SBC6120 "Mini" Front Panel

[atreubig]

Hi Aug,

my SBC6120-RBC runs fine too when operated via the serial port. The problem is that the front panel stops updating when the count wraps from 7777 to 0. The SBC6120-RBC continues to count and seems to do a full count cycle from 0000 to 7777 when then the front panel updates resume.

To replicate the problem, power on the system, but don't type anything on the serial console. Then enter the following count program starting from address 0000 and extended address 0000 via the front panel switches using the DEP toggle. After setting the address back to 0000 verify it via the EXAM toggle. Finally set the address back to 0000 again and then select the AC display via the rotary switch and toggle CLEAR and then CONT to start the program. Observe how the count reaches 7777 and then there is a long pause after which the count resumes at 0000.

Here is the count program (the text after the ";" is the PDP-8 instruction mnemonic):

0000: 7001 ; IAC
0001: 2101 ; ISZ 0101
0002: 5001 ; JMP 0001
0003: 5000 ; JMP 0000

If while the count is paused, you type a single character for example SPACE, the front display resumes and won't ever stop again.

If you enter and start the same program via the serial console, the count increments from 0000 to 7777 and then immediately resumes incrementing from 0000 as you would expect.

Could you please explain why you think that the PAL manufacturer (Lattice vs Atmel) may be the cause of the problem?

I am using the Atmel parts: ATF16V8BQL-15PU and ATF22V10CQZ-20PU

I used the JED files in directory "sbc6120-rbc-edition-archive-10may2018\PLD" from the SBC6120-RBC site to program the PLDs. After programming the PLDs verified OK.

I would appreciate any insights you have.

I also have an original FP6120 front panel from Bob Armstrong with an SBC6120-RBC processor board and it has the same pausing problem.
I believe the original SBC6120 processor board from Bob Armstrong works fine with either Chris's or original front panel boards.

Thanks
Tom

The 22 Lattice pals are slower. 25 vs 20. Somewhat a stab in the dark. Also the SBC uses different memory. It will be a week or so before all parts arrive. Some are still stumbling thru the Aliexpress shipping maze.

Aug
AG5AT

 

[thunter0512]

I hope you can reproduce the problem I am seeing.

 

[thunter0512]

I have swapped the 3 PALs between the working original SBC6120 and the pausing SBC6120-RBC. The hang problem stays with the SBC6120-RBC despite the PALs from the working unit. The original SBC6120 does not exhibit the front-panel pause.
The pause problem is not caused by the PALs.

 

[thunter0512]

I swapped the HD6120 CPUs between working original SBC6120 and the pausing SBC61120-RBC. The problem stayed with the SBC6120-RBC.

Next I swapped the UARTs between the two units. It turns out that the problem follows the UART.

The UART causing the pauses is this one:



I have 3 of these and tried all of them. They work fine as a UART, but they all exhibit this pausing problem when used with either Spare Time Gizmos FP6120 or Hideehoo's front panel.

I subsequently tried an Intersil and also a Harris version of this chip and they works perfectly as a UART, but also these dont cause the pause of the front panel updating.

 

[vrs42]

Even if that's some shady counterfeit part trying to phone home (which makes no sense), I don't get how that could prevent the front panel from updating.

 

[shirsch]

I'm hardly an expert on counterfeits but something about the labeling on that chip looks off. Characters do not exactly line up on a common baseline, for starters.

 

[thunter0512]

Even if that's some shady counterfeit part trying to phone home (which makes no sense), I don't get how that could prevent the front panel from updating.

The front panel updates are mostly done in software. Each update is triggered by the approx 30 Hz trap generated by the 7555 on the front panel which results in CPREQ (Control Panel REQuest) interrupts. This CPREQ interrupt is also generated by a framing error in the UART - i.e. when pressing the BREAK key on the console terminal. Without digging into the BTS6120 software I think it is conceivable that some software interaction with the UART occurs when CPREQ is handled which reveals some incompatibility between various versions of the UART.

---

I found another bad UART which does NOT work - this seems to be a different manufacturer, but same designation "CDP6402CE":




The two good working UARTs which do not cause the front panel to freeze are:





I hope this helps others building Hideehoo's front panel. Stick with Intersil and Harris versions of the UART on the SBC6120-RBC board.

 

[antiquekid3]

If you can take some acetone to the tops of the suspect packages and see if anything comes off onto a paper towel, that is a good sign of a counterfeit. If you can get photos of the bottom side as well, that may help determine it, too.

 

[thunter0512]

If you can take some acetone to the tops of the suspect packages and see if anything comes off onto a paper towel, that is a good sign of a counterfeit. If you can get photos of the bottom side as well, that may help determine it, too.

I doubt that the problematic UARTs are counterfeit. More likely they have been independently implemented by other manufacturers and are meant to be compatible, but are not quite a 100% match possibly because of some under-documented feature of the original. They work fine as UARTs. Only when you fit a FP6120 and run the 4 word count program will you see the problem.

Nothing came off when I tried rubbing acetone on 3 of them.

Here are photos of the backs:



Other than the word "Singapore" none of it means anything to me.