Posts: 1,665
Threads: 142
Joined: Jan 2009
Today I came just a little closer to completion of my collection of HP LED era calculators. My basic goal is to have one sample of each model of the Classic, Woodstock and Spice handhelds as well as associated printing desktops. I just bought on auction a complete working HP 27 set, which has proven to be one of the toughest models to find in good condition. I have allowed that certain models are covered by others, such as the HP 25 is covered by the HP 25C, and indeed sold my HP 25 when I found an HP 25C. I may also sell my HP 33E and HP 38E, since I now have both an HP 33C and HP 38C. The only models I now lack are the HP 29C and HP 70. I own a HP 19C, which somewhat covers the HP 29C, and have passed up opportunities to buy an HP 29C at high prices. As to the HP 70, well that may never happen as I've yet to see a good and complete one show up or anything at a remotely reasonable price. But perhaps never having a complete collection will keep me more interested in this hobby.
Edited: 3 Sept 2009, 9:39 p.m. after one or more responses were posted
Posts: 1,392
Threads: 142
Joined: Jun 2007
Michael, instead of the 70, why not go after a 65? I got one about two years ago and it has become one of my favorites. It was the first programmable, and the card reader/writer would be a welcome addition to many models today.
Posts: 1,665
Threads: 142
Joined: Jan 2009
Don, I already have an HP 65, in fact I've got 3 of them, all in perfect refurbished working condition. About a year ago I sold a fourth HP 65 that I'd fixed up, but was not in as nice cosmetic condition. I agree that it is a milestone product, and a lot of fun to use. I've also got 2 of the HP 67 and one HP 97. So I have all of the card reader models except the pre and post classic models. Like I said, I have all of the Classics, except the HP 70. Perhaps, I should have said one of each model rather than sample. I am also the proud owner of an HP 35 Red Dot. I'm only interested in the HP 70 for its uniqueness, since the HP 80 (I've got 4, including a version 1) completely trumps it in features, and besides mostly I use my HP 12C anyway.
But thanks for your observations.
Michael
Edited: 3 Sept 2009, 9:44 p.m.
Posts: 1,392
Threads: 142
Joined: Jun 2007
Oh, OK, now I see. I don't know if you have run into this on your 65s, but I think there is a problem with the y^x function for larger values when y is 10. For example, on my 65, using y^x, 10^1 through 10^6 are all OK, but 10^7 gives 9999999.99, 10^8 gives 99999999.60, 10^9 gives 1.000000003 09, 10^10 gives 1.000000000 10, and 10^11 gives 9.999999970 10. I would be curious if you get the same results. Maybe this was documented somewhere back in 1974 or 75, but I haven't seen it anywhere.
The EEX function works fine.
Posts: 1,665
Threads: 142
Joined: Jan 2009
Yes, all 3 of my HP 65's give the same results that you have. Two of them are version 1 with the serial number inside the battery compartment and indicating 1973 as date of manufacture and the third is a version 2 with the serial number between the bottom feet below the back label and indicating 1976 as the date of manufacture. I also get the same results with a HP 45 and HP 55, however, my HP 67 from 1978 gives the exact correct results. So, apparently this "bug" was common to the early classics, but later corrected in the second generation mag card models.
Posts: 2,309
Threads: 116
Joined: Jun 2005
I'm still hoping to find an 81 and a 9805A (which is similar to the 46 and 81).
Posts: 1,477
Threads: 71
Joined: Jan 2005
According to Wlodek the 91 was the first HP calculator to correctly deal with integer exponentiation. The HP-91 came out on March 1, 1976, the 27 and 29C and 19C came out later and also handled integer exponents correctly. I don't think that HP revised the ROM's of the earlier Woodstocks (21, 22, 25, 25C) even though they continued to produce them for a couple of years after the 91 came out.
Posts: 222
Threads: 21
Joined: Oct 2006
I have a neat little application on my Palm Treo 650 called CSIM. It simulates the HP-35, HP-45 or HP-55, depending on setup. It also gives the answers above for each model. It is a great little Palm OS program for anyone (like me) not lucky enough to own a real Classic.
Jeff Kearns
Posts: 1,665
Threads: 142
Joined: Jan 2009
Katie,
I tested my HP 21 (s/n 1710A), HP 22 (s/n 1601A), HP 22 (s/n 1711A) and HP 25C (s/n 1703A) and both the HP 21 and HP 25C give the same erroroneous results, however, the HP 22's give the correct exact results. Curious that a business calculator would be more accurate with large numbers than a scientific calculator.
Michael
Posts: 1,477
Threads: 71
Joined: Jan 2005
Michael,
You're right my HP-22 works correctly on integer exponents. The 22 came out before the 91, in fact the serial number on mine is 1601A.... so it was made in January of 1976. I think that the book may have this wrong.
Thanks,
Katie
Posts: 1,665
Threads: 142
Joined: Jan 2009
Quote:
I'm still hoping to find an 81 and a 9805A (which is similar to the 46 and 81).
I also don't have a 9805A, but am not looking for one, since I only care about HP RPN calculators. The 46 is one of my collection favorites, along with the 91, 65/67/97, 19C and 34C.
Posts: 88
Threads: 8
Joined: Jul 2007
Hi Katie, thanks,
And thanks to Gene Wright for pointing out this thread to me.
HP described the increased accuracy of y^x in an article in the HP Journal. (I don't have the exact issue reference to hand, sorry.) The article described the HP-91 as they were writing about the new Topcat series, but I assume that in fact the improved algorithm was also put in the HP-22, which was introduced a few months before the 91, but after the HP-21 and the HP-25. I believe that the HP Journal articles were sometimes written long before they were printed, so someone else could have sneaked a new algorithm into a different model before the article was published, saying the HP-91 was the first model with this algorithm ;-)
Anyway, thanks for pointing this out - something else to change in the next edition,
Wlodek
Posts: 2,309
Threads: 116
Joined: Jun 2005
The 19C, 22, 27, 29C, 67, and 97 use the newer logarithm and exponential code, verified by studying the ROM code.
The 21, 25, and 25C use the older code, verified by studying the ROM code.
I suspect that the 92 and 95C use the newer code, but I haven't verified it.
Posts: 758
Threads: 9
Joined: Jul 2007
I've always found these differences in results for the same calculation to be interesting. Here are three examples that puzzle me a bit. Each is a 2500-iteration Savage Benchmark, where the "perfect" numerical result would be 2500.
HP-25
01 RAD
02 STO 0
03 0
04 STO 1
05 RCL 1
06 1
07 +
08 SQRT
09 X^2
10 LN
11 e^X
12 ATAN
13 TAN
14 STO 1
15 RCL 0
16 1
17 -
18 STO 0
19 X=0?
20 GTO 22
21 GTO 05
22 RCL 1
23 DEG
24 GTO 00
Enter 2500 and press R/S
Result = 2499.079996 after 5940 seconds.
HP-67
01 LBL B
02 RAD
03 ST I
04 0
05 LBL 1
06 1
07 +
08 SQRT
09 X^2
10 LN
11 e^X
12 ATAN
13 TAN
14 DSZ I
15 GTO 1
16 DEG
17 RTN
Enter 2500 and press B
Result = 2499.976148 after 5840 seconds
HP-34C
01 LBL B
02 RAD
03 STO I
04 0
05 LBL 1
06 1
07 +
08 SQRT
09 X^2
10 LN
11 e^X
12 ATAN
13 TAN
14 DSE I
15 GTO 1
16 DEG
17 RTN
Enter 2500 and press B
Result = 2499.969898 after 7530 seconds
The HP-41CX after 3210 seconds and the HP-15C after 5840 seconds produce numerical results identical to the HP-34C.
SUMMARY (Result and time in seconds)
HP-25____2499.079996____5940
HP-67____2499.976148____5840
HP-34C___2499.969898____7530
HP-41CX__2499.969898____3210
HP-15C___2499.969898____5840
The benchmark does not use the Y^X function, so other differences must exist in the numerical algorithms of the HP-25 and the later models.
Suprisingly, the HP-67 result differs slightly from that of all the later pre-Saturn calculators. However, if the benchmark is run for less than 1500 iterations, the HP-67's result is identical to the later calculators!
I don't have a working HP-19C to check results for the 19C/29C models, but I'd bet they show the same strange effect as the HP-67.
Edited: 6 Sept 2009, 12:53 p.m.
Posts: 1,665
Threads: 142
Joined: Jan 2009
I ran the Savage benchmark on my HP-19C, and the result is 2499.976148 as you predicted. I used pretty much the same code as you did for the HP-25, except that the GTO <line number> are replaced with GTO <label> and labels 0,1,2 were inserted. Also, I inserted a R/S before the last line (GTO 0) to prevent it from looping. I didn't time it precisely, but it took approximately 6600 seconds to run.
|