White screen but no problem with Minerva?

[martyn_hill]

Thanks Pr0f !

Now I think about it, my second question about tying the output is nonsensical - thanks for being non-judgmental in your reply

[Peter]

Unlike regular TTL and derived families, which kind of default to high, CMOS is really floating in a VERY high impedance state so all sorts of electromagnetic garbage can influence them resulting in unused gates processing fake signals, increasing power drain and regenerating said EM garbage.

Good to read from you Nasta. I don't want to encourage open inputs at all, but here's a little further remark just for fun. On several occasions I noticed less tendency of HCT to float into the linear mode of operation / oscillation when inputs are left open than HC. There is a subtle difference that could explain this. HCT not only has an enlarged N-channel transistor at the input and a level shifter diode to lower the switching threshold to TTL. It also contains an auxiliary pull-up transistor (with its gate at the output of the second transistor stage) and that seems to have a small stabilizing effect. Which is a poor excuse for the fact that the Q60 contains one HCT where I made the same mistake.

[Peter]

Is a direct connection to the relevant powerline preferred over using a pull-up/down resistor?

No. Both is equally good for HC/HCT. For LSTTL a resistor around 1K to VCC is best.

[Nasta]

Unlike regular TTL and derived families, which kind of default to high, CMOS is really floating in a VERY high impedance state so all sorts of electromagnetic garbage can influence them resulting in unused gates processing fake signals, increasing power drain and regenerating said EM garbage.

Good to read from you Nasta. I don't want to encourage open inputs at all, but here's a little further remark just for fun. On several occasions I noticed less tendency of HCT to float into the linear mode of operation / oscillation when inputs are left open than HC. There is a subtle difference that could explain this. HCT not only has an enlarged N-channel transistor at the input and a level shifter diode to lower the switching threshold to TTL. It also contains an auxiliary pull-up transistor (with its gate at the output of the second transistor stage) and that seems to have a small stabilizing effect. Which is a poor excuse for the fact that the Q60 contains one HCT where I made the same mistake.

Same thing here, I also had some luck when I had to service a LSTTL based circuit but had only HCT chips, and missed one with floating inputs - it worked fine except when I got real close to the pins with my fingers

[mk79]

Same thing here, I also had some luck when I had to service a LSTTL based circuit but had only HCT chips, and missed one with floating inputs - it worked fine except when I got real close to the pins with my fingers

For QL-VGA I once switched to a larger memory chip for production reason, the second halve of the memory is unused. When I tested this a few weeks later the hardware worked pretty well but very sporadically had strange defects in the picture. Suffice to say I should have altered the FPGA code to bind the additional address line to ground right then when I altered the PCB, after all the time it took me a while to get what's going on...

[Popopo]

I think this is the original SuperBASIC program which that page refers to:

Code: Select all

100 REMark 1990 Dominic Brown
110 INK#1,7:PAPER#1,2
120 INK#0,4:PAPER#0,0
130 WINDOW#1,480,200,32,16
140 WINDOW#0,480,40,32,216
150 BORDER#1;1,0,7
160 FOR i=0 TO 1:CSIZE#i,1,0
170 DIM x(16),y(16),ram_chip(16)
180 DIM faulty_address$(20),go_again$(2)
190 RESTORE
200 true = 1
210 false = 0
220 FOR i = 1 TO 16:READ x(i),y(i)
230 REPeat MAIN_LOOP
240   PAPER 0
250   CLS:CLS#0
260   REPeat get_address_loop
270     INPUT#0,'Enter Address reported as faulty ';faulty_address$
280     IF faulty_address$<>''
290       faulty_address = DEC(faulty_address$)
300       IF faulty_address <> 1E9: EXIT get_address_loop
310       PR 2,'Invalid Character used in Hex try again!!!'
320     END IF
330   END REPeat get_address_loop
340   IF faulty_address > 262144
350     PR 7,'Faulty ram is in expansion memory!!!'
360   ELSE
370     d1$=get_data$("Enter first data value  >")
380     d2$=get_data$("Enter second data value >")
385     failmap=0
395     FOR i=1 TO 7 STEP 2:failmap=failmap||xorbyte(d1$(i TO i+1),d2$(i TO i+1))
450     hibank=8*(faulty_address>=196608)
455 PRINT #0;BIN$(failmap,8)
460     FOR i=1 TO 16:ram_chip(i) = false
470     mask%=128
480     FOR i=8 TO 1 STEP -1:IF failmap&&mask% THEN ram_chip(i+hibank) = true:bad_chip=i+hibank:END IF :mask%=mask%/2
490     FOR ic = 1 TO 16
500       IF ram_chip(ic) = true
510         block_colour = 2
520       ELSE
530         block_colour = 7
540       END IF
550       BLOCK 25,45,x(ic),y(ic),block_colour
560       CURSOR x(ic)+5,y(ic)+10
570       PRINT 'IC'
580       CURSOR x(ic)+5,y(ic)+20
590       PRINT ic
600     END FOR ic
610     PR 7,'The error has occured in IC'&bad_chip
620   END IF
630   PRINT#0,'Press Esc to Finish, any other key to continue '
640   go_again$=INKEY$(#0,-1)
650   IF go_again$ = CHR$(27): EXIT MAIN_LOOP
660 END REPeat MAIN_LOOP
670 STOP
680 :
690 DEFine FuNction get_data$(x$)
695   LOCal ans$
700   REPeat l
710     INPUT #0;(x$);ans$
740  REMark   IF res==1E9 THEN PRINT #0;'Invalid hex digit!':NEXT l
750     EXIT l
760   END REPeat l
770   ans$=FILL$('0',8)&ans$
780   ans$=ans$(LEN(ans$)-7 TO)
860   RETurn ans$
870 END DEFine
872 :
875 DEFine FuNction xorbyte(x$,y$)
877   RETurn DEC(x$)^^DEC(y$)
878 END DEFine
880 :
890 DEFine FuNction DEC(A$)
900   LOCal S,T,Q,valid_chr
910   LOCal valid_chr$(22)
920   T=0
930   valid_chr$ = '0123456789abcdef'
940   FOR Q=1 TO LEN(A$)
950     valid_chr=A$(Q) INSTR valid_chr$
960     IF valid_chr = false:EXIT Q
1000     T=T*16+valid_chr-1
1010   END FOR Q
1040   IF valid_chr = false:RETurn 1E9
1050   RETurn T
1060 END DEFine DEC
1070 :
1080 DEFine PROCedure PR(ink_colour,statement$)
1090   INK#0,ink_colour
1100   PRINT#0,statement$
1110   INK#0,4
1120 END DEFine PR
1130 DATA 90,40
1140 DATA 130,40
1150 DATA 170,40
1160 DATA 210,40
1170 DATA 250,40
1180 DATA 290,40
1190 DATA 330,40
1200 DATA 410,110
1210 DATA 90,110
1220 DATA 130,110
1230 DATA 170,110
1240 DATA 210,110
1250 DATA 250,110
1260 DATA 290,110
1270 DATA 330,110
1280 DATA 370,110

Thank you a lot!
I was looking for it to add some improvements.
Ie. Some QLs have RAMs 256KBits each, so the test cannot tell which one is the faulty one due to the change of size.
I wondered if it were possible to modify it and now I can try it out.