Spotlight: HP 35 Calculator (1972)
Here it is, the first every handheld scientific calculator: the Hewlett Packard HP 35 from 1972!
I got it at the HHC 2026 Conference earlier this month, and I thank you, David Ramsey, Gene Wright, Ron Grant, and Mark Sims. (Rest in peace, Mr. Sims)
The 35 came with a working Ni-CAD battery pack and the original plug in charger. The keyboard and the leather case are in excellent condition. I also have the little booklet.
The HP 35 is Version 4, and in this version, the functions are labeled on top of the keys instead of above them. All the labels are in capital letters. In earlier versions, some of the labels were in lower case (see the Functions of the 35 section).
I also have a 3-AAA Alkaline battery adapter from Waterhosko that is for HP classic calculator series (35, 45, 55, 80, 65, 67).
I feel like I have finally arrived.
Quick Facts
Model: HP 35
Company: Hewlett Packard
Type: RPN Scientific
Memory: 1 (store, recall)
Years in Production: 1972 to 1975
Display: 10 digits
Original Price: $ 395.00
Hardware, Simulators, and Emulators
The hardware may not be easy to come by, although there are still some HP 35 calculators available through auction sites such as eBay.
There are plenty of simulators and emulators for the HP 35. For example, the pockemul app, available for both Android, Apple iOS, Windows, and mac0S, contains many past calculators and pocket computers. One of the calculators that are offered at no cost is the HP 35 (the version shown is an earlier version (not red dot) where the functions are labeled above the keys).
Site: https://pockemul.com/
A Javascript simulator, hosted by the MoHPC (Museum of HP Calculators) site, is written by Neil Fraser. The simulator may also be downloaded: https://www.hpmuseum.org/simulate/hp35sim/hp35sim.htm
Functions on the HP 35
The HP 35 has one shift key, and that is only for the trigonometric functions.
[ x^y ]: power key, note the order of the variables
[ log ]: common logarithm (log10)
[ ln ]: natural logarithm, base e
[ e^x ]: exponential function
[ √x ]: principal square root
[ sin ]: sine function, angle is in degrees
[ arc ] [ sin ]: arcsine function, returns the angle in degrees
[ cos ]: cosine function, angle is in degrees
[ arc ] [ cos ]: arccosine function, returns the angle in degrees
[ tan ]: tangent function, angle is in degrees
[ arc ] [ tan ]: arctangent function, returns the angle in degrees
[ clr ]: sets all four stacks (X, Y, Z, T) and the memory register to 0
[ 1/x ]: reciprocal
[ x<>y ]: swap the contents of X and Y
[ R↓ ]: roll down the stack (T → Z, Z → Y, Y → X, X → T)
[ STO ]: store the contents of the X stack into memory
[ RCL ]: recall the contents of memory to the X stack, pushes up the contents up the stack (whatever was in stack T was lost)
[ ENTER ]: pushes the contents of memory form the X stack to the Y stack, everything gets pushed up one level (whatever was in stack T was lost). Entering a new number replaces the copied number in the X stack.
[ CHS ]: changes the sign of the number of the X stack. For the HP 35 only, the sign can be changed before or after the number has been entered.
[ - ], [ + ], [ × ], [ ÷ ]: arithmetic keys
[ π ]: Enters pi (π) to the X stack. (3.141592654)
The Power Function
I hope you have noticed something about the power function. Usually the power function on must calculators (especially non-graphing scientific ones) usually are labeled y^x. On most RPN calculators, the number on the Y stack is treated as base while the number on the X stack is treated as the exponent.
This isn’t the case on the HP 35: the key is labeled x^y. The variables are reversed. On the HP 35’s power function, the number on the number of X stack is the base, while the number on the Y stack is treated as the exponent.
For example, let’s try to calculate 1.8^3.
On the HP 35: 3 [ ENTER ] 1.8 [ x^y ] returns 5.832000001
Now try 3^4.
On the HP 35: 4 [ ENTER ] 3 [ x^y ] returns 81.00000001
But 3^4 = 81. Why the extra .00000001?
Accuracy
Try: 1 ÷ 3 × 3
1, [ ENTER ], 3, [ ÷ ], 3, [ × ]
This returns .9999999999
(we would expect 1)
To put it simply, the HP 35 works with floating point numbers which are presented with a 10-digit mantissa and a 2-digit exponent. Numbers are stored as binary-coded decimal as 56 bit numbers. The calculator rounds off results in at 11th decimal place, leaving a possibility of the 10th place of a possible error of ±1 digit.
With the internal algorithms used with trigonometric, exponential, and other transcendental functions, round off errors can build up and show up in the 10th decimal place.
Something to keep in mind when doing calculations: The HP 35 does not have a fixed decimal place setting, so rounding to more accurate results falls to the responsibility of the user.
For more information:
Cochran, David S. “Algorithms and Accuracy in the HP-35” HP Journal. June 1972. Retrieved from
(Retrieved November 11, 2026)
“HP 35” Wikipedia. Last Edited on December 13, 2025. Retrieved November 11, 2026.
https://en.wikipedia.org/wiki/HP-35
Final Thoughts
The HP 35 is a well-time honored calculator and has brought the world of scientific computing to many in a handheld package. It is a delight to use and the red LCD display is timeless.
This would make an excellent entry-level scientific calculator for those who are learning RPN for the first time, providing a wealth of functions for a wide range of calculations. I recommend this calculator, physical handled and/or emulator.
Eddie
HHC Conference Videos
Please check out the Youtube page hpcalc.org at https://www.youtube.com/@hpcalc .
Videos and talks from HP Calculator conferences until HCC 2025 are posted. Look out for talks from HCC 2026 to posted there soon. Disclaimer: I do not run this account.
All original content copyright, © 2011-2026. Edward Shore. Unauthorized use and/or unauthorized distribution for commercial purposes without express and written permission from the author is strictly prohibited. This blog entry may be distributed for noncommercial purposes, provided that full credit is given to the author.
