HP 17bII+ Silver solver

[pinkman]

I'm very satisfied with my HP 17bII+ Silver, which I find very powerful, but also nice and not looking too complicated (no [f] and [g] functions on keys like the HP 12c or HP 35s that I used to use at work every day, but a really complete calculator except for trigs and complex numbers calculations).
So the 17BII+ is my new everyday calculator. I don't come back to arguments where a good (HP) calculator is a perfect complement or subsitute to Excel.

I chose the 17BII+ after having carefully studied the programs I use in my day to day work:
- price and costs calculations: can be modeled in the solver, with 4 or 5 equations and share vars
- margin calculations: built-in functions
- time value of money: built-in functions
- time functions: built-in functions
For the few moments I need trigs or complex calculations, I always have Free42 or a real 35s / 15c not really far from me.

After having rebuilt my work environment in the (so powerful) solver, I started to study the behavior of the solver, looking at loops (Σ function), conditional branchings (IF function), menu selection (S function), and the Get and Let functions (G(), L()).
I googled a lot and found an interesting pdf file about the Solvers of the 19B, 17B, 17BII and - according to the author, the 17BII+ Silver.
The file is here : http://www.mh-aerotools.de/hp/documents/...ET-LET.pdf

I tried a few equations in the "Using New and Old Values" chapter, pages 4 and following.
There I found lots of differences with my actual 17BII+ Silver, which I would like to share here with you.

For instance, the following equation found page 4:

Code:


A=2*G(A)+B

should return increasing values of A, recursively. It does not. Instead of incrementing A, it affects B to A, and then does not do anything when calculating A.

In the next example found page 5:

Code:


A=G(A)+1

A is set to 1, whatever the initial value is, and the number of calculations you do.

Then the equation :

Code:


Q=L(A:A+1)    ; page 5

effectively increases A and Q, but 2 by 2.

I don't understand the first 2 cases. In the first one, A should be set to -B, not B, if not using the "old" value of A. In the second one, the solver does not use the "old" value, but it does not also solve the equation, as there is no defined solution.
In the last case, I understand that the equation is evaluated twice before finding an answer. So the old value is used there, but not the way I could expect.

I finally found one - and only one - way to use iterations in the solver, with the equation :

Code:


G(A)+1=A

It calculates what I could define as A++ each time the equation is evaluated. Note that it sometimes does not, I still haven't found when and why.
Note that neither A=G(A)+1 or A=1+G(A) or G(A)+2=A works.

I'm not disappointed, as the solver is a really interesting and useful feature of the calculator, but I'm just surprised not having found more working cases of iterations, or a clear understanding of how the solver works.

Comments are welcomed.
Regards,
Thibault

[rprosperi]

The 17BII+ (Silver Edition) is an excellent machine, in fact it has the best keyboard of any machine made today by HP, but the solver does have a bug, and there are also a few other smaller issues making the solver slightly inferior to the 17B/17BII/19B/19BII/27S version.

See these 3 articles for details:

http://www.hpmuseum.org/cgi-sys/cgiwrap/...ead=242551

http://www.hpmuseum.org/forum/thread-657...l#pid58685

http://www.hpmuseum.org/cgi-sys/cgiwrap/...ead=134189

Overall these are not dramatic issues, and once you understand the solver bug, you likely can create equations that can avoid the issue.

I have most HP machines but the 17BII and 17BII+ are the ones I use most often for real work (vs. playing, exploring or following along interesting threads here).

Edit: added 3rd link

[Don Shepherd]

(09-14-2018 09:20 PM)pinkman Wrote:  I'm not disappointed, as the solver is a really interesting and useful feature of the calculator, but I'm just surprised not having found more working cases of iterations, or a clear understanding of how the solver works.

Thibault, when Kinpo built the 17bii+ calculator years ago (both the gold one and the silver one), they basically goofed the solver implementation. This has been discussed at length in the HPMuseum forum. The solvers on the 17b and 17bii work fine, just as you would expect them to. If you plan on making significant use of the solver and its incredible capabilities, forget the + and get an original 17b or 17bii. You won't be sorry.

Also, get the manual (it's on the Museum DVD) Technical Applications for the HP-27s and HP-19b. It applies to the 17b as well.

The Sigma function also works fine on the 17b and 17bii.

[pinkman]

Thanks to both of you for the details, links and advice.
I've read the threads carefully, I did not find them by myself first. It's the end of the night now, I'll try to make few testing later.

[rprosperi]

If you want to really explore the capabilities of the awesome Pioneer Solver, and confidently try to push it without worrying about using L() this way or that, I agree with Don, buy a 17BII and use that for the Solver stuff, but continue to use the 17BII+ for every day stuff.

Here's a very nice 17BII for only $25 (shipping included, in US) and you can even find them cheaper if you're willing to wait:

https://www.ebay.com/itm/HP-17Bll-Financ...3252533550

The 17BII is bug-free for solver use, while the 17BII+ has a much better LCD, readable in a wider range of lighting & use conditions.

In case you haven't seen this yet, here's an example of what can be done with the solver:

http://www.hpmuseum.org/forum/thread-2630.html

[pinkman]

Well I'll try to find one, even if I'm not in the US.
I also want to continue using my actual 17bII+ at work, as the solver is powerful enough for me (for the moment), and it looks really good.

Don and you Bob have done a lot to help understand what the solver can do, that's pretty good stuff.

Regards,
Thibault

[rprosperi]

Nah, Don and Gerson are the real Pioneer Solver Masters, I just have a good collection of links.

Of all the various tools built-in to the various calculator models I've explored, the Pioneer Solver is easily the one that most exceeds it's initial apparent capability. This awesome tool must have been incredibly well-tested by the QA team. The fact that the sheer size (and audacity!) of Gerson's Trig formulas still return amazingly accurate results says more about the underlying design and code than any comments I could add.

Enjoy exploring it, and when you've mastered it (or at least tamed it a bit), come back here and share some interesting Solver formulas. There are numerous folks here that enjoy entering the formulas and running some test cases. (well, I suppose "enjoy" is not really the right word about entering the formulas - I guess feel good about accomplishing it successfully is more accurate).

When I see some of these long Solver equations, it reminds of TECO commands back in the PDP-11 days.

[Thomas Klemm]

If you feel like entering long formulas you can solve the 8-queens problem.

):0)

Cheers
Thomas

[Don Shepherd]

How about a nifty, elegant, simple number base conversion Solver equation for the 17b/17bii, courtesy Thomas Klemm:

BC:ANS=N+(FROM-TO)\(\times \Sigma\)(I:0:LOG(N)\(\div\)LOG(TO):1:L(N:IDIV(N:TO))\(\times\)FROM^I)

Note: either FROM or TO must be 10 unless you are doing HEX conversions

[Thomas Klemm]

(09-17-2018 06:30 AM)Don Shepherd Wrote:  number base (2-10) conversion

(06-19-2014 02:06 PM)Thomas Klemm Wrote:  You could set either FROM or TO to 100. This allows conversions between HEX and DEC.

Example: CAFE_HEX
FROM: 16
TO: 100
N: 12101514
ANS: 51,966

Best regards
Thomas

[Don Shepherd]

(09-17-2018 07:49 AM)Thomas Klemm Wrote:  

(09-17-2018 06:30 AM)Don Shepherd Wrote:  number base (2-10) conversion

(06-19-2014 02:06 PM)Thomas Klemm Wrote:  You could set either FROM or TO to 100. This allows conversions between HEX and DEC.

Example: CAFE_HEX
FROM: 16
TO: 100
N: 12101514
ANS: 51,966

Best regards
Thomas

Wow, I didn't realize that Thomas. Stunning.

Don

[Martin Hepperle]

As has already been written, unfortunately the solver in this calculator is flawed.
Part of the problems comes from the fact that it evaluates the equation twice which leads to incrementing by two instead of one etc. But there seem to be more quirks.
This is too bad, as the solver is a very capable tool (with G() and L()), while still easy and versatile to use (with the menu buttons).
Accomplishing something similar (solve for one variable today, for another tomorrow) with modern tools like Excel is more complicated.

Martin

[rprosperi]

(09-16-2018 10:23 PM)rprosperi Wrote:  Nah, Don and Gerson are the real Pioneer Solver Masters, I just have a good collection of links.

An obvious correction is in order here:

Don, Gerson and Thomas Klemm are the real Pioneer Solver Masters...

No disrespect intended by the omission. A review of past posts of significant solver formulas quickly reveals just how often all three of these three guys were the authors.

[pinkman]

Be sure I have great respect of everyone mentioned above

Thanks for the advices, my new old 17bii is now in my hands ($29 on French TAS), I'm just waiting for the next rainy Sunday for pushing the solver to it's limits (mmh, to MY limits I guess).

[Jlouis]

Just one doubt here, Are the solver of 18C and 19B / BII the same of the 17BII and 27S?

TIA

[mfleming]

(09-28-2018 07:47 PM)Jlouis Wrote:  Just one doubt here, Are the solver of 18C and 19B / BII the same of the 17BII and 27S?

TIA

Check out the document by Martin Hepperle in this thread which lists the functions available for each calculator. The document is an excellent introduction to the Solver application.

Solver GET-LET

~Mark

[Jlouis]

(09-29-2018 01:37 AM)mfleming Wrote:  

(09-28-2018 07:47 PM)Jlouis Wrote:  Just one doubt here, Are the solver of 18C and 19B / BII the same of the 17BII and 27S?

TIA

Check out the document by Martin Hepperle in this thread which lists the functions available for each calculator. The document is an excellent introduction to the Solver application.

Solver GET-LET

~Mark

Thanks mfleming, I should have read the beginning of the manual. It is the same solver for the calculators of my question. I prefer to use the 19bII due the dedicated alpha keyboard, but the 27s is a pleasure to use too.

Cheers

[Don Shepherd]

(09-28-2018 07:47 PM)Jlouis Wrote:  Just one doubt here, Are the solver of 18C and 19B / BII the same of the 17BII and 27S?

TIA

The 27S solver has a random number function; the 17b/bii do not. I'm not sure about the 18c and 19b.

[Don Shepherd]

(09-28-2018 07:47 PM)Jlouis Wrote:  Just one doubt here, Are the solver of 18C and 19B / BII the same of the 17BII and 27S?

TIA

Here is a 3-page cross-reference I created a few years ago of Solver functions present in the 19bii, 27s, and 17bii.


  solver 1.PDF (Size: 1.08 MB / Downloads: 50)

  solver 2.PDF (Size: 968.77 KB / Downloads: 35)

  solver 3.PDF (Size: 494.75 KB / Downloads: 32)

[Jlouis]

(09-29-2018 11:53 AM)Don Shepherd Wrote:  

(09-28-2018 07:47 PM)Jlouis Wrote:  Just one doubt here, Are the solver of 18C and 19B / BII the same of the 17BII and 27S?

TIA

Here is a 3-page cross-reference of Solver functions present in the 19bii, 27s, and 17bii I created a few years ago.

Really thanks Dom for taking time to make these documents.

This is what makes MoHC a fantastic community.

Cheers

JL

Edited: Looks like the 19BII is a little more complete than the 27s and that the 17BII is the weaker one.