Had the hankering to “build something” recently so I built out my second version of the new (Rev .2.0) H8 front panel. In the fall I had built one for display at VCF and ended up pimping that up with blue LEDs and a smoky lens for my “Big Blue” demo:
This time my goal was a little more modest: to replace the front panel on the “gaming” computer that I had set up for VCF. I wanted to keep the original red lens but add more tactile feedback to the keypad. I also decided to add LED backlighting for the keys. Here are the results:
to get a really “clicky” feel I chose to use the Kailh Mechanical Key Switches
which are advertised as Cherry MX Compatible but I found them to be a little different than the Cherry “blue” MX switch spec’ed in the front panel BOM
The Cherry switch has a recess designed to house a small 3mm LED but the Kailh has just a small rectangular opening (a difference I failed to notice in my haste to order parts 😊 ). I did experiment with sanding down a few 3mm LEDs before concluding that wasn’t the way to go. I ended up using surface mounted LEDs from Jameco:
These were a bit tedious to solder onto the board but with a bit of practice I got satisfactory results. I concluded they are fine for the soft “backlighting” effect I got but for true backlit keys the cherry MX with 3mm LEDs are likely to provide more satisfactory results.
I paired the keys with Adafruit’s “relegendable” keycaps
These have a different clearance pattern than the professionally built custom keycaps I used previously, but they worked fine with the Kailh switches and SMD LEDs. For the key labels I simply printed out an old Powerpoint slide I had from back when I built Rusty. Here’s what Rusty’s keypad looks like:
and here’s my new version based on the Adafruit relegendable keycaps:
it was particularly difficult to solder the key switches to be exactly square to the board, so a few seem to tilt one direction or another. If I get ambitious some day I may desolder a few and try to true them up.
The new keycaps do not extend nearly as far from the cabinet as the original Heath ones or the newer professional ones. When depressed the key is just slightly lower than the surrounding bezel. I don’t see this as an issue. The “clicky” feedback gives things a solid feel. With these new keys I never see any errors when keying in a program from the front panel (as opposed to the original Heath Stackpole switches where you had to constantly watch what was being keyed in).
I guess the only other change of note was the use of 3mm LEDs for the 5 status lights on the left of the panel (vs. the 5mm ones originally spec’ed by Heath). For the 7-segment LEDs I used these off eBay:
(by the way, here are the Blue ones: https://www.ebay.com/itm/262002126325 but don’t expect these to work well with the original Heath red lens)
Not sure how many of you out there are building out this new board but just thought I’d share my experience. Details on Norberto’s site
order boards from Todd Goodman:
One other thing to note: the board silkscreen and BOM talk about using a small 8 ohm speaker to replace the original Heath one. I found that didn’t work. The heath speaker had a higher impedance (70 ohm?) and that matters for this application.
As always, questions, comments, suggestions, feedback welcome…
Ah. Yes. the alignment pins. That’s what’s missing! Thanks.
p.s. one more interesting detour in this project concerned the 74LS123. I was using the part shown below, which I think is one I had in my parts drawer so I’m not sure from whence it came:
but it looks suspiciously unlike official TI packaging, no dot or indent to indicate pin 1. In retrospect this certainly looks like a potential counterfeit chip! (?)
So when I first powered up the new panel the keypad was functional but the 7-segment LEDs were not refreshing. Following the troubleshooting flowchart in the H8 operations manual led me to replace the 123 chip. I didn’t have an LS version but I did have a 74123 so I used that. that brought the display alive, but I was still seeing odd artifacts. Digits that should have been dark were partially lit.
in the “lesson not learned” department I should have remembered this discussion:
I was able to steal an actual 74LS123 from an unused board and that got everything working fine.
Note to self: add some 74LS123s to my next Jameco/Mouser/Digikey shopping list!
Yes, the analog characteristics of the 74123 are very different
from the 74LS123. They are not interchangeable. Not sure about
that suspect part, but it certainly does not look like something
TI would allow out the door.
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/153e01d7fdaf%2440e5dff0%24c2b19fd0%24%40gmail.com.
That's really a "problem" with this different 74123 specifications, I had this issue also in other systems like an Apple-1 replica and Altair MPU boards. Do you have an advice how to re-calculate the R and C? Is there a schematic for a small test cirquit to test this 74123 in a "meaningful" way?
The datasheets for the parts usually have a formula for
calculating the delay. You'll see they are different for 74123 and
74LS123. If you're actually going to modify the PCB components
rather than use the part it was designed for, you might need to
calculate the TC for the part intended in the circuit, then use
the other part's formula to determine R/C values that produce the
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/b8e17ccf-453c-45e6-af66-d25df10dd52fn%40googlegroups.com.