Dear ATMEL/MICROCHIP
The design/documentation of the serial coms link to the ATBTLC1000-MR110CA seems to be a total Dogs Dinner. From what I have read I don't think it's just me having trouble, is it?. (Readers if you agree please comment below).
I have designed a 4 layer PCB, for a client, which includes a SAMD21 micro, connected via UART to an ATBTLC1000-MR110CA. I have followed the schematic on page 15 of "Atmel-42514-ATBTLC1000-MR110CA-BLE-Module_Datasheet.pdf" which shows the following connections:
1) UART_RXD
2) UART_TXD
3) UART_CTS
4) UART_RTS
5) Wake
6) Chip_EN
So I have made these connections on my PCB. The datasheet also shows SPI connections but the BluSDK does not support SPI so I have not connected those pins to my SAMD21.
The datasheet calls this connection a 4 wire connection (even though it has 6 wires), which, I assume, means it is the correct connection for use with the BluSDK with the settings:
UART_FLOWCONTROL_6WIRE_MODE=false
UART_FLOWCONTROL_4WIRE_MODE=true
Q1) Please can someone confirm if this is correct?
According to Page 30 of "ATBTLC1000-BluSDK-v5.0-Release-Notes.pdf" The only allowed alternative BluSDK setting is:
UART_FLOWCONTROL_6WIRE_MODE=true
UART_FLOWCONTROL_4WIRE_MODE=false
Q2) What is the required connection between the SAMD21 and ATBTLC1000-MR110CA for this setting? Two UARTS? Connected how? Why is there not a schematic showing this in the datasheet?
Q3) Which of the above two connection methods is recommended going forward for a mass produced product?
With the 4 wire connection described above, I was able to work out conf_serialdrv.h (and conf_console.h) settings for my PCB layout and make progress with my software and get the battery service, Device Information Service and a couple of my own custom services working on this board as a BLE peripheral.
However, I started to get AT_BLE_TIMEOUT errors after the code had been running for a few minutes.
Q4) Does this sound like a UART coms problem? Or is there likely some other explanation, e.g. a bug in my code? Any ideas what it might be?
Annex 12 on Page 30-31 of "ATBTLC1000-BluSDK-v5.0-Release-Notes.pdf" says the following:
Q5) I can't make any sense of this. Can someone translate this into meaningful English sentences so people can understand what it is trying to say, please?
Q6) What is 2 wire mode? Is this another viable connection scheme not shown in the datasheet?
With the 4 wire settings above I stepped through the code and found that pf_cfg.bus_info.bus_flow_control_enabled is being set = false by the BluSDK functions.
Q7) Is this equivalent to "busConfig.bus_flow_control_enabled" in the text above?
However looking at the UART data with a logic analyser I see about 2.2 seconds of solid data download (at 115200 baud) on initialisation of the ATBTLC1000-MR110CA, which I take to be the patch download that the above text says is disabled when busConfig.bus_flow_control_enabled = false.
Q7) Can you explain why this seems to show the above-quoted text is nonsense.
Annex 12 on Page 32 of "ATBTLC1000-BluSDK-v5.0-Release-Notes.pdf" (the same document quoted above, which seems to be full of nonsense), describes a "BTLC1000 Hardware Flow control 4-Wire Mode E-Fuse Write Procedure" with a big red warning about it being irreversible. Since my board seemed to be working OK for a while, before the AT_BLE_TIMEOUT errors showed up, I am in two minds about whether I need to do this procedure on my ATBTLC1000-MR110CA or not. I am using the 4 wire Mode as I understand it.
Q8) Do I have to do the E-Fuse write procedure to correct the coms errors I am experiencing? Or does the level of functionality I have achieved show this is not needed (or has been done) and the problem is in my code?
I wondered if I needed to do the E-Fuse write procedure or if it was already set so I wired up my Atmel-ICE 3v3, GND, SWDIO & SWDCLK to the ATBTLC1000-MR110CA on my PCB.
I then ran the "EfuseBlockProgram.exe" to try to find if the fuse was set already or not. However, it did not find the ICE and I think it said that no Segger J-Link was connected.
Q9) Why was it looking for a Seger J-Link instead of an Atmel-ICE?
So I tried using Atmel Studio to access the Fuse settings via the Atmel-ICE but it would not connect to the ATBTLC1000-MR110CA device either, so I gave up until I get confirmation that I need to do this. Perhaps I have a connection problem.
Q10) The document mentions using a SAM-ICE but I only have an Atmel-ICE (a later model) If I need to set this fuse can I use my Atmel-ICE or do I have to buy another tool? If so which one?
Q11) If this product I am designing goes into production WILL WE HAVE TO DO THIS E-FUSE PROGRAMMING ON EVERY SINGLE UNIT USED IN PRODUCTION ?!?!?!
Q12) Should I use 6 wire mode to avoid haveing to do this?
Page 6 of "Atmel-42526-ATBTLC1000-BluSDK-Platform-Porting-Guide_USERGUIDE.pdf" says the following:
Q13) What hardware fix? When will this come into effect? Has it been done on the unit I have? How will this be handled?
As you can probably tell I am just about in a state of terminal confusion with all of this.
Q14) Why is, what should be a simple serial interface, being made so hard?
Any practical advice on all this will be wonderful.
Thanks.
