OK so here's my 2 cents as they say...... I've spent over 2 years involved with this company and have dealt with 2 of their products the Tronsmart MK908II and the Beelink R89. The issue is both of these could be amazing products..... as the specs suggest.
I've been a computer enthusiast for 32 years and generated an income from it for over 20 years and counting so this type of hardware to me at one time was pure fantasy and I am more than fully aware of what its capable of. Jeez the very first harddrive (hardcard) I had was only 10MB!!!!!! and I cut my teeth on basic as well as unix operating systems. So I think its safe to say I'm no NOOB but I never take a leetist attitude. I've been running XBMC on a Motorola Atrix 4G phone with HDMI out (Dual Core @1.2GHz & 1GB RAM) since some time in 2009.
I first dealt with the Tronsmart (netxeon) MK908II (Quad Core A9 @ 1.6GHz,,Quad core GPU, 2GB DDR3, 8GB NAND) which has basically more than enough features in the SOC (System On Chip) to provide the average person with amazing HD @1080P content... it should blow a Raspberry Pi B+ (ARM V7 @700MHz (single core), 512MB RAM) out of the water BUT it doesn't!!!! Then I dealt with the Beelink (netxeon) R89 (RK3288 Quad core @ 1.8GHz, 8 core GPU, 16GB NAND etc) which in theory should blow both the Raspberry Pi and the MK908II out of the water...........which it doesn't I can assure you.
So first off lets start with the biggest problem the MK908II has...... Initially it appeared fine, it may have stuttered from time to time but that's to be expected with early products (software issue surely!!!) and would be righted in time with updates. The updates came and went, not many I admit but the problems never went away. It took a total of 4 weeks before the MK908II became unusable no matter what firmware was on it, it would suffer random reboots, hangs or no boots and due to a family and increasing personal commitments I was unable to get to the bottom of this, I had my theories but needed considerable time to test these which was in very short supply. During the past couple of months I've since found that time.......
So what was the problem? The NAND.
They decided to use a NAND by a company called Foresee, now I can honestly say these NAND chips are terrible with read and write speeds below the competition but not so bad to have a really negative effect on the stick. So why so bad then? These chips are full of errors, bad sectors however you want to name them. So can I do a full low level format to start a fresh and hopefully let the built in error checking do its job and mark the sectors as unusable, thus having a chip that's not trying to write to bad blocks?................NO there is no Maskrom mode available on this NAND. How can I be certain of this? Because I contacted Foresee (Jeremiah Chou) about this very issue. They sent me a full apparently confidential spec sheet of the NAND and confirmed the NAND cant be low level formatted. Now this bit that follows is purely theory but I believe these NAND's are recycled (ever wondered why your NAND reports as Samsung brand when its Foresee and the labeling of the chip is hardly readable (re-branded?). I'm certain this is the issue of instability because I can have MK908II's running better than they ever have done and reliably by booting them from SD Card and all the problems are gone.....there may be the odd quirk but nothing more than you would get on any Android device classed as stable. I get consistent Antutu scores over 26000!!! and that's on a RK3188T chip.
The T chip doesn't bother me that much because I understand how the chips are made and that during chip production the speeds the chips can run at varies across the big silicone wafer / die that they are manufactured from. Dam many chip manufacturers eg Intel / AMD have even relabeled many higher speed chips to lower speed ones because they haven't had enough of the lower speed chips to meet demand. They cannot guarantee a certain number of chips to run at a certain speed until they are produced and tested, this then states how many 1GHz chips there are or how many 1.4GHz there are, I think you get my point now. A chip speed isn't really known until its been made and tested. It would cost the same to produce a 1.2GHz CPU as a 1.4GHz CPU they are made on the same platter of silicone. Also trust me a quad core running @1.4GHz is more than capable of anything you can throw at these sticks.
Now we come to our second product the Beelink R89
I was asked to join a beta testing program for this product and happily gave my details to Oman of Beelink (Netxeon) ( I place no blame on this individual at all, software guy). Admitted I was excited about this new product having read the specs of the RK3288 SOC and expected to have a few niggles.
I saw this as a perfect opportunity to help prevent Netxeon repeating their mistakes and making a truly fantastic Android box. So with that in mind was ready to get my hands dirty.
It took a while but finally they released pictures to us of the PCB and everything appeared great...........but wait a minute what was that I spied on the board.......no it cant be? the name on the NAND.........you guessed it Foresee. The instant I saw this I was straight on to Oman to explain how that was their greatest mistake to use this brand and also explained the reasons why. I let him know by private message instead of in the forum so as not to discourage people and this issue could be avoided. He even agreed that there was no way to format the Foresee NAND which I believe they have always been aware of. Still informing me they would not change the NAND after knowing these issues. He said they reboot them 1000+ times as tho that's some form of NAND stabiltiy test (not exactly a series of read and write cycles under load)
Anyway the R89 came, I opened it and it did indeed have a Foresee NAND...........how did it run..........even worse than the MK908II
This is where my interest and commitment quickly declined........ they wanted feed back from users yet declined to act on my advice (kinda of pointless taking part really if one of the most critical pieces of hardware is not just faulty but useless)
To cut a long story shorter I now have the R89 running stable and pretty sweet but only by booting it off SD card so no internal NAND usage. May I point out that so far the MK908II's out perform the R89's in terms of usage (not benchmarks)
Just to add a point the last batch of MK908IIs I ordered have since had the packaging changed to show an RK3188T chip and on inspection have Micron NAND's which finally makes the MK908II run as they should and also granting maskrom mode!!
Also the current Micron NAND MK908II's makes a great / fantastic Android device but look how long its took to get to this stage. So it is only now that this product has come close to maturity and its took me a hell of a lot of work.
As a final word I would say I would gladly help Netxeon in future but on one condition they listen to my responses
Laters
LawlessPPC
I've been a computer enthusiast for 32 years and generated an income from it for over 20 years and counting so this type of hardware to me at one time was pure fantasy and I am more than fully aware of what its capable of. Jeez the very first harddrive (hardcard) I had was only 10MB!!!!!! and I cut my teeth on basic as well as unix operating systems. So I think its safe to say I'm no NOOB but I never take a leetist attitude. I've been running XBMC on a Motorola Atrix 4G phone with HDMI out (Dual Core @1.2GHz & 1GB RAM) since some time in 2009.
I first dealt with the Tronsmart (netxeon) MK908II (Quad Core A9 @ 1.6GHz,,Quad core GPU, 2GB DDR3, 8GB NAND) which has basically more than enough features in the SOC (System On Chip) to provide the average person with amazing HD @1080P content... it should blow a Raspberry Pi B+ (ARM V7 @700MHz (single core), 512MB RAM) out of the water BUT it doesn't!!!! Then I dealt with the Beelink (netxeon) R89 (RK3288 Quad core @ 1.8GHz, 8 core GPU, 16GB NAND etc) which in theory should blow both the Raspberry Pi and the MK908II out of the water...........which it doesn't I can assure you.
So first off lets start with the biggest problem the MK908II has...... Initially it appeared fine, it may have stuttered from time to time but that's to be expected with early products (software issue surely!!!) and would be righted in time with updates. The updates came and went, not many I admit but the problems never went away. It took a total of 4 weeks before the MK908II became unusable no matter what firmware was on it, it would suffer random reboots, hangs or no boots and due to a family and increasing personal commitments I was unable to get to the bottom of this, I had my theories but needed considerable time to test these which was in very short supply. During the past couple of months I've since found that time.......
So what was the problem? The NAND.
They decided to use a NAND by a company called Foresee, now I can honestly say these NAND chips are terrible with read and write speeds below the competition but not so bad to have a really negative effect on the stick. So why so bad then? These chips are full of errors, bad sectors however you want to name them. So can I do a full low level format to start a fresh and hopefully let the built in error checking do its job and mark the sectors as unusable, thus having a chip that's not trying to write to bad blocks?................NO there is no Maskrom mode available on this NAND. How can I be certain of this? Because I contacted Foresee (Jeremiah Chou) about this very issue. They sent me a full apparently confidential spec sheet of the NAND and confirmed the NAND cant be low level formatted. Now this bit that follows is purely theory but I believe these NAND's are recycled (ever wondered why your NAND reports as Samsung brand when its Foresee and the labeling of the chip is hardly readable (re-branded?). I'm certain this is the issue of instability because I can have MK908II's running better than they ever have done and reliably by booting them from SD Card and all the problems are gone.....there may be the odd quirk but nothing more than you would get on any Android device classed as stable. I get consistent Antutu scores over 26000!!! and that's on a RK3188T chip.
The T chip doesn't bother me that much because I understand how the chips are made and that during chip production the speeds the chips can run at varies across the big silicone wafer / die that they are manufactured from. Dam many chip manufacturers eg Intel / AMD have even relabeled many higher speed chips to lower speed ones because they haven't had enough of the lower speed chips to meet demand. They cannot guarantee a certain number of chips to run at a certain speed until they are produced and tested, this then states how many 1GHz chips there are or how many 1.4GHz there are, I think you get my point now. A chip speed isn't really known until its been made and tested. It would cost the same to produce a 1.2GHz CPU as a 1.4GHz CPU they are made on the same platter of silicone. Also trust me a quad core running @1.4GHz is more than capable of anything you can throw at these sticks.
Now we come to our second product the Beelink R89
I was asked to join a beta testing program for this product and happily gave my details to Oman of Beelink (Netxeon) ( I place no blame on this individual at all, software guy). Admitted I was excited about this new product having read the specs of the RK3288 SOC and expected to have a few niggles.
I saw this as a perfect opportunity to help prevent Netxeon repeating their mistakes and making a truly fantastic Android box. So with that in mind was ready to get my hands dirty.
It took a while but finally they released pictures to us of the PCB and everything appeared great...........but wait a minute what was that I spied on the board.......no it cant be? the name on the NAND.........you guessed it Foresee. The instant I saw this I was straight on to Oman to explain how that was their greatest mistake to use this brand and also explained the reasons why. I let him know by private message instead of in the forum so as not to discourage people and this issue could be avoided. He even agreed that there was no way to format the Foresee NAND which I believe they have always been aware of. Still informing me they would not change the NAND after knowing these issues. He said they reboot them 1000+ times as tho that's some form of NAND stabiltiy test (not exactly a series of read and write cycles under load)
Anyway the R89 came, I opened it and it did indeed have a Foresee NAND...........how did it run..........even worse than the MK908II
This is where my interest and commitment quickly declined........ they wanted feed back from users yet declined to act on my advice (kinda of pointless taking part really if one of the most critical pieces of hardware is not just faulty but useless)
To cut a long story shorter I now have the R89 running stable and pretty sweet but only by booting it off SD card so no internal NAND usage. May I point out that so far the MK908II's out perform the R89's in terms of usage (not benchmarks)
Just to add a point the last batch of MK908IIs I ordered have since had the packaging changed to show an RK3188T chip and on inspection have Micron NAND's which finally makes the MK908II run as they should and also granting maskrom mode!!
Also the current Micron NAND MK908II's makes a great / fantastic Android device but look how long its took to get to this stage. So it is only now that this product has come close to maturity and its took me a hell of a lot of work.
As a final word I would say I would gladly help Netxeon in future but on one condition they listen to my responses
Laters
LawlessPPC
Comment