cancel
Showing results for 
Search instead for 
Did you mean: 

WICED Studio Wi-Fi Combo

Anonymous
Not applicable

Does anyone test the BCM43362 WIFI-direct data throughput?

0 Likes
1 Solution
Anonymous
Not applicable

It's unreachable.

BCM43362 supports 72.2Mbps as maximum rate.

So we can expect 30~40Mbps on TCP level. But it's not 64Mbps.

By the way, how do you handle the big data on STM32 MCU?

View solution in original post

0 Likes
15 Replies
Anonymous
Not applicable

When WICED runs as a GO, up 10.1Mbps/down 13.8Mbps on TCP is my result on my desk.

It's not in shield room but it's enough for WICED I think.

Is there any problem on your test result?

0 Likes
Anonymous
Not applicable

Is this the application level(based on TCP) throughtput or MAC layer?

0 Likes
Anonymous
Not applicable

It's on TCP.

What is your problem / requirement?

0 Likes
Anonymous
Not applicable

Our user sienario is to transfer big data from Smart Phone to Device(based on BCM43362). For good user expirence, we hope to transfter 8MB/s application level data, that's 64Mbps, how can we achive this?

0 Likes
Anonymous
Not applicable

It's unreachable.

BCM43362 supports 72.2Mbps as maximum rate.

So we can expect 30~40Mbps on TCP level. But it's not 64Mbps.

By the way, how do you handle the big data on STM32 MCU?

View solution in original post

0 Likes
Anonymous
Not applicable

We use Cortex-M4 core which can up to 168MHz, so it's not a problem to handle this big data, we just save the data onto eMMC storage.

People said WIFI-Direct can achive higher throughtput, can we use BCM43362 for that mode?

0 Likes
Anonymous
Not applicable

As I wrote, expected maximum throughput is a half of what you need.

You should consider data compression or some other way.

0 Likes
Anonymous
Not applicable

By using UDP, do you think what's the maximum throughtput? Because the envirement is very clean, the distance is only 1 meter, maybe UDP can be ok.

0 Likes
Anonymous
Not applicable

UDP doesn't change the result much.

STM32F4 supports up to 48MHz@SDIO2.0. It's also one of the bottleneck.

0 Likes
Anonymous
Not applicable

According to some paper, UDP could be 3x faster than TCP when the network is very well.

STM32F4, 48MHz is the clock rate, the data width is 4bit/8bit, so SDIO is of course not the bottleneck.

0 Likes
Anonymous
Not applicable

Throughput of 1x1 11n (20MHz) will be 40~50Mbps as maximum.

It depends on IEEE802.11 MAC and can't be improved.

SDIO48MHz should be a bottleneck. You can test this by changing clock speed.

0 Likes
Anonymous
Not applicable

Before, I test SDIO with 48MHz to write eMMC, the speed is 7.6MB/s, that's 60.8Mbps for application level, I don't test it for WIFI chip, because you know I am not famillar with the protocal, I  need the specification for BCM43362, by this spec, it should describe how to program with BCM43362 SDIO interface, Broadcomm did some driver for it, but the driver is for Linux or others OS, I need a pure low level specification, similar to Bluetooth HCI command Spec, is there this spec for BCM43362?


0 Likes
Anonymous
Not applicable

Anyway, BCM43362(WICED) doesn't have such high throughput (64Mbps).

Could you consider compressing data?

0 Likes
Anonymous
Not applicable

I am closing this thread,
if your problem still persist then it is advisable to create a new thread.

Thanks

0 Likes
Anonymous
Not applicable

What is the throughput on upper MAC layer ?

Thanks

0 Likes