old hp 20s faster then the newer ones?

The Museum of HP Calculators

HP Forum Archive 09

old hp 20s faster then the newer ones?
Message #1 Posted by Mark on 31 Aug 2002, 7:54 a.m.

I have a older model hp 20s (the brown ones) Sn#id94500666 and a newer black model with the purple and green numbers Sn#id03102813. If I run a simple program lbl a + 1 gto A the older model is a lot faster. doing about 2 more operations a second.

Re: old hp 20s faster then the newer ones?
Message #2 Posted by Julián Miranda (Spain) on 31 Aug 2002, 8:21 a.m.,
in response to message #1 by Mark

Perhaps you could post the results of your test. With the same program you posted my 20S (3121S...)does 4176 in a minute. And a 21S (2936A...) does 4235.

Re: old hp 20s faster then the newer ones?
Message #3 Posted by mark on 31 Aug 2002, 10:44 a.m.,
in response to message #2 by Julián Miranda (Spain)

The older model hp 20s (the brown one)Sn#id94500666 got4748 The newer black model Sn#id03102813 got 4670
and i got an even older brown one sn#3420s81550 it got 4796

Re: old hp 20s faster then the newer ones?
Message #4 Posted by Raymond Hellstern (Germany) on 31 Aug 2002, 10:41 a.m.,
in response to message #1 by Mark

ACO redesigned some of the calculators in the last few years. Maybe the 20S has undergone the same process?
On the other hand, someone should try another of the 'new' 20S units, to prove that yours isn't a 'monday' machine;-)
Raymond

Re: old hp 20s faster then the newer ones?
Message #5 Posted by Randy Sloyer (US) on 31 Aug 2002, 1:04 p.m.,
in response to message #1 by Mark

Your actual program as displayed on my machines:
00- 01- 61 41 A 02- 75 03- 1 04- 51 41 A
My results by serial number for 1 minute:
20S 2931A USA 3690 counts 20S 3404S SING 4778 20S ID202 INDO 4574

All batteries where new, 4.66 to 4.70 volts per set. I tried the highest voltage pile in the old US unit - no change in speed.


Re: old hp 20s faster then the newer ones?
Message #6 Posted by Ellis Easley on 1 Sept 2002, 7:56 a.m.,
in response to message #5 by Randy Sloyer (US)

I calculated the ratio of your three results to their average and got 84.9%, 109.9% and 105.2%. My 20S, Indonesia ID836 gave 4749, which is 109.2% of that average.
Here's a guess: I'll bet the frequency is set by an external capacitor and a resistor inside the IC, or possibly an external capacitor and inductor. A resistor inside an IC would probably have enough variation to account for the range of speeds; I think the Voyagers had an external capacitor and inductor, and I know the Classics and Woodstocks did. But however they did it, and assuming they had a design that gave a fairly consistant speed, the easiest way for a sizable reduction in speed to happen by accident would be for someone at the factory to substitute a lower quality capacitor for the one that sets the frequency. You can get capacitors in a tolerance called "Guaranteed Minimum Value", which is sometimes stated "+100 -0%". These are intended as bypass capacitors for low frequencies. Compliance engineers warn against using them in signal filtering applications because they can be self-resonant at the frequencies you are trying to suppress.


Re: old hp 20s faster then the newer ones?
Message #7 Posted by Randy Sloyer(US) on 1 Sept 2002, 9:17 a.m.,
in response to message #6 by Ellis Easley

Hi Ellis!
The 20S is a one chip design, with one 10uf tantalum cap across the battery. No other external parts of any kind. Zero, notta, zip. Amazing. So the oscillator must be a RC type with everything on the silicon?
I thought perhaps battery voltage might affect the speed, but tests showed no change.
Just process variations in the chip?

Re: old hp 20s faster then the newer ones?
Message #8 Posted by Ellis Easley on 1 Sept 2002, 10:14 a.m.,
in response to message #7 by Randy Sloyer(US)

Well, I've always heard that the resistors on chips can vary a lot, like 50%. I don't know about capacitors on chips in this regard, except that the the capacitance value has to be kept small to reduce real estate. Maybe there is a way to design them so the process variation affects R and C oppositely, so their variation cancels in applications like an oscillator.
Speaking of real estate, here's a possibility: maybe the chip went through a die shrink and the capacitor came out different.
Or maybe, even though there is an internal RC for the oscillator, there might be a pin or pins to add external components to change the frequency. Maybe a PCB change resulted in more capaitance on that pin, which might just be connected to a test pad.

[ Return to Index | Top of Index ]

Go back to the main exhibit hall