Chinese ATMega88 clone

[Stonent]

The 8051 is such a wonderful chip. Way ahead of its time.

And way behind modern times.

Almost every tutorial on the 8051 I see on youtube is hosted by someone from India or a nearby country.
So at least some people in India seem to think it's useful. I've had some interest in watching videos on it, but eventually I seem to exit out because I have to concentrate too hard on understanding the instructor.

[ataradov]

There is still obviously a huge market for those chips - instruction set is patent-free and there is plenty of tools.

But I'm not doing any designs on the architecture with a single accumulator register, it is just too much pain for no particular gain.

[Rick Law]

The 8051 is such a wonderful chip. Way ahead of its time.

And way behind modern times.

Almost every tutorial on the 8051 I see on youtube is hosted by someone from India or a nearby country.
So at least some people in India seem to think it's useful...
...

You know, between India and China, you are talking 1/2 the world's population.  I recall reading that India (or China, don't recall which) each year graduate more engineers than the entire engineering population of the USA.  So there is a lot of engineers out there and soon to be a lot more.

China may make "crap" now, so did the Japanese in the 50's, 60's, and perhaps even 70's.  Soon, perhaps, Chinese could become a must-learn for engineers....

[AndersAnd]

They have a comparison sheet on their website.

Internal 32MHz Calibrated Oscillator

It claims both MCUs has a 32 MHz internal oscillator.   ATmega88A only has a 128 kHz [edit: and a 8 MHz] Internal Oscillator. A 32 MHz oscillator wouldn't make sense on an ATmega only specified to 20 MHz anyway.

LGT8F88A is however specified to 32 MHz and the block diagram shows both a 32 MHz and a 32 kHz internal RC oscillator:

[andersm]

ATmega88A only has a 128 kHz Internal Oscillator.

It has two internal RC oscillators, one 128kHz and one 8MHz.

[AndersAnd]

ATmega88A only has a 128 kHz Internal Oscillator.

It has two internal RC oscillators, one 128kHz and one 8MHz.

Yes you're right, but not a 32 MHz oscillator like the comparison table shows.

[firewalker]

Have they issued AVRGCC patches for their mcu?

Alexander.

[ataradov]

It is fully compatible with all existing compilers. At least that's what they claim.

[AndersAnd]

LogicGreen has another MCU series named LGT8F0XA too, with an MVR8X 8-bit RISC core.
Anyone know if these are clones of other MCUs too and if so which ones? Are these ATtiny clones or something?
These are only specified up to 3.6 V though.
http://www.mcugreen.com/?page_id=4431

LGT8F08A – SOP28L
LGT8F04A – SOP24L/SSOP24L/SOP20L/SSOP20L
LGT8F02A – SOP14L
LGT8F01A – SOP8L

0 ~ 16MHz @1.8V ~ 3.0V
0 ~ 25MHz @3.0V ~ 3.6V

[ataradov]

They are AVR clones as well, but more sophisticated than ATTiny, it might be some earlier ATMegas or not a complete clone (peripheral-wise).

[ataradov]

They actually provide a header file for this part. I don't read Chinese, but it looks like it is a modified version of atmega164 with reduced memory.

[westfw]

I'm not doing any designs on the architecture with a single accumulator

Your choice, of course.  But that does does eliminate a LOT of 8bit microcontrollers!
No Freescale 8bit controllers.
No Microchip 8bit controllers.
No 8051 based controllers from anyone
No STM8.
I think it pretty much only leaves AVR and some 8080/Z80 derivatives (Zilog, Renesas) ?

(actually, I'd claim that learning a "single accumulator architecture" is pretty important if you're going to bother with 8bit microcontrollers at all (an arguable topic, of course), and that that a lot of the *interesting* architectural differences show up in the things that are done to entice embedded developers, like the bit processor in 8051, or the direct test instructions on "memory.")

What do you care, if you're programming in a HLL anyway?

[ataradov]

Yes, and I try avoid all mentioned MCUs. I've started with singe accumulator MCUs and I'm not sure if studying them can be of any benefit, except for broadening the knowledge, of course.

It also leaves ARM, which is my current first choice. I like to have free and open development tools that I can run on any platform I choose.

[andersm]

It is fully compatible with all existing compilers. At least that's what they claim.

A patch with the changed instruction costs may let the compiler generate better code for their chip.

[amyk]

A patch with the changed instruction costs may let the compiler generate better code for their chip.

I think there's something great about some unknown Chinese company being able to come up with an AVR-compatible core and making it faster the Atmel's. Almost like Intel vs AMD...?

FYI for the 8051 and related single-accumulator architectures (6502 comes to mind), the first 128 or so bytes of "memory" are to be treated as registers.

[ataradov]

There is plenty of open AVR cores out there, you don't need to invent anything.

"Faster" is also questionable. Just yesterday I ran Xmega chip at 48 MHz (32 MHz maximum frequency according to the datasheet). I just needed to test that clock for USB is ok. This does not mean that Xmega can run at 48 MHz over entire temperature range.

And Atmel is liable if parts do not work as claimed in the DS. Do you trust that some Chinese company, for which there is no contact information to be found, made all the test and can guarantee that parts actually perform according to the DS (which is only available in Chinese)?

[Rick Law]

A patch with the changed instruction costs may let the compiler generate better code for their chip.

I think there's something great about some unknown Chinese company being able to come up with an AVR-compatible core and making it faster the Atmel's. Almost like Intel vs AMD...?

FYI for the 8051 and related single-accumulator architectures (6502 comes to mind), the first 128 or so bytes of "memory" are to be treated as registers.

If memory serves, AMD was a "second source" manufacturer making the 486 (386?) under Intel license.  So they AMD not start from scratch.  There were however a few other manufacturers (Cyrix, VIA...)  I believe at least one of these two is Chinese owned.

[andersm]

"Faster" is also questionable. Just yesterday I ran Xmega chip at 48 MHz (32 MHz maximum frequency according to the datasheet). I just needed to test that clock for USB is ok. This does not mean that Xmega can run at 48 MHz over entire temperature range.

Several instructions execute in fewer clock cycles on the LGT8F88A, so clock-for-clock it is faster than the ATMega88

[Rick Law]

There is plenty of open AVR cores out there, you don't need to invent anything.

"Faster" is also questionable. Just yesterday I ran Xmega chip at 48 MHz (32 MHz maximum frequency according to the datasheet). I just needed to test that clock for USB is ok. This does not mean that Xmega can run at 48 MHz over entire temperature range.

And Atmel is liable if parts do not work as claimed in the DS. Do you trust that some Chinese company, for which there is contact information to be found, made all the test and can guarantee that parts actually perform according to the DS (which is only available in Chinese)?

Why not?  One just needs to know who is behind the guarantee.

If you own a laptop, chances are, (80-90% probability?) they are made by a Chinese manufacturer.  Got a microwave oven?  Chances are, they are made by a Chinese manufacturer.  So plenty of the Chinese manufacturers and designers seem to be doing an adequate or even a good job doing what they are/were contracted to do, and their US counterpart feels the stuff done is good enough that they have no problem standing behind their products made by their Chinese contractors.

Obviously, HP (or GE, or...) wouldn't just do a web-search to decide who would be their OEM manufacturer.  They choose carefully and the quality of the choice is reflected by the quality of their own decision making process.

There are good companies there as well as here (where ever here is), and there are bad companies there as well as here.  One just need to be careful and know the company (brand/dealer) one deals with.

eBay or AlliExpress permit a kind of business model that can be abused by crooks easier than other business models.  I would not judge all Chinese companies by the crooks and products one can find on web-retailers.  Unfortunately we being forum readers on the internet likely let internet experience weight too much in our judgement.

When I was working in New York City over 10 years ago, I could readily buy a "genuine Rolex" for $10 within blocks from my office.  I can also get camcorders, VCR, walkman tape players...  I am sure I will get what is coming if I spend $6 for a "genuine new Sony Walkman" and expected a real Sony Walkman.  Shopping in NYC then (and perhaps now) was no different than strolling down the web.  You got to do your homework - and if the deal seems to good to be true, it probably is too good to be true.

[ataradov]

Ah, that. That is also a characteristic of the available open cores, they often don't follow original timings and sometimes they are faster. We can give them that, but I don't really think it is important. You don't use AVRs for high performance applications.

[ataradov]

There are good companies there as well as here (where ever here is), and there are bad companies there as well as here.  One just need to be careful and know the company (brand/dealer) one deals with.

I'm not saying that all Chinese companies are bad, and I see that a lot of electronics comes from China.

But this particular one is bugging me more than any other. For one, it is not clear what the legal status of their product, we'll have to wait for Atmel's reaction.

I personally don't understand what exactly is patentable about MCU architecture/instruction set, but something is holding everyone else from making AVR-compatible chips. And I also know that there is something in ARMv3 and above that prevents open implementations, so there is something you can patent and I'd really like to hear an explanation of what it is exactly.

[firewalker]

Is there any source code/make files available for their clone. I wasn't able to locate an AVRGCC example.

I they are offering more options how can the compiler make use of those option without being patched for?

-mmcu=

Alexander.

[dannyf]

"-mmcu"

what did they say this chip is a clone of?

[ataradov]

I they are offering more options how can the compiler make use of those option without being patched for?

They are not offering anything extra, they just have some instructions that take less time to execute than their counterparts from the "prototype" device.  So just select mega88 and you will be fine.

[baljemmett]

If memory serves, AMD was a "second source" manufacturer making the 486 (386?) under Intel license.

Going back further than that, too -- I've seen AMD-fabbed 80186s, for instance.  Not sure about anything older than that, though.