hardware design with tms570ls3137 pge

Yong,

Did you check the state of the nPORRST pins and nRST pins on the device to make sure that they are actually released?

How about the crystal you are using.  Did you confirm that it is running?

When you mention JTAG does not connect

  - Which emulator are you using?

  - Where does it fail?   Did you try the "Test Connection" in the Target Configuration Window (Advanced Tab) first?

-Anthony

Hi,Anthony

Thanks for replying.

The nRST pin is low when the power on, so the warmreset button seems no use.The crystal is working(16M) through the oscilloscope. The Jtag is just like the arm 20pin in the HDK,and the emulator is j-link.The "Test Connection" button is gray in CCS,so I cannot click it. I just click the debug button and the CCS says that unable to connect.

Now, I want to weld a minimum system board, absolutely, the source of 3.3v and 1.2v is essential. What about the crystal? Does it essential too for the minimum system of the TMS570LS3137? What else is in need about the minimum system board?

Yong

Hi,

Now,I build the minmium system of the TMS570LS3137, the nRST pin and the nPORST pin now has high voltage, but the initial state of the nERROR pin is low(the error LED lights,the schematic circuit is just like the HDK schematic), I don't know whether it's correct.

The worse is that the JTag still cannot link with the PC through the Segger Jlink.

Can you TI guys give any suggestion to find out the problem?

Yong.

Yong,

The Crystal isn't essential, but JTAG gets trickier if you clock the part at a low frequency <<1MHz like you would if the part fell back to the on-chip clock.

You might have trouble with JTAG in this case if you don't have adaptive clocking setup.  Without adaptive clocking, the rule of thumb is that the part needs to be running 8x faster than the JTAG clock.   So if JTAG clock is even 1MHz you need the part running at least 8MHZ.   that's why I asked.  But if you see 16MHz anyway you can rule that out.

Do you see the nTRST pin being pulled up when you try to connect with the Segger?   I think there might be something different about how nTRST is driven.   We need to have a pulldown on nTRST at some time for the board to start up but then nTRST needs to go high before you can connect the emulator.  Emulators sometimes are different in whether they drive this pin or they just have an open collector output and need a pullup on the board...

Other than that I don't know what to say without seeing your schematic.

-Anthony

Thanks for answering, the nTRST pin is always low whenever I try to connect with the Segger.

the HDK board with the arm 20pin interface. So I don't know how to do next now. The picture below is the arm 20pin designed, the arm 20pin is DIP20 foot

print, The emulator just connect it through the JTag flat cable.

By the way, I want to know the initial out state of the nERROR pin , is it low or just depends on the MCU's (may be  different with the different states of other pins)?

yong

Yong,

You probably need a pullup on nTRST then;  or it might be some other pins that need pullups.

If you are using a J-LINK then the segger documentation:  http://www.segger.com/admin/uploads/productDocs/UM08001_JLinkARM.pdf  has a table of the pinout in chapter 10 and it mentions which pins need pullups on the target board.

Just note that with the TMS570, nTRST needs to be asserted once after every power on reset or the on-chip emulation logic may not be reset.  So it would be best if you made your board so that the pullup is only there when the segger is connected.  Maybe put it on an adapter between the header and the segger...

 EDIT: / ADDITION.

Sorry I didn't answer the nERROR question.  nERROR shouldn't be asserted but since the contents of the 570 are not known / valid, it's likely you will get nERROR asserted until you have a good program in the flash.  ie. this should go away once you program valid code... but you have to obviously get JTAG connected first before you do this.

nERROR by itself won't prevent JTAG from connecting.  nPORRST would,  nRST would make it hard to get control of the CPU.

Hi,

I've tried to pull up the TMS, TDI,TCK,nTRST pins to 3.3 voltage through a 10k risistor, and added pull down of the TCK through a RC(risistor and cap) circuit, just like the picture below.

When I power on the TMS570, the pin voltage of the TDI and TMS is 3.3v, while the nTRST and the TCK is 2.4v, I don't know what maks that happen.

I've referenced the documention of the segger jlink. It suggested that the nTRST and the TMS should be pulled up, the other pins should be in a defined state.But now the TCK and nTRST pins' voltage is neighter high nor low, does it possible the TMS570 output a voltage makes that?

When the chip is powered first time, does it have any pin which can promote a square wave automaticlly? I'll check whether the chip works fine.

yong

Yong,

The TMS570 has internal pullups/pulldowns on various pins.   If you are getting 2.4v then likely your external pull is not strong enough to override the internal pull.

In fact I think we recommend about 2K to override the internal pull.   10K is too weak.

No square wave that I'm aware of...

-Anthony

Hi,Anthony

I've tried to measure the voltage of the HDK's TDI, TMS, TCK and the nTRST pins, I found that the voltage in the HDK is just the same with mine. When the JTAG is not connected to the PC, the nTRST and the TCK pins' voltage are about 2.4v(the HDK's is 2.2v),the TMS and TDI pins are 3.3v, while, they're just the same with the HDK board, and they are still same when the USB of the Jlink connected to the PC. The difference is just that the HDK board can connect the PC through the Jlink and mine not.

I've checked the reset circuit and it's working fine(the voltage of the reset is low when the button is pushed), when the nPORST pin is low(cold reset button is pushed), the nError pin of the chip outputs a 1.6v ,while the HDK board is 3.3v when nPORST pin is low. I don't know why.

The R225, R226 and R132 ,C29 on page 7 of the TMS570 HDK schematic need to be placed close to the TMS570, does it matters a lot? In my board, it seems not so close.

I've tried some IO pins' voltage of the TMS570(CAN1RX,CAN2RX, CAN3RX,CAN3TX,MII_RXD0...),the default state of these pins are just the same with the "TMS570LS3137 16/32-Bit RISC Flash Microcontroller" mentioned about, except the CAN3RX and CAN3TX pins(the datasheet says these two pins' default state are pulled up, while actually they are low). Does it mean that the TMS570 chip is not working or may be have some problem? How can I make sure that the chip is working properly?

The crystal's input and output waveform are just the same with the HDK board. So, I think the crystal is working properly.

Great need of your help! 

yong

One more to add is that I don't use the every pin of the TMS57LS3137-PGE, the unused pins are left float,so does it  matter the 570 MCU‘s proper working?

Do your unused pins have internal pullups / pulldowns?   if so they will not be floating, they will be pulled up or pulled down.

If not then you probably would want to add a pu/pd just to keep the current lower;  floating inputs can cause higher current consumption.

Alternatively you could program the pins that are 'floating' to be outputs but this will not be effective during the time the part is in reset.

Eh.. I've checked the unused pins, most of them have internal pu/pd, some of the unused pins like the MibACD module are floating. What I mean is that do these floating pins may lead to the MCU's proper working? Now, the MCU cannot connect the Jlink.

I mentioned above about the CAN3TX and CAN3RX default state are not like the said on the datasheet. And the nERROR pin outputs voltage is 1.6v when the nPORST pin is low. Do these mentioned above can make sure that the 570MCU is working fine or not?

I've tried every possible way to make it work, but still failed. I don't know how to do it next. Do you have any suggestions?

yong

Yong,

MibADC should not stop the MCU from connecting to JTAG, and these pins do not go to a normal input buffer so you won't have the current problem anyway.

Error pin sounds normal,  from  TMS570LS31x/21x 16/32-Bit RISC Flash Microcontroller Technical Reference Manual (SPNU499):

CAN pins - can't tell without knowing if your observation was made during reset or after the part was released from reset.

How many boards / chips have you tried?  It is possible to damage chips by various mechanisms.

Do you have an of our development boards to compare against?

I have the TMS570LS31 HDK board, but it's have programe in it, so it's not the default state. The error pin is 3.3v when the nPORST pin is low, but what I got in my board is 1.6v, when relase the nPORST, the error pin turns low, this is just the same with the HDK board.

The can3 pins goes high when the nPORST is low, and go high when release them, while the can1,2 pins are not, they are always high whether the nPORST pin is low or not.

I've just try one chip and board, the samples I got is few, so I want to know whether there are any other mistake on the board. If no mistakes, the only way is to try a new board.

yong

Hi Yong,

Can you connect to the TMS570LS31 HDK board with the same emulator that you are trying to use to connect to your board?

If so, then something is probably wrong on your board.  Did you assemble the board by hand?  Have you double checked all the solder joints, etc.

If you have a hi-res photo you can upload I'll look at it.   Not sure what else to do though.

I guess the last question is whether you got the samples from TI or a reputable distributor...

EDIT:  Sorry I missed your question about R225, 226 etc. above.  How far are these from the MCU on your board?   The R225,226 are source terminators, and ideally they're at the source (the MCU).  If they're say an inch or so away it is probably ok.    You don't want them 6 inches away.

The other thing you could do short of trying another board is to capture some oscilloscope shots of the JTAG lines as you try to connect.  We might see something odd there.

Also, regarding CAN3:

yong zhang2 said:

The can3 pins goes high when the nPORST is low, and go high when release them, while the can1,2 pins are not, they are always high whether the nPORST pin is low or not.

Was one of the 'high' supposed to be 'low'?   All of the CAN pins have pullups internally though, so if there's no program running on the MCU I can't think of a reason why any would be changed to 'low'.  

The emulator is ok when connectted to the HDK board. The TMS570 MCU is the sample from TI, so I think it's ok.

I actually assemble the board by hand, the CAN3 pins' go low when release the nPORST button. It's very strange. The R225,226 etc are about 5-7 centimeters away from the MCU, I don't know whether it matters a lot.

The PCB is not that perfect because the pull up/down resistors are missing when I design it, so I added the pull up/down resistors in another board and connect it with fly lines.

Now, I planned to design the PCB again to test it. It seems it's the only way I can do now.

yong

Yong,

Ok, if you want to send schematics we can take a quick look before you iterate again.