar, with a notable exception being Cherry."

wiki/Keycap_material#PBT

User avatar
infodroid

02 Oct 2017, 16:16

I hear this excuse all the time, but it doesn't make any sense to me. Why should shrinkage affect the ability to produce large PBT keys?

Surely manufacturers can simply account for the shrinkage in their measurements. If the underlying reason is, as I suspect, that the shrinkage is random and unpredictable, then why is it so hard to make the shrinkage uniform and predictable?

User avatar
ShivaYash

02 Oct 2017, 16:20

infodroid wrote:I hear this excuse all the time, but it doesn't make any sense to me. Why should shrinkage affect the ability to produce large PBT keys?

Surely manufacturers can simply account for the shrinkage in their measurements. If the underlying reason is, as I suspect, that the shrinkage is random and unpredictable, then why is it so hard to make the shrinkage uniform and predictable?
Agreed. Paying £££ for a board should mean top quality caps.


Sent from my iPhone using Tapatalk

User avatar
Techno Trousers
100,000,000 actuations

02 Oct 2017, 16:59

As far as I'm concerned, if a company like Unicomp can successfully manufacture PBT space bars, then anyone should be able to do it.

User avatar
hansichen

02 Oct 2017, 17:14

Some companies don't care about their products. Just take signature plastics, they produce pbt caps but even the normal modifiers have warping issues, the space bar is even worse. Other companies like cherry proved that it's possible, in the end we just lack a company who wants to invest time and money into good keycaps.

User avatar
E3E

02 Oct 2017, 17:58

Cherry, NMB, IBM

rich1051414

02 Oct 2017, 18:48

PBT has tendencies to deform. The larger the injection mold, the higher probability it will deform. AFAIK, this is one of the primary reasons PBT costs more.
I think the problem is primarily to do with the cost of quality control. Using ABS instead cuts a LOT of corners, which is too attractive for a lot of keyboard manufacturers to ignore.
A great number of my PBT spacebars have some deformation in them, but not enough to be very noticeable or affect performance.

User avatar
infodroid

02 Oct 2017, 19:21

I don't think a material like PBT deforms at random. Surely there are factors that cause the deformation, which can be controlled. It is not clear why they can't be controlled, or controlled in a cost-effective way.

It is strange that an expensive high-end product like a Topre keyboard can't ship with PBT space bars when even Unicomp can do it, as Techno Trousers reminds us.

User avatar
hoaryhag

03 Oct 2017, 07:07

I've worked at an injection molding plant making industrial shipping containers. You make a lot of bad parts due to bubbling, warping, color changes, and the machines in general needing lots of tweaking. I would imagine if pbt is so sensitive to variables you would need a skilled tech that knows the machine, a good machine to begin with, a properly made mold for the design, the right color feed, and an employee at the machine to monitor it and discard bad parts. In other words, a good company that knows their shit.
http://www.injectionmoldingplastic.com/ ... rpage.html

It's much easier to go with what consistently makes a good product without babysitting the machine all the time. Small, fast parts that are difficult to QC are problematic. You would HAVE to have a separate mold for the spacebar if the keys were pbt wouldn't you. Otherwise, when you have problems, you end up with a shit ton of warped spacebars and the rest of the keys pass QC. Then, they think, why not go with the easier to manage abs, if you are gonna have a separate mold anyway.

andrewjoy

03 Oct 2017, 12:00

Techno Trousers wrote: As far as I'm concerned, if a company like Unicomp can successfully manufacture PBT space bars, then anyone should be able to do it.

That is so true, i would not trust unicomp to make a door stop :P if they can manage it... ANYONE should be able to do it.

And its not as if ABS does not warp its terrible for it, just look at 3d printing , to print ABS correctly you need a heated chamber and cool it consistently, now sure this is due to the process of FDM printing but its not as if its a factor you can ignore.

The model M case is PC right ? Why not use that for spacebars , better than ABS.

Findecanor

03 Oct 2017, 16:42

andrewjoy wrote: The model M case is PC right ?
My old Model M cases are marked "PVC". That plastic has largely been phased because its production has waste products harmful to the environment. It also releases some quite noxious gases when burned.

The Wiki mentions that Model M's were made of ABS or PC-ABS, where PC-ABS is just a mix of A, B, S and PC.
I think that IBM Ambra PC's and Model M2 keyboards were made of PC-ABS. I think Unicomp uses ABS.
Signature Plastics' clear keycaps are made of PC.
Last edited by Findecanor on 03 Oct 2017, 16:51, edited 1 time in total.

User avatar
seebart
Offtopicthority Instigator

03 Oct 2017, 16:47

hoaryhag wrote: I've worked at an injection molding plant making industrial shipping containers. You make a lot of bad parts due to bubbling, warping, color changes, and the machines in general needing lots of tweaking. I would imagine if pbt is so sensitive to variables you would need a skilled tech that knows the machine, a good machine to begin with, a properly made mold for the design, the right color feed, and an employee at the machine to monitor it and discard bad parts. In other words, a good company that knows their shit.
http://www.injectionmoldingplastic.com/ ... rpage.html

It's much easier to go with what consistently makes a good product without babysitting the machine all the time. Small, fast parts that are difficult to QC are problematic. You would HAVE to have a separate mold for the spacebar if the keys were pbt wouldn't you. Otherwise, when you have problems, you end up with a shit ton of warped spacebars and the rest of the keys pass QC. Then, they think, why not go with the easier to manage abs, if you are gonna have a separate mold anyway.
Very interesting. Good to get some first-hand information from someone who's actually worked in this field.
andrewjoy wrote: That is so true, i would not trust unicomp to make a door stop :P
As bad as it sounds me neither!

User avatar
Daniel Beardsmore

03 Oct 2017, 19:07

hoaryhag wrote: I've worked at an injection molding plant making industrial shipping containers.
Most Alps clone switches are devoid of branding and, unlike Cherry MX clones that all look different, Alps clones tend to all look virtually identical. The only identifying marks are the mould cavity numbers. From what Tai-Hao were saying (and they didn't clarify this in detail), a mark in the form "A23" means mould 1 (A), cavity 23. For some reason most SKBL/SKBM switches are mould 5, and less commonly mould 4, but seldom if ever moulds 1–3. Cavity numbers can go over 100. With Alps SKCL/SKCM, the extra letters within circles seem to be rods inserted into the mould to denote batch/shift, as there already is a letter-number pairing (e.g. "10D") in each cavity. (The really weird ones is where there is only a single number (e.g. 3 or 13) repeated in opposite corners, e.g. top-left 3 and bottom-right 3)

What puzzles me is when this numbering is not consistent. For example, older Alps clones of the Himake type tended to just have a single number (e.g. "51") and then they went up to complex numbers like "1H37" (three tiers of identification).

Some changes appear to be logistical (I guess the move from "51" to "1H37" denoted an increase in the number of mould machines to satisfy a huge increase in demand), but I've wondered how many cycles a Japanese or Taiwanese mould is likely to do before you have to start over. This may give a very vague idea about which if any changes to moulds were done out of a need to retool.

Off-hand I can't think of any examples of totally unexplained mould changes. I do know of a switch series where the mouldings are pitted and rough, where the moulds seem to be very old — the branding has been removed, and the moulds previously belonged to a different switch manufacturer.

Just rambling really as it's been so long since I was looking at this stuff that I forget what mysteries there were.

User avatar
hoaryhag

04 Oct 2017, 07:30

Daniel Beardsmore wrote:
hoaryhag wrote: I've worked at an injection molding plant making industrial shipping containers.
Most Alps clone switches are devoid of branding and, unlike Cherry MX clones that all look different, Alps clones tend to all look virtually identical. The only identifying marks are the mould cavity numbers. From what Tai-Hao were saying (and they didn't clarify this in detail), a mark in the form "A23" means mould 1 (A), cavity 23. For some reason most SKBL/SKBM switches are mould 5, and less commonly mould 4, but seldom if ever moulds 1–3. Cavity numbers can go over 100. With Alps SKCL/SKCM, the extra letters within circles seem to be rods inserted into the mould to denote batch/shift, as there already is a letter-number pairing (e.g. "10D") in each cavity. (The really weird ones is where there is only a single number (e.g. 3 or 13) repeated in opposite corners, e.g. top-left 3 and bottom-right 3)

What puzzles me is when this numbering is not consistent. For example, older Alps clones of the Himake type tended to just have a single number (e.g. "51") and then they went up to complex numbers like "1H37" (three tiers of identification).

Some changes appear to be logistical (I guess the move from "51" to "1H37" denoted an increase in the number of mould machines to satisfy a huge increase in demand), but I've wondered how many cycles a Japanese or Taiwanese mould is likely to do before you have to start over. This may give a very vague idea about which if any changes to moulds were done out of a need to retool.

Off-hand I can't think of any examples of totally unexplained mould changes. I do know of a switch series where the mouldings are pitted and rough, where the moulds seem to be very old — the branding has been removed, and the moulds previously belonged to a different switch manufacturer.

Just rambling really as it's been so long since I was looking at this stuff that I forget what mysteries there were.
You would know better than I. I don't own any boards with alps clones. If they all look indentical, it's a good indication they were all made in the same place with the same color, the same plastic, and the same molds. Molds can be used for many years if they survive and can be repaired when they are damaged. They are all well made and the big ones are very expensive, 500k - 1mil a mold and more. Small repairs like broken mold pins and added tooling that blocks off portions of the mold are done in house. Cracks and chips on the mold are repaired by the company that makes them and they have to be sent off. Molds can be damaged for several reasons. The mold could be dropped by the setup crew changing out the molds. The machine might decide to slam shut, instead of locking together gently, and destroy mold pins, the mold, and hurt the machine. Most of the time its just some broken pins, but the machines can spaz out and damage molds. The mold getting too hot and cooling down too fast, cracking the mold. Weak points on the mold due to design. Previous repairs failing, age, bad setup of the mold or machine when the mold is installed, flawed mold, etc. That said, molds aren't damaged badly enough to send off very often.

They would increase the number of lines running the switches as they received more and larger orders. Since they were smaller molds and probably running around the clock, they might have had a back up mold to keep production going while one was being repaired. Designations would get more complicated over time as they increased their output. Down the line, if problems are noticed with the switches due to the mold, they can pin it down to when and where.

The switches had to be assembled. Particular molds, machines may have consistently provided the parts for a particular switch type being assembled nearby. If things are running smoothly, a mold could stay running on the same machine for quite a long time.

sandylu1995

07 Dec 2017, 12:08

I like ABS more. I don't prefer PBT because it glitters after using too much... looks oily too me.

Post Reply

Return to “Keyboards”