16F1827 table not behaving as expected

[Bruce]

For LOOKUP2 your table is limited to 85 list items unless you're using an 18F part..

[J. Mark Wolf]

For LOOKUP2 your table is limited to 85 list items unless you're using an 18F part..

Hi Bruce

The compiler was not balking at the table size, and I confirmed with Darrel that the table size limit has been expanded, and that the manual has not yet been updated to reflect this.

[Bruce]

Hi Mark,

You can run a simple test to see what the limit is with the table attached.

Setup a loop to call your LOOKUP2 routine with an index value from 0 to 99.

Where does it fail?

[Darrel Taylor]

Hi J.,

When you asked about LOOKUP2 previously, I had to ask.
The answer I got was 256.

That was only half right.
The compiler executable was changed ... but the macro's were not.
So while the compiler won't complain up to 256 ... the macros only work up to 85. (256/3)

Charles says the executable will be changed back to 85.
I think the macros should be changed to work with 256, but I'd have to do it in that case.
Unless Bruce beats me to it.

Are those EXT tables looking better yet?

[Bruce]

Unless Bruce beats me to it.

Nice try, but I'll wait for the patch....

If you could yank Charles off those dang race cars, I bet he could fix it ASAP...;o)

[J. Mark Wolf]

Hi Mark,

You can run a simple test to see what the limit is with the table attached.

Setup a loop to call your LOOKUP2 routine with an index value from 0 to 99.

Where does it fail?

Hi Bruce

I thought the same thing driving home from work.

I tried it and it does indeed go "off in the weeds" after reading 85 elements.

So I take it that mere mortals cannot tweak the macros?

I can decimate my table to remove the leading zero "placekeepers" and redundant upper section to get below 85 elements, but then I'll have to play some other games to assert some artificial offsets to enter the table correctly under differing circumstances.

This is a case where the compiler/macros really should be tweaked to allow larger tables, to allow things like thermistor tables, etc.

Maybe small table limits made sense back when the devices had little code space, but that's not as true anymore.

[Bruce]

Hi Mark,

You can tweak the macros if you want, but I don't recommend it unless you're as good as Darrel...

The 85 limit is because the compiler generates 3 table entries per list value. 3 x 85 = 255, which is the largest number that fits in the W register.

It would be easier to just use three lookup2 tables with a select case to jump to the appropriate lookup2 routine.

[J. Mark Wolf]

I tried breaking up my 254-element table into four 64-element tables as mark_s suggested in another thread, but it's giving more trouble and now "Crossing Page Bounday errors".

It behave as though PBP only allows one LOOKUP2 table construct per program. The DacDwell variable returns with the proper value only if the Measurement variable is contained within the original "called" table.

I've inserted some serial traps to show if the index is working properly between the tables (it is), and to show if the table result is working properly (it isn't)

You can see on the right side of the JPG the serial messages showing the Measurement variable changing to the appropriate table as the Measurement variable increments/decrements, but the DacDwell variable shows zero when it references the second table.

When the Measurement variable increments to 64 or above, it would normally advance to the next table and return the DacDwell values found there, but as youcan see it returns zeros.

I only have room for DwellTable 3 & 4 in the screenshot (DwellTable 1 & 2 contain nothing but zeros for padding), but if I force the Measurement variable to DwellTable to 3 or 4, it will return the proper values until it crosses to the next table construct

So far I'm damned if I do and damned if I don't with respect to these tables.

I'm staring to consider one of the newer "K" parts, but I'm afraid I'll run into compiler deficiencies there as well.

Still not sure if someone is going to fix the macros to match the executable, or drop back to a 85-element limit.

Can anyone offer a way forward?

JPG and code below:




[CODE]'
DwellTable3:
'if measurement > 191 then goto DwellTable4
lookup2 Measurement,[$5161,$4c48,$47ce,$43d1,$403f,$3d09,$3a21,$377d,_ '128 - 135 1.5pps - 2.2pps
$3513,$32dd,$30d4,$2ef3,$2d36,$2b99,$2a18,$28b1,_ '136 - 143 2.3pps - 3.0pps
$2761,$2626,$24fe,$23e7,$22e1,$21e9,$20fe,$2020,_ '144 - 151 3.1pps - 3.8pps
$1f4d,$1e85,$1dc6,$1d10,$1c63,$1bbe,$1b20,$1a89,_ '152 - 159 3.9pps - 4.6pps
$19f9,$196e,$18ea,$186a,$17ef,$177a,$1708,$169b,_ '160 - 167 4.7pps - 5.4pps
$1632,$15cc,$156a,$150c,$14b1,$1458,$1403,$13b0,_ '168 - 175 5.5pps - 6.2pps
$1360,$1313,$12c8,$127f,$1238,$11f4,$11b1,$1170,_ '176 - 183 6.3pps - 7.0pps
$1131,$10f4,$10b9,$107f,$1047,$1010,$0fda,$0fa6],dacdwell3 '184 - 191 7.1pps - 7.8pps
serout2 portb.5,396,["Measurement = ", dec measurement, 13, 10]
serout2 portb.5,396,["DacDwell3 = ", isHEX4 DacDwell3, 13, 10]
serout2 portb.5,396,[13, 10] '
pause 1000
return
'goto Endtable
'
DwellTable4:
lookup2 Measurement,[$0f74,$0f42,$0f12,$0ee3,$0000,$0000,$0000,$0000,_ '192 - 199 7.9pps - 8.2pps & 900Hz w/ no mod
$0000,$0000,$0000,$0000,$0000,$0000,$5161,$4c48,_ '200 - 207 900Hz w/no mod & 1.5pps - 1.6pps
$47ce,$43d1,$403f,$3d09,$3a21,$377d,$3513,$32dd,_ '208 - 215 1.7pps - 2.5pps
$30d4,$2ef3,$2d36,$2b99,$2a18,$28b1,$2761,$2626,_ '216 - 223 2.6pps - 3.3pps
$24fe,$23e7,$22e1,$21e9,$20fe,$2020,$1f4d,$1e85,_ '224 - 231 3.4pps - 4.1pps
$1dc6,$1d10,$1c63,$1bbe,$1b20,$1a89,$19f9,$196e,_ '232 - 239 4.2pps - 4.9pps
$18ea,$186a,$17ef,$177a,$1708,$169b,$1632,$15cc,_ '240 - 247 5.0pps - 5.7pps
$156a,$150c,$14b1,$1458,$1403,$13b0],dacdwell4 '248 - 253 5.8pps - 6.3pps
serout2 portb.5,396,["Measurement = ", dec measurement, 13, 10]
serout2 portb.5,396,["DacDwell4 = ", isHEX4 DacDwell4, 13, 10]
serout2 portb.5,396,[13, 10] '
pause 1000
return'[CODE]

[mark_s]

Hi Mark,
Those error messages are okay.

If I am reading your code correctly you are not off setting the index value. Each lookup2 table has a possible index address of 0 - 85. In your case you need to offset each table by 64

Something like this

If index > 64 then table2
if index > 128 then table3
if index > 192 then table4

Table1: 'No offset

Lookup2 index[1-64]var
return

Table2:

Index = index -64 'offset
Lookup2 index[65 -128]var
return

Table3:

Index = index -128 'offset
Lookup2 index[129-192]var
return

Table4:

Index = index -192 'offset
Lookup2 index[193-256]var
return

[Bruce]

Something like this should work so you don't alter your index variable.

main:
  for Measurement = 0 to 17
      gosub DwellTable
  next Measurement
  goto main
 
DwellTable:
  IF Measurement > 5 THEN Dwel2
  lookup2 Measurement,[$0110,$0120,$0130,$0140,$0150,$0160],dacdwell
  RETURN
 
Dwel2:
  IF Measurement > 11 THEN Dwel3
  lookup2 Measurement-6,[$0111,$0121,$0131,$0141,$0151,$0161],dacdwell
  RETURN
 
Dwel3:
  lookup2 Measurement-12,[$0112,$0122,$0132,$0142,$0152,$0162],dacdwell
  RETURN

[J. Mark Wolf]

Hi Mark,
Those error messages are okay.

If I am reading your code correctly you are not off setting the index value. Each lookup2 table has a possible index address of 0 - 85. In your case you need to offset each table by 64

Something like this

If index > 64 then table2
if index > 128 then table3
if index > 192 then table4

Table1: 'No offset

Lookup2 index[1-64]var
return

Table2:

Index = index -64 'offset
Lookup2 index[65 -128]var
return

Exactly right Mark! I was forgetting to offset the Measurement variable back to zero relative to each table.

The Page Boundary errors went away too!

Thanks for looking!

[J. Mark Wolf]

Hi Mark,
Those error messages are okay.

If I am reading your code correctly you are not off setting the index value. Each lookup2 table has a possible index address of 0 - 85. In your case you need to offset each table by 64

Something like this

If index > 64 then table2
if index > 128 then table3
if index > 192 then table4

Table1: 'No offset

Lookup2 index[1-64]var
return

Table2:

Index = index -64 'offset
Lookup2 index[65 -128]var
return

Exactly right Mark! I was forgetting to offset the Measurement variable back to zero relative to each table.

The Page Boundary errors went away too!

Thanks for the tips. I've been working on this for hours!

I think my vector is still off by one count, but substantially working code below:

'*******************************************************************************
' 
DwellTable1:
if measurement > 63 then goto DwellTable2 
'
Vector = measurement
'
lookup2 Vector,     [$0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '0 - 7       
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '8 - 15 
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '16 - 23
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '24 - 31 
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '32 - 39
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '40 - 47
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '48 - 55
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000],dacdwell '56 - 63
goto Endtable
'
DwellTable2:
if measurement > 127 then goto DwellTable3 
'
Vector = measurement - 64
'
lookup2 Vector,     [$0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '64 - 71
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '72 - 79
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '80 - 87
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '88 - 95
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '96 - 103
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '104 - 111
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000,_ '112 - 119
                     $0000,$0000,$0000,$0000,$0000,$0000,$0000,$0000],dacdwell '120 - 127
goto Endtable
'
DwellTable3:
if measurement > 191 then goto DwellTable4
'
Vector = measurement - 128
'
lookup2 Vector,     [$5161,$4c48,$47ce,$43d1,$403f,$3d09,$3a21,$377d,_ '128 - 135  1.5pps - 2.2pps
                     $3513,$32dd,$30d4,$2ef3,$2d36,$2b99,$2a18,$28b1,_ '136 - 143  2.3pps - 3.0pps
                     $2761,$2626,$24fe,$23e7,$22e1,$21e9,$20fe,$2020,_ '144 - 151  3.1pps - 3.8pps
                     $1f4d,$1e85,$1dc6,$1d10,$1c63,$1bbe,$1b20,$1a89,_ '152 - 159  3.9pps - 4.6pps
                     $19f9,$196e,$18ea,$186a,$17ef,$177a,$1708,$169b,_ '160 - 167  4.7pps - 5.4pps
                     $1632,$15cc,$156a,$150c,$14b1,$1458,$1403,$13b0,_ '168 - 175  5.5pps - 6.2pps
                     $1360,$1313,$12c8,$127f,$1238,$11f4,$11b1,$1170,_ '176 - 183  6.3pps - 7.0pps
                     $1131,$10f4,$10b9,$107f,$1047,$1010,$0fda,$0fa6],dacdwell '184 - 191  7.1pps - 7.8pps
goto Endtable
'
DwellTable4:
'
Vector = measurement - 192
'
lookup2 vector,     [$0f74,$0f42,$0f12,$0ee3,$0000,$0000,$0000,$0000,_ '192 - 199  7.9pps - 8.2pps & 900Hz w/ no mod
                     $0000,$0000,$0000,$0000,$0000,$0000,$5161,$4c48,_ '200 - 207  900Hz w/no mod & 1.5pps - 1.6pps
                     $47ce,$43d1,$403f,$3d09,$3a21,$377d,$3513,$32dd,_ '208 - 215  1.7pps - 2.5pps
                     $30d4,$2ef3,$2d36,$2b99,$2a18,$28b1,$2761,$2626,_ '216 - 223  2.6pps - 3.3pps
                     $24fe,$23e7,$22e1,$21e9,$20fe,$2020,$1f4d,$1e85,_ '224 - 231  3.4pps - 4.1pps
                     $1dc6,$1d10,$1c63,$1bbe,$1b20,$1a89,$19f9,$196e,_ '232 - 239  4.2pps - 4.9pps
                     $18ea,$186a,$17ef,$177a,$1708,$169b,$1632,$15cc,_ '240 - 247  5.0pps - 5.7pps
                     $156a,$150c,$14b1,$1458,$1403,$13b0],dacdwell     '248 - 253  5.8pps - 6.3pps             
Endtable:
'
return
'
end