Teensy 4.0 First Beta Test

Status
Not open for further replies.
@Paul (and others) One thing I noticed with both this Beta 2 board as well as Beta 1 (at least in the adapter boards). If I have the something like a USB to Serial adapter plugged into lets say Serial4 connection (using CP2104 adapter). And then plug in the T4 into USB, the board is not detected by PC... Not sure if it booted or not, but no sounds showing Serial port plugged in... But if the adapter is not connected comes up fine.
My guess is it has to do with the voltage on the RX pin back feeding?

I've not seen this problem, and I have done a *lot* of testing with a FTDI TTL-232R-3V3 cable connected to Serial4. But I'm using Linux (Ubuntu 18.04.2 x86-64) and I have both the FTDI cable and the T4 plugged into a USB2 hub.

Maybe I need to get one of those CP2104 adaptors? Which one are you using? Any idea if this happens with all OS, or only Windows?
 
I've not seen this problem, and I have done a *lot* of testing with a FTDI TTL-232R-3V3 cable connected to Serial4. But I'm using Linux (Ubuntu 18.04.2 x86-64) and I have both the FTDI cable and the T4 plugged into a USB2 hub.

Maybe I need to get one of those CP2104 adaptors? Which one are you using? Any idea if this happens with all OS, or only Windows?

Please see post #2373 - looks like an issue I posted some time back - using extra Teensy's for UART<>USB on Windows:
in Post #6 :: teensy_ports issues with 3+ boards, #1330, #1398,

That sounds like behavior I was seeing many weeks back? Where I had to have other Teensy's OFFLINE in order for the T4-1 so show up. Those posts suggest it failed to start … @KurtE put a blink to see if it is running to confirm similarity
 
Opps, I probably should have made one with all 10 pins for you. They were all done the same and soldering extra pins times 9 boards was time I didn't have. :(

I'll send you a package today with extra pogo pins and another bare PCB. Some assembly required. That's the most I can do at this moment. If you're still needing Serial6 & Serial7 hardware, ping me in a week or so and I'll get another board soldered up for you.
Sounds good , hopefully enough pins will be accessible to test out SPI1
 
@defragster
I will take a look but it seems to be reading about 15-20deg higher than what I am reading with my temp gun. Keep checking the code but the only thing I come down to is the calibration data that is set in one of the fuses. It seems to be reading it ok but...

Heres' the quote from the manual
The temperature calibration point fuse values are available in the OCOTP_ANA1 register.
All I am doing is:
Code:
uint32_t calibrationData = HW_OCOTP_ANA1;
I think that is all I need to read the fuse.
 
@defragster
I will take a look but it seems to be reading about 15-20deg higher than what I am reading with my temp gun. Keep checking the code but the only thing I come down to is the calibration data that is set in one of the fuses. It seems to be reading it ok but...

Heres' the quote from the manual
All I am doing is:
Code:
uint32_t calibrationData = HW_OCOTP_ANA1;
I think that is all I need to read the fuse.

Yeah - about 20 deg C diff.

Just started it on T4-1 to see if anything changed in move to 1062. Oddly it is failing … stuck in my cooling loop() where I have :: asm volatile ("wfi");

I commented that out on T4-1 and it is now running - that runs fine on T4-2 ???? ? ? ? ?

My first test with 'wfi' early in T4-1 was odd - duplicate interrupts and odd behavior … but it worked as expected when I threw it into temp testing and temp dropped 5 deg C in the one doing no real code but waiting.

Will let T4-1 cycle and heat soak as it was a room temp. The first pass completed with max temp of 50.5C … back to update shortly … opps restarted - that was the 1306 version … have to remove 'wfi' again.
 
I've not seen this problem, and I have done a *lot* of testing with a FTDI TTL-232R-3V3 cable connected to Serial4. But I'm using Linux (Ubuntu 18.04.2 x86-64) and I have both the FTDI cable and the T4 plugged into a USB2 hub.

Maybe I need to get one of those CP2104 adaptors? Which one are you using? Any idea if this happens with all OS, or only Windows?

I am using the Adafruit CP2104 adapter used for USB Host problems earlier)...

@defragster - I am pretty sure blinking LED won't work. When I plug it in, nothing happens. No debug output on Serial4, Serial port not created on Windows, and pressing program button does nothing

Pressing the power button for about 10-15 seconds appears to turn off power, pressing it again, appears to turn it back on, but still nothing else.

Windows 10, I think both are plugged into an USB3 hub
 
@mjs513 - etc - CoreMark w/TempMon results T4-1 .vs. T4-2 odd results?

> 'wfi' stalls sketch on T4-1 - during cooling loop
> T4-1 showing 'Errors detected' once soon after shift to 804 MHz in first 10 secs {one out of 194 passes in 30 seconds} : 4 out of 4 times and NONE on T4-2
> Both units completing similar number of CoreMark passes of 300 iterations
> Temps below are results at 600 MHz - with same code running - sitting side by side
--> The T4-1 does have larger PCB for larger heat dissipation - both are mounted on BreakOut board.
T4-1:: THERMO read at 800 MHz 38.3 C
F_CPU MHz 600 Sec=10 CM#=53 [Max C=47.897198] AFTER :: °C=47.897198
F_CPU MHz 600 Sec=20 CM#=106 [Max C=48.551403] AFTER :: °C=48.551403
F_CPU MHz 600 Sec=30 CM#=159 [Max C=49.205608] AFTER :: °C=49.205608
// and 800 MHz
F_CPU MHz 804 Sec=30 CM#=194 [Max C=54.439251] AFTER :: °C=54.439251
25 sec Cooling Period ... °C=53.785046 After Cool :: °C=44.626167
T4-2:: THERMO read at 800 MHz 41.4 C
F_CPU MHz 600 Sec=10 CM#=54 [Max C=56.043480] AFTER :: °C=56.043480
F_CPU MHz 600 Sec=20 CM#=107 [Max C=56.652172] AFTER :: °C=56.652172
F_CPU MHz 600 Sec=30 CM#=160 [Max C=56.652172] AFTER :: °C=56.652172
// and 800 MHz
F_CPU MHz 804 Sec=30 CM#=194 [Max C=62.130436] AFTER :: °C=62.130436
25 sec Cooling Period ... °C=60.913044 After Cool :: °C=53.000000
 
I am using the Adafruit CP2104 adapter used for USB Host problems earlier)...

@defragster - I am pretty sure blinking LED won't work. When I plug it in, nothing happens. No debug output on Serial4, Serial port not created on Windows, and pressing program button does nothing

Pressing the power button for about 10-15 seconds appears to turn off power, pressing it again, appears to turn it back on, but still nothing else.

Windows 10, I think both are plugged into an USB3 hub

Seems like a repro of the prior posted issues I had. Not sure if TeensyLoader verbose log will show anything - I think I posted them in mine 1K posts back. Quick Next question(s) - what happens if TeensyLoader and or IDE is closed when you repeat? Something is triggering the T4 on startup - that's why I was looking to see updated USB Stack - but going to Sublime compile with Compile.cmd/TyComm and no IDE and PJRC tools it went away. Having many online with IDE would show some go away oddly from list - TyComm didn't show that.

<edit> @KurtE - what is your added SerMon program? I should try to repro - and be sure it isn't TyComm that was causing my issue. Is TyComm active on your system?

See post 2391
 
Last edited:
About to connect T4-2 Serial Updated Post #3

BETA2 T4 1062::
Code:
        UART6 Serial1 [0/1]   UART4 Serial2 [7/8]
        UART2 Serial3 [15/14] UART3 Serial4 [16/17]
        UART8 Serial5 [21/20] UART1 Serial6 [25/24] 
        UART7 Serial7 [28/29]

Found this Ref for 1062 mapping - post 1917

Based on Paul's post 1909

Not seen @KurtE's latest spreadsheet. If above post 1917 table is correct that needs to get back to Post 3.
 
I can't seem to reproduce the I2C problem. Have tried many things. Here's what I see when I run Frank's modified code, using the Synthesis > Guitar example.

sc.png

I spent some time looking at the waveforms too. As nearly as I can tell, everything looks good.

file.png

But the audio definitely is not working. I hear strange noises and some static, but not even a hint of the guitar sound. :(

Out for time for now. Will pick this back up in several hours. Very mysterious problem....
 
I can't seem to reproduce the I2C problem.
....

I got i2c SSD1306 working no issues when I got usable libraries [_GFX , SSD1306] it ran their sample Adafruit demo sketch using Wire.h.

I just got my Prop Shield out and removed EEPROM calls and the CalibrateSensors is running with copious data- Assume that is a second test of i2c?
NXP_CAL_t4b2.jpg
 
About to connect T4-2 Serial Updated Post #3

BETA2 T4 1062::
Code:
        UART6 Serial1 [0/1]   UART4 Serial2 [7/8]
        UART2 Serial3 [15/14] UART3 Serial4 [16/17]
        UART8 Serial5 [21/20] UART1 Serial6 [25/24] 
        UART7 Serial7 [28/29]

Found this Ref for 1062 mapping - post 1917

Based on Paul's post 1909

Not seen @KurtE's latest spreadsheet. If above post 1917 table is correct that needs to get back to Post 3.
@defragster what should be updated? Serial8 is no longer valid as the two pins on bottom that had the Hart was replaced with the can pins.

I have updates which I should PR, for update Serial2 and remove Serial8 for1062...

@Paul I did a PR: https://github.com/PaulStoffregen/cores/pull/365
But it has some conflicts with your previous one... Same changes then more additions
Let me know if you need me to make a new PR ...
 
@defragster what should be updated? Serial8 is no longer valid as the two pins on bottom that had the Hart was replaced with the can pins.
...

I posted that for Serial T4-2 above T4-1, what will need updated I think is the 1052 versus 1062 layout in that Page 1 table post.

The linked Table in FrankB's post if it matches your XLSheet could be copied to that post I think. I'll move it there and then it can be reviewed for edits/adds as needed.

<EDIT> Copied into place
 
No Start T4 with Serial Teensys on UART<>USB

@PAUL - the below is truely interesting - but first I just saw this:

Standing OVER the T4-2 - thanks to the GREEN Power LED on the T4-2 Breakout Board::
Using a subset of that BELOW as follows:
When the Unpowered T_4-2 Serial#1 Rx line is connected to the Powered T_3.1 Serial#1 Tx the Green Breakout Led GLOWS DIMLY? - with common GND connected.
If I put the T_3.1 into Bootloader - the Glow stops and the T4-2 powers up normally
I don't know where the BB Green LED gets fed from and what that means to T4 startup ... pretty sure you can take it from there and Debunk or Resolve …

**Note: I put the T4-1 on [with and without] Breakout and same connections (GND, Tx>Rx) from that T_3.1 don't create this behavior

--- THIS IS THE BELOW-----
Got this set up again looking like before on T4-1 (not using the same single hub initially) and like KurtE AFAIK.

NOTE: I did not reboot Windows though:: IDE, TeensyLoader and TyComm are closed and TaskMan shows no orphaned processes

I connected - on the same powered HUB:
> T_3.1 to T4b2 Serial #1 :: Rx, Tx and GND
> T_3.6 to it's Serial #4 :: Tx and GND
>> T4 stalls/hung on USB Pluginstart - or Power On after Off - the Green light does come on, but no T4. And Program Button does nothing to the Red Bootloader LED.

>remove T_3.6 Power - T4 Still fails to power on with T_3.1 powered
>remove T_3.1 Power - T4 powers on with T_3.6 powered and arrives at USB

 
Last edited:
I'm still having SD problems with T4 beta1.

Solved:


Problem was caused by Kingston uSD. Used a SanDisk uSD and it worked fine on T4 beta1. For whatever reason, T3.2 could read the Kingston uSD just fine, and T3.6 could read the Kingston uSD both from on-board uSD slot and from sparkfun SPI SD breakout. Maybe T4 SPI SS pin toggling is too fast for Kingston?
 
Last edited:

Solved:


Problem was caused by Kingston uSD. Used a SanDisk uSD and it worked fine on T4 beta1. For whatever reason, T3.2 could read the Kingston uSD just fine.

Glad you got it solved - I tend to always use SanDisk uSD's so have not run into that one before.
 
@KurtE

I just pushed an update to the ILI9488 library to use SPI1/SPI2 on the T4B2 that works. EDIT: Just remember to specify the MISO, MOSI, SCK pins for SPI1/SPI2 then it works perfectly.
 
For whatever reason, T3.2 could read the Kingston uSD just fine.

How long were the wires? Was ground routed close to the clock?

I'm continuing to look into the audio problems. It's really starting to look like a signal quality issue. Turns out the 1052 chip on beta1 has the exact same problem as 1062 on beta2 when I replace the 5 mux chips with wires.

I soldered another board just now, using 33 ohm resistors in place of the mux chips. It's drying in the oven now, will test with it in ~15 min.

Edit: beta1 audio works with the 33 ohm resistors...
 
Last edited:
How long were the wires? Was ground routed close to the clock?

The initial test was on the T4 beta1 breakout + audio shield, so there were "no wires" and T4 beta1 was failing with Kingston uSD. For later tests there were long wires (12") to the pins on the back of the audio shield. There were shorter (4") jumpers to the sparkfun SD adapter. The SanDisk worked with T4 even with the long wires to the back of the audio shield. FWIW, with the long-wire connections to shield and T3.2, I occasionally had failures with Kingston too.
 
How long were the wires? Was ground routed close to the clock?

I'm continuing to look into the audio problems. It's really starting to look like a signal quality issue. Turns out the 1052 chip on beta1 has the exact same problem as 1062 on beta2 when I replace the 5 mux chips with wires.

I soldered another board just now, using 33 ohm resistors in place of the mux chips. It's drying in the oven now, will test with it in ~15 min.

Edit: beta1 audio works with the 33 ohm resistors...

Well think its good you seem to isolated the problem but... Can this be fixed by changing the pad configs or is this going to require a hardware change of some sort?
 
I've cut the I2S traces to the serial pins with a sharp knife - the audio works much better now, at least for the first seconds. Then, the noise comes back.

I've designed a similar board for a friend - it uses a Teensy 3.2 and the audioshield - this works good. The 3.2 is a little bit nearer to the shield, and the traces are some cm shorter.

The I2C problem remains.. sometimes I get all 1s too - if I disconnect the usb and reconnect it, I get some 0s or it gets stuck during initialization. I'll try an other audioshield.
Edit:Nope, I2C much better now!
 
Last edited:
Looks like the problem is only with MCLK. I tested many resistors. 22 ohms to 4.7K seem to work fine. 10 ohms or less definitely does not work.

To make this change on the new board, I first cut the MCLK trace.

mod1.jpg

Then I tinned the bare copper and soldered a 100 ohm resistor across the gap.

mod2.jpg

Unfortunately this appears to be the only way to get the audio shield to work with the beta #2.
 
Status
Not open for further replies.
Back
Top