Header Ads

Mediatek MT2811, the new Bluetooth chip in Sony's Wireless headsets

Mediatek MT2811 Sony WF-1000XM3

Mediatek tweeted last week that their chip powers the WF-1000XM3, but we already knew this back in August in the teardown article from 52audio. What Mediatek did not provide was anymore details like specifications or a datasheet in the tweet.

So what is known about the MT2811? Trying to find any information on any English sites yields nothing, so I decided to do some investigating. To start of with what we already know, the MT2811 offers Bluetooth 5.0, has AAC and SBC support and the earbuds can work independently. It has been stated this might be Airoha's MCSync technology. This would make sense, the MT2811 was described as an Airoha chipset.

In an article on 52audio last month, Airoha confirmed that the WF-1000XM3 was using the AB155x platform, which has support for MCSync, so we can confirm that is how Sony was able to get the earbuds to work independently.

Airoha AB155x 52audio.com
Source: http://www.52audio.com/archives/27130.html
The slide above from 52audio has a quick overview of the chip and the features do appear to match with the WF-1000XM3 quite well, like voice assistant support, Hi-Res codec support and hybrid ANC support. Searching Airoha's site for the AB155x does not produce any results, the closest model is the AB1552x which looks to have identical specifications to the AB155x. So I feel it would be safe to assume the AB1552x is just one model in the AB155x line, possibly supported by another slide in the article where it shows the AB1552/155x grouped together. Likewise in that same article in the WF-1000XM3 slide (see slide), "MTK/Airoha AB155x advantages" appears, so this might support the idea of the MT2811 of being a rebranded AB155x.

One question I have seen come up in various videos and posts, why did Sony not just use Qualcomm's new Bluetooth chips, the QCC5100, specifically the QCC5121 which has many features Sony would want, like hybrid NC, small chip size, low power draw, and Qualcomm's Truewireless plus. Even I assumed back when the chip was announced that Sony would use it but perhaps there was some aspect of the chip that prevented Sony from using it.





The answer lies in an AV-Watch interview, Sony's engineers stated that using Qualcomm's new Bluetooth chipsets was not possible since in order to have both earbuds work independently, the device sending the signal (ie phone, walkman, etc) would need at least a Snapdragon 845 or 855. This is not possible in all scenarios, like in an iphone, laptop, or Sony's own Walkmans, so another solution was needed.

Sony WF-1000XXM3 Teardown
Source: 52audio.com
Enter Mediatek and their new MT2811 Bluetooth chip. Looking back at the FCC documents, the MT2811 was mentioned in the WF-1000XM3 reports back in late March, but I did not notice this. Once we knew the MT2811 existed after the teardown, noticing the chip in the FCC documents was much easier and I pointed it out in the WF-H800 post. I extracted out the specific part of the FCC document below for both models. Checking both the WI-1000XM2 and WH-H910N FCC documents, the MT2811 is not mentioned anywhere, so perhaps this software is only used for truly wireless headsets.

FCC Sony WF-1000XM3 WF-H800

Focusing back on the AB155x, as I mentioned Airoha bundled this with the AB1552x in that one slide, so perhaps this is the MT2811 in the WF-1000XM3? At 6.2mm x 6.2mm, I was not sure if it was the correct size, so I decided to see if I could find the actual size of the MT2811.

In the WI-1000XM2 FCC documents there is a photo of the PCB with the MT2811 and some measurements.  I cut the photo out and posted it below for reference, but a larger size can be see in the link above.
Sony WI-1000XM2 Teardown
Sony WI-1000XM2 PCB
I am not sure how accurate the measurements on the photo are but I wanted to see what results I would get. Apart from the 2 measurements, I decided to use the dimensions of the USB type-C port as well just in case the measurements from the photo were wrong, with the assumption that the size of the port would be standard but I discovered this was not the case.

Before I started, I first needed to know if the MT2811 was a square chip (the photo was not clear on this), and I found on another certification site schematics which shows the various chips and components. In the file 007-AH0156_01_001, the MT2811 looks to be a square, so with that out of the way, it was time to calculate. Using the 5.6 cm and 1.5cm lengths, I came up with 2 different sizes, so it looks like the measurements on the photo were not accurate. Next up was the USB type-C port. Using a width of 9.2mm, the estimate was about 5mm x 5mm, and using a length of 10.5mm, the estimate was 5.9mm x 5.9mm. This difference was puzzling but from other teardowns on 52audio.com, it seems that the length of the USB type-C port can be different, so this explains the difference. So the USB port was out as well.

There are 2 more options, the first using the WF-1000XM3 teardown from 52audio.com and using Q128FWY chip as a reference for size. And the second option is using the WF-H800 certification, which has the shapes and sizes of the chips from the WF-H800 with the Q128FWY again as the reference for the size.

Source
The Q128FWY is on the IC203 on the photo above and the ratio of length to width is 1.25. On one site, I found dimensions for the Q128FWY of 3.2mm x 4mm, which matches this ratio. Using those measurements, the MT2811 should be 4.8mm x 4.8mm when scaling to the chip.





Using the QN1e dimensions (3.4mm x 3.4mm), I was able to find out the width of the flat flex cable connecting the 2 boards in the WF-1000XM3, it came out to 5mm. Then using this as a reference for the second board, I was able to calculate the MT2811 to 4.8mm x 4.8mm, which matches the top estimate. Likewise based off the Q128FWY size on the WF-1000XM3, I get the same size as well. So I feel 4.8mm x 4.8mm is the correct size.

Photo originally from 52audio.com
So with 4.8mm x 4.8mm, this does not match the 6.2mm x 6.2mm size of the AB1552x but again thanks to a 52audio.com article (1) (2), we have a product brief for the AB155x below. We can see the different packages offered and right in the middle is the 4.82mm x 4.82mm package which matches my estimate, so it is safe to assume that we are looking at the same specs as the AB1552x.

AB155x product brief
source: 52audio.com
It is interesting that the AB155x is branded as the MT2811 in the WF-1000XM3; I am sure there is a perfectly good reason for this but I do not expect Mediatek/Airoha to reveal that. Yes I know Mediatek owns Airoha but then why not just have the AB155xin the Sony and not the MT2811. Perhaps in the end it does not matter, and OEMs can buy the chip as either an Airoha or Mediatek, since the specifications are the same regardless.

With the chip identified, I wanted to address one more issue I have seen/read online, and that Airoha chips are low cost and are only used in cheap Chinese Bluetooth headphones. I will agree, there are quite a lot of low end Chinese wireless headsets using Airoha's chips, especially the AB1526, AB1526P and the new AB1532. But I do not blame them, the new AB1532 has MCSync and this works with anything, unlike Qualcomm's solution which requires a high end Snapdragon 845/855 SoC. Besides I am sure the Qualcomm option was more expensive and every dollar counts when selling cheap wireless headsets.





I am not sure many Chinese OEMs will go for the high end AB155x option, every dollar counts and I would imagine with the specifications of the AB155x, it does not come cheap. With premium feature like 24-bit 192KHz HD Audio and Hi-resolution codec and Hybrid Active Noise Cancellation, I can see why Sony has been using this for all of their recent high end devices. So far, we are looking at the WF-1000XM3, WI-1000XM2, WF-H800, WH-H910N and WH-H810. I suspect Sony will also use this chip for the WH-1000XM4 once that comes out, but that is only speculation from the fact the WI-1000XM2 and WH-H910N have used it.

Here is the tweet btw.



2 comments:

  1. Good job on finding all those information on the MT2811. I was looking everywhere on google but I couldn't find any information on it. I just don't know how good it will be to use it on the XM4s. If Sony truly said that " Qualcomm's new Bluetooth chipsets was not possible since in order to have both earbuds work independently..." then why not use Qualcomm on the XM4 since XM4 cans don't need to work independently?
    I don't expect Sony to improve the sound quality on the XM4 but with this new chip that was not on the XM3, I don't know if the sound quality on the XM4 will be as good as the XM3...
    Then again, great job on doing all the work on this article.

    ReplyDelete
    Replies
    1. I suspect cost and ease of firmware updates. Sony can order a lot of the mediatek chips for all their new devices and save on cost since they are buying a lot more. Second, I noticed in many of the teardowns for the new devices a lot of parts are shared and using the same Bluetooth chip would help in development and updates to firmware.

      Delete

Powered by Blogger.