Happy day - Triple speed 15C - after 22 years



#17

Hello.

On a BBS, I've got an information that some old HP calculator will be sold at a radio parts shop in Akihabara Tokyo, last week.
When I visit the shop, which is very next to the shop I bought 15C twenty two years ago, I found several old HP-15Cs between HP-35 and HP-65 (It still work!), and met a young man who had posted the information.
After an hour of negotiation talk with the proprietor, I purchased an old HP 15C at reasonable price, and brought it back home.

Since I got a spare 15C, I made speed-up modification to it while doing cleaning and maintenance.

HP 15C - LC resonator replacement for triple speed 15C

It seems to work well until the battery voltage drops under 3.0V at triple speed, with double current drain of 2.5mA(at 4.7V, while running).

After all, I got 15C that is working at triple speed, and made the new young friend who is trying to use good-old-16C from now. It was a happy day.


#18

Wow! you have a spare 15C to experiment with overclocking?

cool,very informative, thanks

#19

Nice work!

Three years ago I made a similar experiment:

http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/archv016.cgi?read=99306

I didn't use any inductor because battery consumption would increase, but perhaps I just didn't have proper ones. I found double speed to be most realiable. I don't have it anymore (traded it with a 16C and have received no complain so far :-) I did the same procedure on one of my three units (easily done as the SMD capacitor I had to replace was on the PCB visible side. It's ok after two and a half years of very moderate use, except that the batteries went dead last week...

Regards,

Gerson.


#20

Thanks Gerson.

I've ever read your article before, and it was very informative reference to me.

Regards,
Lyuka.

#21

This post prompted me to ask a question I have been wondering - does anyone know of a hardware speed-up for the 48sx? I have tried searching the archives w/o success so far.

I think I did read about a software speed-up, but I cannot remember the name of it, nor find it in the archives. I understand SpeedUI only works on the gx.

Anyone? Thanks.


#22

Hello Martin,

48 S Series overclock is possible just changing crystal oscillator.

32kHz is used and must be changed to 64kHz.

clock time will also be modified.

regards


#23

Thanks, Nacho. Do you know if there was ever a thread on how to do this?

#24

Quote:
I think I did read about a software speed-up, but I cannot remember the name of it, nor find it in the archives. I understand SpeedUI only works on the gx.

I remembered it - Meta Kernel. Anyone have experience with this? Is it really faster than the native 48sx user interface? Once you load it onto a 128k card, and leave the card in, can you swap out cards in the other slot, without disturbing the Meta Kernel interface?
#25

Quote:
I understand SpeedUI only works on the gx.

I am not sure but that might be down to memory requirements rather than specific G features. You certainly wont fit it on an SX without a memory card and what I consider the basic configuration takes about 60K on my GX. Worth having though without any doubt.

Mark

#26

Lyuka --

Nice work and photos!

I'd like to make a few comments:

1. Your modifications -- replacement of two capacitors and one inductor, plus addition of a parallel capacitor -- were more extensive than the simple parallel-inductor addition described in Ken Sumrall's writeup hosted by Eric Smith.

2. Was the purpose of lowering the oscillator capacitance values to achieve further increases in resonant frequency while maintaining an L/C ratio of the oscillator closer to its unmodified value? The better-matched L and C components would help to retain more energy-exchange capability and resulting magnitude of the oscillator voltage.

3. Was the bypass capacitor replaced with a higher-capacitance part in order to help compensate for the other changes?

4. Both of your HP-15C's are of the revised (1986-89) Voyager circuit design, with most or all of these parts atop the circuit board that allows easy replacement. The modifications you made appear not to be suited to the original (1981-85) circuit design:

5. Original-design Voyager units lack the same external connections for the bypass capacitor shown on the right side of the photos, as well as for the lower oscillator capacitor on the left side. Also, the oscillator capacitor is of a large "can" style residing in the cutaway of the circuit board, below the battery (shown on your older 1986 calculator). The oscillator inductor is very difficult to access, residing below the circuit board.

6. Your photos did not include one of the unmodified older (1986) calculator. I assume that its oscillator capacitor was can-style like those of the original design. (I have a 1985 unit that has the cutaway as well as the same type 1LM2-0001 CPU).


I based a speedup of my 1985 original-circuit HP-15C on Ken Sumrall's approach. At first, I simply connected a 22-mH inductor in parallel, to give a 3.3x increase in speed. However, the calculator later malfunctioned at healthy battery voltage, so I removed the extra inductor, which restored proper operation. Next time, I will install a 33-mH part for a 2.75x increase in speed.

-- KS

Edited: 18 Oct 2009, 9:58 a.m.


#27

Hi Karl,

Thank you for your comment.
I'm sorry, but I forgot to add your 'successful HP-11C speed up' article to the reference.

I know that some* early model of the HP 15C has one aluminum electrolytic capacitor (which has poor characteristics) for bypassing.
And later model has improved PCB design, and has one solid tantalum (which has good characteristics) bypass capacitor close to the CPU.

Though the solid tantalum capacitor has good characteristics, to increase the operating frequency up to 2.8MHz, it should better have one or two relatively high capacitance ceramic capacitor close to the CPU for mainly stability and EMI reason.
(100uF/6V/MLCC that I used is over-quality and expensive, 47uF/10V/Ta with one 100nF/16V/MLCC in parallel may be enough.)

The series resistance of the inductor and L/C ratio are important for LC resonator. These parts should better be chosen to keep its maximum Q(quality factor) greater than several tens at oscillation frequency for stable oscillation.

regards,
Lyuka

some* - some model uses dip solid tantalum (good characteristics)

Edited: 19 Oct 2009, 6:09 a.m. after one or more responses were posted


#28

Hi Lyuka,

Here is my current sped-up 15C, SN 2740B30986:

I only replaced the SMD capacitator that was soldered on the lower pads with the ceramic 27 pF capacitor you see. Considering the original 180 uH inductor, the speed-up factor I have is 2.5, not 2 as I thought (I will check this later). Have you measured the original current?

Regards,

Gerson.


#29

Hi, Gerson

Yes, the original current have been measured as noted in the URL shown above.

Quote:
Power supply current (Ta=26¡ëC, Vbattery=4.72V(SR44¡ß3), fLC=884kHz(original), Agilent 34401A)

Irun = 1.2mA, Iidle=7¦ÌA, Istandby<0.1¦ÌA


Your experience that 33pF gives exact x2 factor makes sense, since,



Factor_speedup = sqrt((180pF + C_inductor + C_line) / (33pF + C_inductor + C_line))

where

C_inductor is the parallel capacitance of the inductor, assumably 4pF from inductor's self resonant frequency.

C_line is the line capacitance, 12pF by your measurement.


Factor_speedup = sqrt((180 + 4 + 12) / (33 + 4 + 12)) = sqrt(196 / 49) = sqrt(4) = 2 (Great! :-)

Regards,
Lyuka


#30

Quote:
Yes, the original current have been measured as noted in the URL shown above.

Thanks for the nice documentation! I had tried to go beyond 2.5x before but did not succeed. Your choice of components to achieve maximum quality factor is interesting. I will try your mod later.

Regards,

Gerson.


#31

Hi Gerson,

The bypassing capacitor is getting more important at higher frequencies.

At least one 100nF/MLCC should better be added in parallel with 22u/10V/Ta capacitor.

Regards,
Lyuka

#32

Hi, there.

I did further experimental speed-up modification on a 15C.

Experimental modification for quad-speed 15C

It seems that it's stable at 4X speed with fresh SR44 batteries, so far.

Regards,

Lyuka

Edited: 23 Oct 2009, 8:06 a.m.


Possibly Related Threads…
Thread Author Replies Views Last Post
  An amazing day: Giving a talk at HP about their calculators Geir Isene 9 4,683 12-16-2013, 06:14 PM
Last Post: aurelio
  HP Prime: Quite Happy Helge Gabert 13 5,513 12-10-2013, 08:19 AM
Last Post: Eddie W. Shore
  48G vs 49G+ User RPL Speed Comparison John Colvin 7 2,706 11-16-2013, 10:07 PM
Last Post: Han
  HHC 2013 Day 2 Highlights Eddie W. Shore 6 2,655 09-23-2013, 04:03 PM
Last Post: Kimberly Thompson
  HHC 2013: Day 1 Highlights Eddie W. Shore 28 8,131 09-23-2013, 03:22 PM
Last Post: Brad Barton
  WP-34S: Speed of y^x Marcel Samek 1 1,187 09-14-2013, 07:31 PM
Last Post: Paul Dale
  WP-34S function execution speed ? Gene Wright 4 1,782 09-04-2013, 05:40 PM
Last Post: Paul Dale
  HP 11C Batteries have lasted 22 years nink 3 1,508 06-23-2013, 06:51 PM
Last Post: Les Koller
  Happy Mother's Day! Eddie W. Shore 1 1,237 05-12-2013, 11:35 AM
Last Post: Walter B
  happy fibonacci day 5/8/13 Allen 8 2,955 05-09-2013, 01:48 AM
Last Post: Gerson W. Barbosa

Forum Jump: