[TI-H] TI-89 HW2 overclocking


[Prev][Next][Index][Thread]

[TI-H] TI-89 HW2 overclocking



Hi, I've found several pages on TI-89 overclocking. This one in particular 
has a table of values of capacitance vs. MHz:

http://translate.google.com/translate?hl=en&sl=fr&u=http://planete89.free.fr/hardware.php3&prev=/search%3Fq%3Dti- 
89%2Boverclock%2Bc10%26num%3D50%26hl%3Den%26lr%3D%26ie%3DUTF-8%26oe%3Dutf- 
8%26sa%3DG

The original in French is:
http://planete89.free.fr/hardware.php3

They are replacing C10, so they're overclocking a HW1 calculator.

Here is the table: (monospaced)

Capacity		Time					Frequency
2.7pF			do not go				do not go
3.3pF			(some bugs) - 7s			22.9MHz
5.6pF			(of small the bugs) - 7.5s	21.3MHz
10pF			8s					20MHz
12pF			9s					17.8MHz
22pF			11s					14.5MHz
normal		15s					10.7MHz
100pF			27s					5.9MHz
120pF			30s					5.3MHz
390pF			205s					0.8MHz

where "Time" is the time to complete the benchmark of graphing z=sin(x) 
+cos(y) and Frequency is
measured by TiBench 1.0.

I'm wondering about the relevancy of this information to overclocking an 
HW2 calc (except replacing
C4 instead of C10). My TI-89 does the z=sin(x)+cos(y) graph in 11s and 
TiBench 1.5 says 12.6MHz,
implying that this HW1 data can't be directly compared to a HW2 calculator.

So, my question is: have enough people overclocked their HW2 TI-89 on this 
mailing list to construct
this same data? (obviously differences in battery voltage will cause some 
variance in the statistics)
Also, I haven't seen any other pages that recommend using smaller than a 
22pF cap.

(And I've got some ideas for getting to the "do not go" speeds involving 
some microsoldery and some
55ns SRAM -- anyone know how fast the mc68sec000pb16 can be clock-chipped 
with fast RAM? and if I'm
not mistaken, according to Motorola's datasheet, that's a 16MHz chip, as 
they don't make a 12MHz
version.)

(Oh, and one last postscript to the inevitable "but why would you want to 
overclock a calculator"
questions: ever tried solving a nasty definite integral from 0 to x = some 
value, for x? or taken
a leslie matrix to around the 40th or 50th power? or graphing? hasn't TI 
heard of an optimizing C
compiler? </rant>)

-- 
B.A. Baracus
The A-Team



Follow-Ups: