TI officially announced the TI-82 Advanced Edition Python today by adding it to their lineup of graphing calculators on education.ti.com.



Critor broke the news this morning over on TI-Planet with a full article going into the advertised specs and some speculation about its hardware. I strongly recommend you go take a look at that thread to get all the details but here is a summary of noteworthy points.
  • It seems likely this contains the hardware of a TI-83 Premium CE Edition Python inside given the advertised specs and the relatively small amount of time between the release of the TI-83 PCEEP and the TI-82 AEP.
  • I'm surprised they decided to stick with the TI-84 Plus style case as opposed to the newer/slimmer CE style case, but I think this is because of the next point:
  • It will be powered by AAA batteries, not a Li-Ion battery like the CE Blink
  • It will be considerably cheaper than the TI-83 PCEEP, at an estimated 60€ (approx $72 USD), which is surprising for something that seems to have the same hardware, but I guess that's just TI's insane margins Rolling Eyes
  • It will be available for back to school 2021.
  • It will have 154k of ram and 3MB of flash, just like a CE...
  • It looks like TI plans to continue making the TI-82 Advanced because they haven't removed it, but I'm not so sure about that because the newest we've seen was manufactured in early 2018...
  • It will have an exact math engine like the TI-83 Premium CE and TI-83 Premium CE Edition Python.
  • Of course, like the old TI-82 Advanced, the Python Edition will have the USB port on the top, so it can't be the exact same board as a TI-83 Premium CE, but I would expect just a rearranging of the board rather than a full redesign.
  • Just like the TI-82 Advanced, there's no assembly language support.
Just saw the announcement on TI-Planet!
Now if only TI's naming schemes would start making sense...

Do we know if this will be using an ez80 like the CE or something more akin to the CSE? (Minus the python co-processor of course)
Oxiti8 wrote:
Just saw the announcement on TI-Planet!
Now if only TI's naming schemes would start making sense...

Do we know if this will be using an ez80 like the CE or something more akin to the CSE? (Minus the python co-processor of course)

I've stopped trying to understand the naming scheme Laughing
I can't imagine it would be a z80 because the python is already so unbearably slow... But then again TI isn't exactly good at noticing this kind of thing...
I've done some speculations over at TI-Planet, which I'll summarize here:

This looks like a heavily cost-optimized TI-83 PCE. The case is derived from the TI-82 Advanced and uses AAA batteries that are not included, which would be cheaper than a Li-ion battery solution. The lack of TI-Innovator support suggests that there's no USB-OTG support, which combined with the AAA battery would greatly simplify the power circuits (no Li-ion battery to charge, no USB-OTG devices to power from the battery...). The screen and firmware seems lifted from the TI-83 PCE.

That might be just that, but given the price they might have optimized further in order to maintain their margins. Having both a custom ASIC with an ez80 core and an off-the-shelf ARM Cortex-M0 microcontroller is costly. There are ways to cut that down (one off-the-shelf Cortex-M0 microcontroller with an ez80 emulator for the firmware and direct execution of MicroPython, integrate the Cortex-M0 core to their custom ASIC alongside the ez80 core, port MicroPython to the ez80 and ditch the co-processor altogether), but that is pure speculation at this point. Heck, they could even ditch the RTC if they are really stingy.

For me, this is an answer to the NumWorks calculator for the French market, with a product roughly similar to the N0110 specs-wise (which has no USB-OTG and a rather simple PCB to match) with a lower price. TI finally overhauled its long-overdue entry-level model and I do admit I'm pleasantly surprised so far, because I would've expected them to be lazy and stick an ARM Cortex-M0 co-processor to the TI-82 Advanced if I were to guess.
boricj wrote:

That might be just that, but given the price they might have optimized further in order to maintain their margins. Having both a custom ASIC with an ez80 core and an off-the-shelf ARM Cortex-M0 microcontroller is costly. There are ways to cut that down (one off-the-shelf Cortex-M0 microcontroller with an ez80 emulator for the firmware and direct execution of MicroPython, integrate the Cortex-M0 core to their custom ASIC alongside the ez80 core, port MicroPython to the ez80 and ditch the co-processor altogether), but that is pure speculation at this point. Heck, they could even ditch the RTC if they are really stingy.

They didn't ditch the RTC if the paragraph on their website is correct.
Historically, redesigning asics has been a fairly cost-prohibitive endeavor for TI so I'm not convinced they would have integrated the arm coprocessor but that could potentially explain the very significant price difference (almost half the price of the 83PCEEP). Doing so would also finally make the python much faster overall, which I would hope is high on their list of priorities... Rolling Eyes
If they went through the trouble of doing that, I would expect them to also release a new revision of the 83PCEEP / 84+CE-TEP to use the new asics since they would be much cheaper and faster than the old asic + coprocessor. Either way, its a good deal for the end user and I would have no problem recommending it at that price.
On the topic of redesigning ASICs, I don't know what kind of process nodes they are using for the TI-8x series, but I'd be very surprised if it's not from the 20th century. Process nodes that old are quite cheap (relatively) to design, validate and manufacture with. Besides, TI is a major semiconductor designer and manufacturer, they already have tons of engineers on hand for that sort of work.

According to Datamath, the TI-82 Advanced ASIC is the same as the one from the TI-84 Plus SE calculator from 2008 onward. The TI-83 Premium CE ASIC is probably from 2014.

TI may be legendary stingy when it comes to ASICs, but Python competition on calculators is really heating up, at least in France. If they can't get a MicroPython port up and running on the ez80 ISA then they know their current Python patch job is a dead-end, especially in the mid-range... But they could also deem the current ASIC+co-processor solution to be good enough for an entry-level graphing calculator for a very long time. Which, to be honest, it would probably be.
it has a program button but I wonder if it's displays asm or python files ... this is basically a backup for the TI-84PCE, since security is getting worse on the TI-84 plus CE they're deciding to use a backup so they can still make money off testing calculators.

"I think ASM may be coming back to TI-84PCE, war production might be discontinued <_<"
It looks like the battery life should be comparable to the CE's. TI's rechargeable batteries are 4.4 Wh, while alkaline AAAs are 1.3-1.8 Wh each (according to Wikipedia), for a total of 5.2-7.2 Wh. Of course, that still means having to change batteries rather than just recharging them & also relies on TI's power circuitry being similarly efficient in both cases. Or use rechargeable AAAs, but it looks like some of those have less capacity than TI's battery—Wikipedia lists 0.42-1.6 Wh/cell, for a total of 1.68-6.4 Wh for 4 of them.
This is a bizarre launch to see. It seems like the exact reversal of the CSE -> CE situation. It will interesting to see performance numbers as we start to get our hands on this. With the CSE's release (interestingly, my first graphing calculator) it was clear that the z80 they decided to keep for the model was severely underpowered. I certainly don't own a more agonizingly slow calculator in my collection. They very promptly fixed this with the CE, immediately discontinuing the slow, bulky, but still expensive CSE.

I don't know what part of the market this seeks to occupy, especially if it will be as slow as the CSE was. The PCEEP is only nominally more expensive than the $72 mrwompwomp is quoting here.

100% chance I'll still buy one tho

Also, while typing this I was looking up calc specs, turns out both me and womp both maintain the TI calc comparison Wikipedia article.

tr1p1ea from the future wrote:
Well I'm pretty sure it's CE level hardware in a CSE aesthetic, so the performance should be on par with the CE? The specs suggest it will have 154k of ram and 3MB of flash, just like a CE.
If we're using old calc styling anyway, why not do something sick like the TI-85? That took AAAs too, and it was thinner. Darn TI-84 is a chunk for no very good reason. While we're at it, get me full color Voyage 200.
Well I'm pretty sure it's CE level hardware in a CSE aesthetic, so the performance should be on par with the CE? The specs suggest it will have 154k of ram and 3MB of flash, just like a CE.
Sam wrote:
This is a bizarre launch to see. It seems like the exact reversal of the CSE -> CE situation. It will interesting to see performance numbers as we start to get our hands on this. With the CSE's release (interestingly, my first graphing calculator) it was clear that the z80 they decided to keep for the model was severely underpowered. I certainly don't own a more agonizingly slow calculator in my collection. They very promptly fixed this with the CE, immediately discontinuing the slow, bulky, but still expensive CSE.

I don't know what part of the market this seeks to occupy, especially if it will be as slow as the CSE was. The PCEEP is only nominally more expensive than the $72 mrwompwomp is quoting here.

100% chance I'll still buy one tho


Like tr1p1ea said, I would expect at the very least CE hardware judging by the specs. The redesigned ASIC would be even better but I doubt it.
I think the market they are aiming for is for middle-school/algebra. More or less filling the hole left by the TI-73 Explorer that was discontinued years ago. At least in the European market...
Regarding the price. I initially said "about half" but its more like 75% the price of a PCEEP which is still a very nice price-point IMO.

Sam wrote:
Also, while typing this I was looking up calc specs, turns out both me and womp both maintain the TI calc comparison Wikipedia article.


Yeah I sometimes contribute to calculator stuff on wikipedia, not very frequently though and mostly just when something is literally wrong or the articles are missing very large chunks of relevant information. Speaking of which, the TI-Nspire CAS Touchpad was released in 2010, not 2007 Laughing I'll fix it.

Sam wrote:
tr1p1ea from the future wrote:
Well I'm pretty sure it's CE level hardware in a CSE aesthetic, so the performance should be on par with the CE? The specs suggest it will have 154k of ram and 3MB of flash, just like a CE.
If we're using old calc styling anyway, why not do something sick like the TI-85? That took AAAs too, and it was thinner. Darn TI-84 is a chunk for no very good reason. While we're at it, get me full color Voyage 200.


You're like the 3rd person to bring up refreshing the 68k lineup. I strongly doubt TI is headed in that direction, but I don't hate the idea assuming it would be implemented reasonably well... I suspect that if TI did cobble together a full color 68K calc, it would be the CSE all over again.
Quote:
I suspect that if TI did cobble together a full color 68K calc, it would be the CSE all over again.

A 68000 @ 12 MHz would produce a better screen fill rate than that of the 84+CSE, but I agree: if TI were to make a 68k-based calculator with a color screen, in order to improve performance, they'd better use a newer 68k processor, though the classic 68000 can be overclocked somewhat, and implementations of the 68000 in newer foundry technology could run at higher clock rates.
However, they won't make such a model 15+ years after the la(te)st OS upgrade for the TI-68k series. ARM processors replaced 68000 processors in TI's graphing calculators lineup, as well as pretty much elsewhere.

Fill rate estimations:
* 320x240x16bpp = 153600 bytes, i.e. 76800 words or 38400 long words;
* move.w dn,(an)+ is 8 clock cycles, which means that there can be 1.5M such instructions per second @ 12 MHz;
* move.l dn,(an)+ is 12 clock cycles, so 1M such instructions per second @ 12 MHz;
* movem.l to memory are 12 to 16 clock cycles + 8 x number of registers;
* branch instructions (need to throw one in the mix once in a while !) are 8 to 12 clock cycles, decrement and branch instructions are 10 to 14 clock cycles.
=> for a 320x240x16bpp screen, a 68000 @ 12 MHz can achieve >15 FPS maximum full screen fill rate with word writes, 25 FPS with long writes, and 30+ FPS with multiple move long writes.
Small update today:
The first actual price has been published here. It is 64,95€ (taxes included) and 51,95€ for bulk purchases. Of course, it can only be pre-ordered currently.
This is about 20% cheaper than an 83PCEEP (79,95 €) and 25% cheaper for bulk purchases (69,95 €).
Not bad for what is likely the same hardware without a battery Evil or Very Mad
  
Register to Join the Conversation
Have your own thoughts to add to this or any other topic? Want to ask a question, offer a suggestion, share your own programs and projects, upload a file to the file archives, get help with calculator and computer programming, or simply chat with like-minded coders and tech and calculator enthusiasts via the site-wide AJAX SAX widget? Registration for a free Cemetech account only takes a minute.

» Go to Registration page
Page 1 of 1
» All times are GMT - 5 Hours
 
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum

 

Advertisement