HP35 with "the Bug" and the CHS-EEX

[Dan W]

Hi all,

I have one HP35 with "The Bug" and one without. The behavior of the CHS key with the EEX key is different with the buggy HP35 than the other. But I haven't found this difference written up anywhere.

On the buggy HP35, if I enter a number, then EEX, I do NOT see two digits off to the right until I type them in. On the non-buggy HP35 I see two zeros to the right as soon as I hit EEX.

On the buggy HP35 I must hit the CHS key immediately after the EEX key to get a negative exponent. When I do this I get -00 displayed, then I can punch in the actual exponent. But after that the CHS key does not do anything - I can't change the (-) to a (+).

On the non buggy, I can enter the EEX digits and hit CHS at will, reversing the EE sign each time.

Is this known behavior of some HP-35's or related to The Bug? Or is this a different problem?

Thx

-- Dan

[Eric Smith]

I think that was a separate change, which occurred in a release before the famous bug fix. There were at least three public releases of the HP-35 firmware.

If you ever have occasion to open them, please note the part numbers on the ROMs (round metal cans, part number 1818-xxxx).

[Dan W]

The ROMs for the one with the bugs are:

K7239FC 1818-0020
K7233GJ 1818-0017
K7231BF 1818-0006

The numbers are on the side, not the top.

Is there a document describing in detail the different ROMs sold and the bugs or errata in each set?

Thanks,

-- Dan

[Eric Smith]

There are at least three public releases, and possibly four. There were internl engineering releases, of which it is possible that some prototype units may exist.

Prototypes: combinations including 1818-0001 through 1818-0019

Earliest known release: 1818-0006, 1818-0017, 1818-0020

Intermediate: (unknown, possibly unreleased)
combinations including 1818-0021 through 1818-0025

Bug fix: 1818-0024, 1818-0026, 1818-0027

Revision, purpose unknown: 1818-0024, 1818-0026, 1818-0028

Peter Monta has dumped the (6, 17,20) set, and a later release that is believed to be the (24, 26, 27) set, so these can be compared in simulation in Nonpareil. At some point it will be a good idea to dump the sets with 27 and 28 for verification.

[Dan W]

Thanks Eric.

This bug affecting the operation of CHS with EEX has much more user impact than the 2.03 ln e^x bug, but is less well known. Was it resolved quickly in an intermediate release before the more famous bug?