Let's create the FSSK/FEXT = DONE !!!!

[chzel]

Looking good i$!
I can't wait to get my hands on mine!
Regarding the USB connection, maybe a solution like the one Phosphorglow used on his Universal M Controller would be good?

Essentially a small breakout board, with USB B connector and holes spaced correctly for the pins on the SSK case.
Phosphorglow uses nylon risers slit along their length as clamps to hold the PCB down.


post219719.html#p219719

Also, regarding mixing flippers, on my AT I have a couple of XT flippers with absolutely no issues.

[0100010]

Looks like I have 229 Model F flippers. Enough for 2 sets of 84, with 61 left over.

Nice! Where did they all come from? I have an old XT I'll probably sacrifice, then bite the bullet and buy some Ellipse flippers for a backup FSSK.

All the flippers used by an xwhatsit controller shall be similar. There is one single voltage threshold define for all keys. I do not know if mixing them will work.

The FSSK 0001 uses flippers from an XT and a bigfoot donors. The issue that I reported before with the self repeating "1" and the adjustment that I had to do to the threshold, was very probably because the mix.

They mostly came from a couple 3178s. So, to have two sets of 84 flippers that all match, is no issue (I think one is 87 pieces, the other 88). Looks like Murium and pcaro called dibs; will PM you guys.

[0100010]

DT.

[idollar]

Looking good i$!
I can't wait to get my hands on mine!

I will ship it on Monday very probably.

Regarding the USB connection, maybe a solution like the one Phosphorglow used on his Universal M Controller would be good?

Essentially a small breakout board, with USB B connector and holes spaced correctly for the pins on the SSK case.
Phosphorglow uses nylon risers slit along their length as clamps to hold the PCB down.

I will check it. Perhaps the best is to create a version of the xwhatsit, specially designed for the Ms, using a solution similar to Phosohorglow. If we add a connector to the original matrix also, It could be used with the regular matrix also. The controller can sens a normal matrix also

Perhaps this is the next step to do, my next project... I will think about it.

[idollar]

The shipping of the patched and tested (without controller, just conductivity) is ready.
Please check your PMs ....

[vsev]

You're the guy man, just able to write those letters but hey ( wcass also in) IBM lovers will thank you so much ...

[idollar]

First batch of prototypes in their way home.

[idollar]

Dear prototypers:

A couple of hints on how to use your PCBs:

1. Attached you can find the xwhatsit configuration that I am using to type this post
2. The voltage threshold that I am using is 154
3. You have to isolate the bottom metal plate to ensure that there is not shortcuts with the PCB. I have used packing tape to cover the side of the metal place facing the PCB
4. I have also added the rubber layer of the membrane between the metal plate and the PCB. I like more the feeling of the keyboard with it.
5. The big-top insert holes in the PCB that you will receive are just enough to go through the inserts of the case. File them a little bit. If you do not do it, taking the PCB off will be very difficult. The PCB shall sit smoothly.
6. There are some holes that need adjustment. The final version will adapt just a little bit their position. They are on the left side of the keyboard. Before you start screwing the PCB in the board, ensure that it is a align as possible.
7. I highly recommend to use a connector -- note that my configuration will work only if you use the connector in the middle --. This will allow you to mount/umount the board. It will also allow you to reuse the xwhatsit in other keyboards. I will find out the references for the three parts that I have used and post them here.
8. Do not tight the screws too much. The pressure will change the distance between the PCB and the flippers affecting the sensing. If a key senses at a lower voltage ("1" in my case) or at a lower voltage ("u" in my case!) tighten just a little bit the screws under the key to increase the sensitivity of it. I actually do not know how will Ellipse resolve this issue without screws. I am afraid that it may be difficult to fine-tune the sensing of the keys in his project. If I would design a keyboard, I would use screws. That's for sure.
9. The patches are working perfectly in the FSSK0001. I keep typing with it

[XMIT]

What patches are you talking about? I'm still not sure if the FSSK boards themselves need any rework.

[shreebles]

What patches are you talking about? I'm still not sure if the FSSK boards themselves need any rework.

He means this

In the down side:

There are two buts in the PCB

a) A missing trace
b) A top trace in the down button connected to the wrong pad

I could solve both with a little of wire and some handcraft ability with in soldering. I will offer the repair to the prototype contributors.

[idollar]

What patches are you talking about? I'm still not sure if the FSSK boards themselves need any rework.

The FSSK prototypes need some rework. We are still testing them.
I am fine tuning the controller also.

Once I am sure that all the issues are fine, I will open the group buy.

[hammelgammler]

Just received my prototype pcb, great packaging! Thank you for that.

Regarding the xwhatsit controller, doesn't it need any ground connection? I can't really see one in your pictures.
Little bit confusing, as I thought it always needs a ground like on normal Model F's?

[idollar]

Just received my prototype pcb, great packaging! Thank you for that.

Regarding the xwhatsit controller, doesn't it need any ground connection? I can't really see one in your pictures.
Little bit confusing, as I thought it always needs a ground like on normal Model F's?

Pins 1,2,29,30 and two in the middle, separating rows and columns are ground.
I have not grounded the metal plate and my prototype works. I am having some issues with the key "1" which sometimes repeats by itself. Now that you tell us, perhaps grounding the metal plate will improve it.
I will test it.

You will notice that some of the pads are scratched. I removed the isolation layer in some corners to check the for continuity before mounting the set.

I recommend to use a connector. This will ease the testing process.

Do not hesitate to ask questions.

Do not forget to register your working FSSK

[hammelgammler]

Oh alright then. Yeah I will look for a connector, sadly I don't have any at hand.


Do you think it's really necessary to isolate between the PCB and the backplate? I mean, in normal Model F's there no isolation whatsoever. Sure pre-caution is always better, but I thought about because Model F's don't have any.

[fohat]

Do you think it's really necessary to isolate between the PCB and the backplate? I mean, in normal Model F's there no isolation whatsoever. Sure pre-caution is always better, but I thought about because Model F's don't have any.

Nay, nay. Between the PCB and the back plate is a clear sheet of mylar, and between the PCB and the front plate is a foam mat.

[hammelgammler]

Oh really? Then I hadn't reminded it correctly, my bad. Or maybe I never saw that there was something between them.
Thanks!

[wcass]

I've seen some with transparent plastic sheet between PCB and back plate - and some without. You might have better luck one way or the other, so try both.

[idollar]

The following shows the very thin layer of foam that I have added between the metal plate and the PCB.
I also isolated the metal plate with wide tape, to ensure that the bottom pads of the PCB would not short-circuit.

DSC_4869.jpg

(in a previous post)

This is why I did.
I have also added the original mat layer in the sandwich between the PCB and the Metal plate.
In total I am testing tree layers: - tape to isolate - VERY thin foam and the mat.
If I would do it now, I would tape the plate and use the mate only.

[idollar]

Do you think it's really necessary to isolate between the PCB and the backplate? I mean, in normal Model F's there no isolation whatsoever. Sure pre-caution is always better, but I thought about because Model F's don't have any.

Nay, nay. Between the PCB and the back plate is a clear sheet of mylar, and between the PCB and the front plate is a foam mat.

References:

here what is on top of the PCB

here what is under the PCB

I need to check what will happen if we place the foam mat between the flippers and the PCB ...

[hammelgammler]

DSC_2867.jpg (408.87 KiB) Viewed 10803 times

I realized something you can see in i$ picture. In the last row, there are two plastic nubs where the flipper don't rest perfectly like every other one.
I will try to burn them off/flat, so that the flipper rest like normal. These nubs shouldn't do anything as far as I can tell.

The difference in feel wasn't the worst of all time, but it should let them feel different because of the lower "travel".

I will let you know when I burned them off and/or I'm finished with the build.

[wcass]

Check to see if those nubs are the back side of the "space bar stabilizer hold-downs".

[hammelgammler]

Seems like I have a problem, or did something wrong.
I have used the mat like i$ said, so it's like barrel-frame -> switches -> pcb -> mat -> plate. The top part of the pcb I isolated because it touches the top part of the plate, everything else should be isolated with the mat, doesn't have any contact as far as I can tell.

But the ibm-caspense software doesn't seem to register my keypresses. At around 140 threshold every key is "pressed" (no keycaps on the switches), but when put some keycaps on, there are still pressed.

I have not grounded anything, it's more or less like i$.
Here's a picture of my messy job.
Any idea what the problem could be?

[XMIT]

Grounding is often the issue. Did you install the bolts for the bolt mod? Are they making contact with grounding pads on the PCB? "Let ground abound" as analog designers often say.

I would isolate the PCB from the plate with a sheet of mylar or a bunch of Kapton (polyimide) tape and get rid of the mat.

Are you sure your barrel plate is in good condition, and free from cracks or anything that could cause some difference in dimensions? Are you certain that you soldered in the xwhatsit in the correct orientation? I believe you have from the photo but it is always good to check.

[idollar]

I realized something you can see in i$ picture. In the last row, there are two plastic nubs where the flipper don't rest perfectly like every other one.
I will try to burn them off/flat, so that the flipper rest like normal. These nubs shouldn't do anything as far as I can tell.

The difference in feel wasn't the worst of all time, but it should let them feel different because of the lower "travel".

I will let you know when I burned them off and/or I'm finished with the build.

Yes, I have also cut them flat.

[idollar]

Seems like I have a problem, or did something wrong.
I have used the mat like i$ said, so it's like barrel-frame -> switches -> pcb -> mat -> plate. The top part of the pcb I isolated because it touches the top part of the plate, everything else should be isolated with the mat, doesn't have any contact as far as I can tell.

But the ibm-caspense software doesn't seem to register my keypresses. At around 140 threshold every key is "pressed" (no keycaps on the switches), but when put some keycaps on, there are still pressed.

I have not grounded anything, it's more or less like i$.
Here's a picture of my messy job.
Any idea what the problem could be?

All right. Lets start with the things that I would check / comment:

1. I have used a ribbon cable for a good reason: the position of the ground, columns and rows are fixed. If you look at the following you will find out that:

   

   2016-02-04 00_19_59-PCB Layout - [E__FSKK - PCB - from scratch_Diptrace_FSSK v0_49.dip].jpg (14.13 KiB) Viewed 10786 times

   • the two extreme pins in both sides are ground (1,2,29,30)
   • the pins separating rows and columns are also ground (11,12)
   This strategic relative position will not be maintained if single cables are not used.
   Anyhow, I do not think that this is the problem, but if all the other checks fail, you may at least try to "order" your connections:
2. Check the connections:
   • 1,2,11,12,29,30 and the ground in the controller, the external of the usb connector all all ground.
   • NONE of the other pins shall be connected to the ground
   • Starting with 3, check if there is a connection with 4,5,6 .... 30
   • Continue with 4, checking 5,6,7,8 ....
   • Do this until you check the 29 which shall be connected to the 30
3. check the connections between the controller and the PCB
4. Ground the bottom plate (I have not done it yet)
5. Disable the last columns or even better, load the configuration that I posted before. The key allocation will not be the same as the FSSK0001 as you do not have the connector that I used in the middle, but you will be using a configuration in the controller that should work. You can solve the key allocation later.

If bad comes to the wort you can always cut one cable and check if the corresponding column/row shows non-pressed.
I do not recommend to cut near by the controller: it is tricky to remove the cable and use a new one. You may damage the components that are near by the connector.

Tell what is the result of the above checks.

I hope that this helps to get the FSSK002 working

[vivalarevolución]

Lost track of this project over the past couple weeks as other matters have grabbed my attention, but great to see things are continuing to move forward!

[XMIT]

hammelgammler, did you have any luck in getting your board to work?

idollar told me he is on vacation this month. We should see him again in March.

[shreebles]

idollar told me he is on vacation this month. We should see him again in March.

Well-deserved!

I bit the bullet and bought flippers from Ellipse...
Looks like I will need to bite the bullet again to buy an xwhatsit.

My wallet is crying. "Shh, next month will be better", I say.
Yeah right ;D

[Muirium]

How much is he selling his flippers for? Better not have cost you as much as a donor XT!

[hammelgammler]

hammelgammler, did you have any luck in getting your board to work?

idollar told me he is on vacation this month. We should see him again in March.

No I haven't as for now. I asked idollar what the problem might be, but it seems like I'm to stupid for it.