DM32 review

[Vincent Weber]

Hi all,

I have received today my brand-new DM32.

This is my very first Swissmicro product. Why this one? Well, I didn't want the 42 since I can't leave without the Plus42 extentions and since Free42 isn't even up to date of the last DM42 firmware; I had read that the DM41X had had unfixed bugs for 2 years; and I have ordered the 15C CE already, so I didn't want to buy it twice... On the contrary, the 32 intrigued me: what a brave attempt to completely recreate a calculator from scratch!

So, here are my first impressions.

Overall, I really like it! The hardware is small, cute, yet solid, powerful, reliable. The screen is very readable, great. Classy! The simulation software seems really good. Save for a very minor point (see below) I could not fail the normal 32SII operation so far. The fractions feature passes all the tests that are on Craig Finseth page, exact same results as my (latest ROM) real 32SII; the state-switching feature is awesome and really makes up for the lack of labels (if not of registers). Same as an HP-67 with lots of cards!

For a first version, this is really satisfactory. I have also some (small) regrets and suggestions for improvements:

-The keys are a bit too stiff;
-The key labels do not show a distinct sign for menus (the real 32SII kind of underlines them) ;
-The big screen is completely wasted in program mode: only one step is shown, where 4 (or more with a different font?) could be shown;
-It is even more wasted in equation mode, where not even the stack is shown - just blank! Why not showing the equation list?
-A very minor point: on the original machine, trying INPUT or LBL in direct mode briefly shows a warning (the exclamation mark annunciator). No such thing on the DM32. I suspect it is because it is so fast that you don't have time to see it anyway. Cases where the annunciator is sticky (like the conditionals in direct mode) are correctly handled on the DM32.

A question I have that is not answered in the manual: it seems you can't OFF the machine completely, there is always a random picture on screen. This must consume some battery? Anyway to change the pictures and their order?

All in all, a very solid job for a first version, congrats Swissmicro! What's next? A "35S done right" would be great!

Cheers

[rprosperi]

Thanks for the review comments, lots of readers here are interested in such details and in a newcomer's (to SM) impressions.

Here's a couple quick replies, other users will surely jump in with more detailed comments and answers.

1. The LCD used really does not use any power when off, even when showing a display. Really. Zero, none. Tested, proven, etc.

2. You bought a BETA product, as clearly described in the description on SwissMicros' website. As such, it is, by definition, not complete. Many of the f/w shortcomings you note will be addressed in subsequent f/w releases.

Note that the above comment is not meant to say such comments are inappropriate or unwelcome, most new users react the same way. I only bring it up to say: Stay Tuned, I think you'll be happy.

Thanks again.

[Vincent Weber]

Many thanks Bob!
I actually said that for a first version it is great
I was just a bit surprised that the product picture on SM web site shows a multiline program edition, which wetted my appetite, but I could only get one-line edition. Well if this feature is coming that's great!

Cheers

[jonmoore]

(05-30-2023 09:07 PM)Vincent Weber Wrote:  A question I have that is not answered in the manual: it seems you can't OFF the machine completely, there is always a random picture on screen. This must consume some battery? Anyway to change the pictures and their order?

Cheers

The off state of the LCD is actually a feature of the LCD (and a damn fine one to boot). The LCD has a behaviour much like that of an eBook. The LCD has a permanent off state that effectively acts as a screensaver would (not that an LCD's screen gets burn-in in the same manner as a computer monitor).

In fact, when the DM42 was first released, one of the save states was of a blank image, which really messed with people's grey matter! I'm uncertain if the feature uses zero power or virtually zero power. I suspect it's the former as since getting my 41X (pre-launch beta device, 2 years-ish ago) my 42 gets barely any use, and it's still on the same battery as when I acquired my 41X.

The last thing you need to be worried about is battery drain if the new SoC performs as well as the old hardware. I'm guessing it's actually improved - based on a vague memory based on something Michael might have said in this HPCC presentation. The whole video is worth watching, but Michael's part is in the second half.

[Vincent Weber]

Very interesting, thanks a lot! This is worth a note in the manual!

[rprosperi]

(05-30-2023 09:42 PM)Vincent Weber Wrote:  Many thanks Bob!
I actually said that for a first version it is great
I was just a bit surprised that the product picture on SM web site shows a multiline program edition, which wetted my appetite, but I could only get one-line edition. Well if this feature is coming that's great!

Cheers

I like to think of it as the 0.8th version... but why bicker over definitions...

As for the editor, perhaps that photographer had a time machine...

[jonmoore]

Bob, you posted your reply whilst I was drafting mine, so apologies for any crossover.

One thing for Vincent to note though, is that I have a very early beta of the 41X (as you know), and I'm delighted to say that the promised improvements arrived, and they were worth waiting for. I think I joined the beta approximately 6 months before the public launch, so I'm sure Vincent won't be kept waiting too long.

[Vincent Weber]

Good to know about DM41X, many thanks!

[Vincent Weber]

Another small question, rather than issue.

When the function comes in a menu, the keys on the second row acts as on the original, i.e as soft keys to access the functions. Long press shows the function name, and, as an enhancement, returns to NULL as on the 41. So far so good.

But I noticed you can also access the functions using the top row (F-keys), as it is quite intuitive to do so. The behavior is different though, in that case the function is immediately executed no matter what, no long press behaviour.

Is this intended?

[Vincent Weber]

The more I use it, the more I like it!
I think the DM32 has the potential of becoming the "35S done right".
I have been thinking of 2 kinds of improvements : non-intrusive ones, gently expanding the functionality while retaining full compatibility. And intrusive ones, that would break compatibility, probably through a new special mode.

I) Non-intrusive:

- Double the numbers of labels and directly accessible registers with lowercase letters (using shift keys) ;
-Increase the number of indirectly accessible registers. Just store 100 in i, and RCL(i) would be valid. Don't bother with the double indexing (I and J) of the 35S, too disruptive for little benefit;
-Introduce labels that are local to a lettered program (00...99) using shift keys;
-Enhance equation editing not to have to delete from the right, with left and right arrow keys, that could simply be mapped to 2 of the F-keys;
-Enhance equation functionality to match the 35S', by allowing STOring into temp variables, and pressing RDN to use (read-only) REGX, REGY, REGZ, REGT and possibly REGL (alternative would be to use LASTX as allowed in the 35S);
-Expand COMPLEX to support ENTER, X<>Y, LASTX, RDN, RUP... for full 2-level complex stack, as pointed by JF;
-Expand complex functions to reverse trigs and hyperbolics;
-Introduce vectors? Not sure, they are pretty useless on the 35;
-Introduce physics constants in a separate menu.

II) Intrusive:

This would mainly dealing with introducing one-line complex numbers, to bring the functionality closer to the 42S', maybe without the little flaw of having to use 2 stack levels to enter a complex number. Maybe COMPLEX, in this mode, could accept One-line entry, like 1 COMPLEX 2 on one line would generate 1+2i. Some thinking has to be done, probably we would need a polar mode and annunciator, and a way to extract parts from a complex number, unlike the biggest design flaw of the 35S!

What do you guys think?

[Dave Britten]

(05-31-2023 09:52 AM)Vincent Weber Wrote:  The more I use it, the more I like it!
I think the DM32 has the potential of becoming the "35S done right".
I have been thinking of 2 kinds of improvements : non-intrusive ones, gently expanding the functionality while retaining full compatibility. And intrusive ones, that would break compatibility, probably through a new special mode.

I) Non-intrusive:

- Double the numbers of labels and directly accessible registers with lowercase letters (using shift keys) ;
-Increase the number of indirectly accessible registers. Just store 100 in i, and RCL(i) would be valid. Don't bother with the double indexing (I and J) of the 35S, too disruptive for little benefit;
-Introduce labels that are local to a lettered program (00...99) using shift keys;
-Enhance equation editing not to have to delete from the right, with left and right arrow keys, that could simply be mapped to 2 of the F-keys;
-Enhance equation functionality to match the 35S', by allowing STOring into temp variables, and pressing RDN to use (read-only) REGX, REGY, REGZ, REGT and possibly REGL (alternative would be to use LASTX as allowed in the 35S);
-Expand COMPLEX to support ENTER, X<>Y, LASTX, RDN, RUP... for full 2-level complex stack, as pointed by JF;
-Expand complex functions to reverse trigs and hyperbolics;
-Introduce vectors? Not sure, they are pretty useless on the 35;
-Introduce physics constants in a separate menu.

II) Intrusive:

This would mainly dealing with introducing one-line complex numbers, to bring the functionality closer to the 42S', maybe without the little flaw of having to use 2 stack levels to enter a complex number. Maybe COMPLEX, in this mode, could accept One-line entry, like 1 COMPLEX 2 on one line would generate 1+2i. Some thinking has to be done, probably we would need a polar mode and annunciator, and a way to extract parts from a complex number, unlike the biggest design flaw of the 35S!

What do you guys think?

All fine ideas, though I would probably be a bit more conservative. I always liked the 32SII as is, it just needs (lots) more memory. And that problem has already been solved here!

- I would definitely add more storage registers by extending the range of indirect addressing, as you've suggested. MAYBE add an additional flag to enable/disable this if there are program compatibility concerns.
- Originally I thought adding local labels, more global labels, etc. would be a good idea, but I think the quick swapping of active states is a more elegant way of doing it. It's like having multiple program cards for your HP 67, or the program areas of the (unreleased) 95C, but with more convenience and storage space. So I'd keep the labeling as it is now for simplicity and direct compatibility with the 32SII.
- However, I would like to see the option of swapping states while retaining the current storage registers, i.e. having one set of shared storage registers and only swapping out the program/equation storage. And maybe the option of swapping states under program control. Not must-haves or anything, but it would be a nice enhancement.
- Expand complex support - definitely, no reason not to, in my opinion.
- Equation-editor enhancements - yes, arrow keys on the top row function keys would be a nice addition.

[Eric Rechlin]

(05-30-2023 09:07 PM)Vincent Weber Wrote:  -The key labels do not show a distinct sign for menus (the real 32SII kind of underlines them) ;

This was something that disappointed me with the DM42 as well. The actual Pioneer calculators have a slightly different color background behind the menu keys to differentiate them. I assume it was an intentional omission, perhaps as a cost-saving measure to avoid having to pay for one more color of silk screening on the keyboard. But at the price of these I agree that it should have been included. The WP 34S accomplishes this with an underline, which is simpler to do but less attractive.

[J-F Garnier]

(05-31-2023 02:59 PM)Eric Rechlin Wrote:  

(05-30-2023 09:07 PM)Vincent Weber Wrote:  -The key labels do not show a distinct sign for menus (the real 32SII kind of underlines them) ;

This was something that disappointed me with the DM42 as well. The actual Pioneer calculators have a slightly different color background behind the menu keys to differentiate them.

Actually, the different background for menus on the 32SII/42S is hardly visible and depends on the lighting conditions.
In fact, you have to know the difference exists to notice it.

J-F

[Eric Rechlin]

(05-31-2023 03:30 PM)J-F Garnier Wrote:  

(05-31-2023 02:59 PM)Eric Rechlin Wrote:  This was something that disappointed me with the DM42 as well. The actual Pioneer calculators have a slightly different color background behind the menu keys to differentiate them.

Actually, the different background for menus on the 32SII/42S is hardly visible and depends on the lighting conditions.
In fact, you have to know the difference exists to notice it.

J-F

I've never had problems seeing it; I thought it was quite noticeable. But maybe SwissMicros never noticed it!

[J-F Garnier]

(05-31-2023 03:34 PM)Eric Rechlin Wrote:  

(05-31-2023 03:30 PM)J-F Garnier Wrote:  Actually, the different background for menus on the 32SII/42S is hardly visible and depends on the lighting conditions.
In fact, you have to know the difference exists to notice it.

I've never had problems seeing it; I thought it was quite noticeable. But maybe SwissMicros never noticed it!

It may depend on the production batches/locations.
See my picture posted here, the menus can't be seen on the 32S and just visible on the 32SII's.
On the contrary, on this highly contrasted picture from jonmoore, the menus are well visible on the 32SII but less visible on the 32S.

j-F

[jonmoore]

(05-31-2023 03:46 PM)J-F Garnier Wrote:  

(05-31-2023 03:34 PM)Eric Rechlin Wrote:  I've never had problems seeing it; I thought it was quite noticeable. But maybe SwissMicros never noticed it!

On the contrary, on this highly contrasted picture from jonmoore, the menus are well visible on the 32SII but less visible on the 32S.

j-F

As it happens, my photo is shot in natural daylight with minimal processing, but the daylight is causing a significant soft shadow from the left side (low hazy Sun, shining directly through a kitchen window). The labelling on the 32S only features a secondary 'button' ink on the labels that aren't single function, e.g. Last X doesn't feature a secondary background ink but Parts beside it does. It's subtle but very effective - even though it would fail any contemporary colour blindness accessibility test. The shadowing on the alpha labels is most definitely accentuated by the soft shadow. In reality, this is only significant on the top A to F labels which are a reverse of the body labelling on everything below.

My 32Sii by comparison has spent far too many abusive days in school bags without a protective case!

[Vincent Weber]

(05-31-2023 02:39 PM)Dave Britten Wrote:  All fine ideas, though I would probably be a bit more conservative. I always liked the 32SII as is, it just needs (lots) more memory. And that problem has already been solved here!

- I would definitely add more storage registers by extending the range of indirect addressing, as you've suggested. MAYBE add an additional flag to enable/disable this if there are program compatibility concerns.
- Originally I thought adding local labels, more global labels, etc. would be a good idea, but I think the quick swapping of active states is a more elegant way of doing it. It's like having multiple program cards for your HP 67, or the program areas of the (unreleased) 95C, but with more convenience and storage space. So I'd keep the labeling as it is now for simplicity and direct compatibility with the 32SII.
- However, I would like to see the option of swapping states while retaining the current storage registers, i.e. having one set of shared storage registers and only swapping out the program/equation storage. And maybe the option of swapping states under program control. Not must-haves or anything, but it would be a nice enhancement.
- Expand complex support - definitely, no reason not to, in my opinion.
- Equation-editor enhancements - yes, arrow keys on the top row function keys would be a nice addition.

Excellent feedback, thanks!
Yes, I was thinking that the programmability of the 32sii is quite similar to the one of the 67. With state files playing the role of the cards, the similarity is even greater. Indeed, being able to share registers ("data cards"?) among state files would be great!

The 67 has 2 programming features that the 32sii is lacking:
-Rapid backwards branching n steps, when - n is stored in I. This is supposed to save labels and be faster, but is not very elegant nor readable. Since the DM32 is super fast and has 6 more global labels than the 67, I don't see this as a big deal;
-User entry flag, to test the entry of a number, thus allowing solvers à La 12C (with a same key, if a number has been entered, store it, otherwise solve for the variable). Could it be useful on the 32SII, which philosophy relies more on INPUT for data entry... Not sure about that, what do you think?

Cheers

[WillMarchant]

I didn’t see any bit operators like AND, OR, XOR, and bit shifting. And the fixed word size won’t let this replace my HP-16C for my day to day work. Maybe these could be added in as “expansions”? But the Hp-16C has those functions so conveniently on the main panel…

[Dave Britten]

Yeah, the 32SII had no bitwise operators. Thankfully there's more than enough memory on the DM32 to implement them as programs.

[jonmoore]

(06-01-2023 12:40 AM)WillMarchant Wrote:  I didn’t see any bit operators like AND, OR, XOR, and bit shifting. And the fixed word size won’t let this replace my HP-16C for my day to day work. Maybe these could be added in as “expansions”? But the Hp-16C has those functions so conveniently on the main panel…

To be fair, the DM32 (and original HP32 series calculators) don't pretend to crossover in the specialist binary functionality of the HP-16C. For that, you can use the excellent emulator of the HP16C for the DM41X created by Gregory J. McClure and Ángel Martin (and all 41 series hardware). This is a 16C emulator and so much more. And it has a permanent place in my DM41X ROM Map.



By comparison, the DM32 has simple but effective binary conversion functions. The DM42 is a step-up and provides bit operators too (but it's still in no way a HP-16C alternative). And in the realm of HP calculators, the 48 series has a wealth of binary functions out of the box. The downside of the 48 series is that they don't handle 2's complement signed binary conversions properly within the default toolset. But's it's simple to resolve with some simple bespoke RPL routines. All the RPN calculators I've mentioned handle 2's complement signed binary conversion accurately with their native toolset.

Anyway, getting back to the DM32, it doesn't pretend to offer a significant computer science toolset (like the HP16C), but what it does, it does well.