Chinese ATMega88 clone

[Jope]

Itead Studio is currently running an Indiegogo campaign for their Arduino-like 'Iteaduino Lite' board.
The interesting thing is that they don't use an original Atmel AVR, but a Chinese clone of an ATmega88, the LGT8F88 from a company named LogicGreen.
They have a comparison sheet on their website.

[Svuppe]

Wow. They have removed 2 out of 3 gnd pins, and added 8 high sink port pins. I would expect massive ground bounce problems.

[dannyf]

Chinese clone

Bad for Atmel giving that a lot of their business comes from within China which is driving their success so far.

However, I am not sure to what extent volume buyers will embrace a clone. For the one-off market (aka arduino) i don't know who's willing to put up with the potential hustle of a new clone to save a few cents.

massive ground bounce problems.

Be real.

[miceuz]

I wander what's ADC performance with no separate power pins for ADC and switcher as the main power source...

[tszaboo]

They keep the cost low by welding the components on, not soldering:
"these pins will not be welded on it,"
I wonder why don't they have someone read it before writing these kind of mistakes. It is not like they cannot e-mail it to some Indian proofread company to correct it for 1 USD per page. Does anyone know if the memory size is useful? I've only programmed Atmegas in ASM.

[Svuppe]

Be real.

I am quite serious. 8 pins sinking up to 80 mA each, and everything has to exit through one single gnd pin. I wouldn't trust the ADC for anything, and I would have my doubts about other parts too.

[andersm]

I fully expect there to be all kinds of problems with the first iterations of the chip, but I'm actually kind of excited to see a real second source for a microcontroller on the market. The instruction timing differences means it's not a 100% drop-in replacement, but it should work for the majority of applications. I wonder what this will do to the counterfeit AVR market?

[westfw]

It's NOT a "real second source."  It's not quite compatible.  (in relatively interesting ways, actually.  Implying that it's not just a clone, either.  (particularly interesting is the lack of bootloader support, the odd eeprom size, and a different multiplier.)

The single set of power pins IS rather troubling...  Maybe you can do that if you implement the core in smaller geometries, and just dedicate extra transistors to the output drivers?

[dannyf]

and everything has to exit through one single gnd pin.

Sure. But none of that necessarily leads to "massive ground bounce problems".

I wouldn't trust the ADC for anything, and I would have my doubts about other parts too.

You are obviously free to speculate. But unless you have hard data, it remains your speculation.

[dannyf]

instruction timing differences means it's not a 100% drop-in replacement

It is instruction compatible with avrs. I doubt there are a lot of code out there that take (dis)advantage of original avr's timing to the extent to break its execution on this clone.

[MatCat]

I especially love where it says only LogicGreen's ISP will program it... I think I would rather stick with Atmel for now.

[baoshi]

Their ISP is a VUSB based board, perhaps a USBISP clone. You can take a look here, last 3 items are Emulator, Programmer and Demo board.

http://www.mailshop.cn/product_list.jsp?class_id=3640&super_id=1024

I heard the dev platform is AVRGCC??!

[dannyf]

I heard the dev platform is AVRGCC??!

Yes, according to their own comparison, all toolchains for avr are good for this clone - how that's going to sustain in a legal challenge is unknown.

[baoshi]

The company does not have an address, contact number is a mobile, support is on an forum. If I were Atmel I'll go for their customer (if any ).


Sent from my iPhone using Tapatalk

[Rick Law]

Anytime there is more competition, the consumers benefit.  As long as it is a legitimately done clone, I think it is a good thing for the industry.  I am sure there will be tons of issues to iron out.  This may be their learning process.

I would not find the clone they have depicted as interesting.  When sold installed in a board, it diluted whatever price-advantage their CPU offers.  Beside, I wouldn't want a board that I cannot replace a blown CPU.

Personally, I prefer to see a dip drop in replacement that enhances the original - like the way Intel (and competitor) CPUs were in the early x86 era.  Motherboard permitting, one can drop in a 90MHz and replace the 66MHz P4/P5 and off you go with bigger/better system.  Say, with this enhance CPU drop in, you have 2K of ram instead of 1, or 48K of flash instead of 32K, or a 12 bit ADC instead of 10...  That approach may take the market by storm.  Those with a Arduino board may find it a lot easier to buy another chip to plug into their existing board then purchasing yet another.  With that, they can builds experience for their designers, market understanding, and builds name recognition for their product (AMD, etc vs Intel) along the way.  Eventually they would earn enough name recognition that they can sell a CPU entirely of their own design instead of a work-alike of someone elses' design.

[andersm]

I heard the dev platform is AVRGCC??!

Yes, according to their own comparison, all toolchains for avr are good for this clone - how that's going to sustain in a legal challenge is unknown.

It at least used to be the case that you couldn't copyright an instruction set, although parts could be covered by patents.

[Svuppe]

Sure. But none of that necessarily leads to "massive ground bounce problems".

Well, perhaps "massive" was too harsh a word.
And after reading the datasheet, I see that the intention is to use the 80 mA pins for multiplexing common-cathode LED displays, which means only one pin sinking the high current at any given time. Not all 8 pins with 80 mA each. That is something different altogether.
Unfortunately, the datasheet does not specify the total current allowed for the entire chip.

[dannyf]

That approach may take the market by storm.

Given how mcus are used, that kind of market and demand must be tiny.

[poorchava]

but I'm actually kind of excited to see a real second source for a microcontroller on the market

This is not very prevalent in retail sale, but when you go into volume production, such things happen quite often. Granted - the different sources are almost never 100% compatible (99% would be my guess), but they are usable.

[Rick Law]

That approach may take the market by storm.

Given how mcus are used, that kind of market and demand must be tiny.

Yeah, I suppose you are right.  I was thinking merely of the experimenter market rather than the whole market.  Consumers isn't likely going to choose a Microwave oven based on the MCU they use.

[westfw]

I would expect that the chip is aimed at whatever market is responsible for all those "ATmega8 for $1" China gray-market deals you see on eBay (but I don't know what those are!)

[amyk]

Not surprising, Holtek made PIC clones before so I guess the next logical step is an AVR clone.

But the Chinese still seem to prefer 8051 clones (AFAIK Intel doesn't make them anymore) for "general purpose" MCU.

[dannyf]

But the Chinese still seem to prefer 8051 clones

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

[ataradov]

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

And way behind modern times.

[dannyf]

Sure, if you let mkting think for you.