IBM 4978 "Beamship" keyboard review (Beamsprings)

User avatar
Chyros

01 Aug 2020, 14:57

Today I show you one of the coolest keyboards in my collection IMO; the IBM Beamship :D . Hope you enjoy the video, this is a pretty special one!

User avatar
dcopellino

01 Aug 2020, 18:07

Wow Thomas, with this review you are adding another piece to archeological industrial history of times gone. Your sources reconstruction appears quite well funded and from the old articles being shown in the videoclip It seems the current nomenclature with alphabet letters was already in use at the time of those magazines as good reference for identifying the different ibm keyboard generations. So, what about series "E", quoted by the article?
Perhaps, it distinguishes the prototypes that gave origin to the 4707 "bank" keyboard series including kishaver, F77 and the 107 keys (or cloned pingmasters)? Mine is only a very shallow deduction that lets you the last word about.

User avatar
ddrfraser1

01 Aug 2020, 18:48

I wonder if models C through E were botched attempts at cheaper variants til they landed on the buckling spring. Maybe foam an foil? poor performing leaf spring? Who knows what else. Perhaps the switches from the PC Jr were derived from one of these failed attempts.

User avatar
Redmaus
Gotta start somewhere

01 Aug 2020, 21:51

Just an unreal board.... I have been eyeing the one in JP!'s collection since I first saw it.
A beamspring battleship, so cool! You should list another video after you convert it and possibly add a solenoid driver. One question I have to ask, where did you score this?

User avatar
Chyros

02 Aug 2020, 00:00

Redmaus wrote:
01 Aug 2020, 21:51
Just an unreal board.... I have been eyeing the one in JP!'s collection since I first saw it.
A beamspring battleship, so cool! You should list another video after you convert it and possibly add a solenoid driver. One question I have to ask, where did you score this?
Got this one from eBay :) . A converter is already underway, and I hope to add a solenoid to it too, yes :) . Hopefully it'll cameo again in another video down the line!
ddrfraser1 wrote:
01 Aug 2020, 18:48
I wonder if models C through E were botched attempts at cheaper variants til they landed on the buckling spring. Maybe foam an foil? poor performing leaf spring? Who knows what else. Perhaps the switches from the PC Jr were derived from one of these failed attempts.
I'm almost certain that C and E were intermediate or just other alternatives, yes. Whatever they are, they're old; the CBS patent dates from 1977!
dcopellino wrote:
01 Aug 2020, 18:07
Wow Thomas, with this review you are adding another piece to archeological industrial history of times gone. Your sources reconstruction appears quite well funded and from the old articles being shown in the videoclip It seems the current nomenclature with alphabet letters was already in use at the time of those magazines as good reference for identifying the different ibm keyboard generations. So, what about series "E", quoted by the article?
Perhaps, it distinguishes the prototypes that gave origin to the 4707 "bank" keyboard series including kishaver, F77 and the 107 keys (or cloned pingmasters)? Mine is only a very shallow deduction that lets you the last word about.
A lot of this history dates from some right sterling work done here at DT. That internal IBM article and the manual about Type A and B keyboards were especially revelatory :D .

kmnov2017

02 Aug 2020, 00:43

Redmaus wrote:
01 Aug 2020, 21:51

.....One question I have to ask, where did you score this?
I saw this in May, it's probably the same one https://www.ebay.com/itm/IBM-TERMINAL-K ... cvip-panel

kmnov2017

02 Aug 2020, 00:45

Chyros wrote:
02 Aug 2020, 00:00
.... A converter is already underway, and I hope to add a solenoid to it too, yes :) ...

The universal through hole converter with QMK will work on this beamspring.....

User avatar
PlacaFromHell

02 Aug 2020, 01:04

ddrfraser1 wrote:
01 Aug 2020, 18:48
I wonder if models C through E were botched attempts at cheaper variants til they landed on the buckling spring. Maybe foam an foil? poor performing leaf spring? Who knows what else. Perhaps the switches from the PC Jr were derived from one of these failed attempts.
Behold this thing patented a year after the beamspring, by the same dude who designed the beamspring switch. Maybe this one is our model C?
Spoiler:
Image
I also found two other switches by Richard H. Harris, but one has just a description of how actually works.

https://patents.google.com/patent/EP0075088B1/en

https://patents.google.com/patent/GB136 ... Ibm&page=4

User avatar
depletedvespene

02 Aug 2020, 01:25

so... many... alphas...

User avatar
ddrfraser1

02 Aug 2020, 05:24

PlacaFromHell wrote:
02 Aug 2020, 01:04
ddrfraser1 wrote:
01 Aug 2020, 18:48
I wonder if models C through E were botched attempts at cheaper variants til they landed on the buckling spring. Maybe foam an foil? poor performing leaf spring? Who knows what else. Perhaps the switches from the PC Jr were derived from one of these failed attempts.
Behold this thing patented a year after the beamspring, by the same dude who designed the beamspring switch. Maybe this one is our model C?
Spoiler:
Image
I also found two other switches by Richard H. Harris, but one has just a description of how actually works.

https://patents.google.com/patent/EP0075088B1/en

https://patents.google.com/patent/GB136 ... Ibm&page=4
Interesting. Looks like an early attempt at a plate spring. I wonder if plate spring corresponds to a letter 🤔

User avatar
Chyros

02 Aug 2020, 10:51

PlacaFromHell wrote:
02 Aug 2020, 01:04
ddrfraser1 wrote:
01 Aug 2020, 18:48
I wonder if models C through E were botched attempts at cheaper variants til they landed on the buckling spring. Maybe foam an foil? poor performing leaf spring? Who knows what else. Perhaps the switches from the PC Jr were derived from one of these failed attempts.
Behold this thing patented a year after the beamspring, by the same dude who designed the beamspring switch. Maybe this one is our model C?
Spoiler:
Image
I also found two other switches by Richard H. Harris, but one has just a description of how actually works.

https://patents.google.com/patent/EP0075088B1/en

https://patents.google.com/patent/GB136 ... Ibm&page=4
Very interesting! Yes, that's definitely a possibility I think :) . Shows that they had been thinking of buckling springs for a long time - ten years before actual production!

User avatar
shine

02 Aug 2020, 16:31

beamship is the absolute dream for a beamspring collector, at least mine :)

so envy! enjoy it!

John Doe

02 Aug 2020, 16:57

Why you guys always could hunt some endgame stuff on ebay before I even have glance at it?

Damn... :evil: :evil: :evil:

User avatar
hellothere

04 Aug 2020, 03:31

As if I had a spare $1500 to spend on a keyboard :D.

User avatar
Invisius

04 Aug 2020, 06:03

Talk about a pristine example, glad it ended up in good hands. Was worried for a sec the desk was about to collapse after that drop test :lol: I look forward to seeing it converted, and wonder if you'll be able to get the macro keys working.

For the letter variants, I think there's a separate thing going on than the typical "model F/M" thing we're used to. IBM catalogs of the 70s-80s frequently used "Type A/B/C" for base variants of accessory options, and "1/2/3/4" for smaller variants - see https://www.ardent-tool.com/ibm5550/option_1-e.html for a quick reference

IBM used the Model A-D designations for typewriters launched between 1948 and 1967, coincidentally around the time the Beamspring switch was released. The Selectric was developed and released years before the model D, so It's not clear whether they switched to Model E internally to reference the Beamspring, or planned to use one of those prototypes PlacaFromHell linked, but ended up skipping a whole generation. Since they did just this with the jump from Model F->M, it's not unreasonable.

User avatar
sharktastica

04 Aug 2020, 11:51

Invisius wrote:
04 Aug 2020, 06:03
For the letter variants, I think there's a separate thing going on than the typical "model F/M" thing we're used to. IBM catalogs of the 70s-80s frequently used "Type A/B/C" for base variants of accessory options, and "1/2/3/4" for smaller variants - see https://www.ardent-tool.com/ibm5550/option_1-e.html for a quick reference
IBM product support manuals also used a similar nomenclature when it was appropriate. For example, for the 6580 Displaywriter, the designations "Type A" and "Type B" were used to refer to beam spring and Model F variants of the system's keyboards respectively - as per Chapter 5, pages 63 and 64 of the bitsavers' PDF scan of the IBM Displaywriter System Product Support Manual (Feburary 1983 revision); figures 5-1 and 5-2.

User avatar
depletedvespene

04 Aug 2020, 15:15

Invisius wrote:
04 Aug 2020, 06:03
Talk about a pristine example, glad it ended up in good hands. Was worried for a sec the desk was about to collapse after that drop test :lol: I look forward to seeing it converted, and wonder if you'll be able to get the macro keys working.
I'd love to see a picture of the PCB as well, to think about how the bottom row could actually made to be, if chopping off the relevant bits of the top case weren't a sacrilegious proposition.

User avatar
Noobmaen

04 Aug 2020, 15:38

It's as flexible as it gets, by shortening the spacebar one can get a few more modifiers without modifying the top case, but it would be cramped. Image Image

User avatar
depletedvespene

04 Aug 2020, 15:44

GFL, that's amazing. One could even do a split space bar on that PCB.

Excuse me while I hop on to my time travel machine...

John Doe

04 Aug 2020, 16:26

Plus with extra keys and caps the keyboard could be added more keys judging from the PCB and plate? Cool.

User avatar
SneakyRobb
THINK

06 Aug 2020, 00:17

Great video and very informative as always. Well done!

NGL pretty jelly of that board.

PlacaFromHell wrote:
02 Aug 2020, 01:04
ddrfraser1 wrote:
01 Aug 2020, 18:48
I wonder if models C through E were botched attempts at cheaper variants til they landed on the buckling spring. Maybe foam an foil? poor performing leaf spring? Who knows what else. Perhaps the switches from the PC Jr were derived from one of these failed attempts.
Behold this thing patented a year after the beamspring, by the same dude who designed the beamspring switch. Maybe this one is our model C?
Spoiler:
Image
I also found two other switches by Richard H. Harris, but one has just a description of how actually works.

https://patents.google.com/patent/EP0075088B1/en

https://patents.google.com/patent/GB136 ... Ibm&page=4
Hi these great links and cool stuff. For the one with only a description you can click the links for "External linksEspacenet" in the blue window thing. This will often open very useful information. A lot of old patent documents are a mess of work to get through and google only scratches the surface.

https://worldwide.espacenet.com/patent/ ... GB1363777A

In this case that second one is the beamspring switch. You can even see how the patent attempts to cover stuff like variations on the theme. Upside-down beams for instance
upsidedown beamspring.JPG
upsidedown beamspring.JPG (20.32 KiB) Viewed 277 times

For the linear beamspring switch, this one
linear beamspring.JPG
linear beamspring.JPG (17.13 KiB) Viewed 277 times
https://worldwide.espacenet.com/patent/ ... US3693059A

You can read the description where Richard says:
Spoiler:
In view of the foregoing and other problems in the prior art, it is an object of this invention to achieve an almost linear capacitance variation in response to actuator or key stem displacement so that the "on" and "off" sensing positions or capacitance values will be separated from one another over a relatively large distance so that good physical key hysteresis is produced in an improved way.

It is another object of this invention to allow significant amounts of overtravel of the actuator member beyond the point necessary to actuate the switch in an improved way so that good "key feel" is imparted to the operator.

Additionally, it is an object of this invention to make an improved capacitive switch in which the undesired effects of surface irregularities or dirt on the capacitive coupling plates are overcome in an improved manner.

Still another object of this invention is to simplify and improve the structure by reducing the number of active parts and by increasing the number of functions served by some parts in order to achieve an overall improved switch.
Spoiler:
"While mechanical means for providing physical hysteresis have been successful in the past, and devices for overcoming the effects of surface irregularities and dirt have been developed, the devices so produced are generally complicated in structure with an increased number of parts, are difficult to assemble and are costly to service. In particular, to overcome capacitive variance due to surface irregularities and dirt, it has been necessary to produce very flat coupling surfaces and then isolate these surfaces from dust and dirt. Both of these necessary precautions are expensive and make such devices less desirable."
then he says
Spoiler:
"Additionally, it is an object of this invention to make an improved capacitive switch in which the undesired effects of surface irregularities or dirt on the capacitive coupling plates are overcome in an improved manner."
Then
Spoiler:
"This invention satisfies the foregoing objects and solves the above mentioned problems by utilizing a deformable or conformable, resilient, generally bowed or U-shaped spring member as both a capacitive coupling plate member which conforms to the surface irregularities of the coupling members which it couples, and as a key restoring bias means. In this invention, the U-shaped spring is mounted on the end of a key stem or actuator which is vertically movable in relationship to two co-planar, dielectrically insulated capacitive plates. When the operator displaces the key stem, the U-shaped is brought into increased contact with the dielectric layer overlying the capacitive plates. As additional force is applied to the key stem, the U-spring flattens out and conforms to the surface of the dielectric over the capacitive plates, thereby increasing capacitance in a nearly linear fashion and also increasing the amount of force required to displace the key further as a function of key travel. To prevent erroneous keying, the key is given an initial preload by displacing the key permanently by a few hundredths of an inch, thereby creating an initial spring force which must be overcome before the key is moved further."

So if we are reading this right, this patent is for a near-linear capacitive beamspring where Richard basically removes almost all the parts from the beamspring switch and linearizes it. Wtf right?

User avatar
darkcruix

Yesterday, 19:02

It would be possible to even get a standard NAV-Cluster created with some re-arrangement of the switches... Looking so forward to see a video about the working condition of it and how it is programmed in the end. Insane!

Post Reply

Return to “Reviews”