Segger Debug Out on DK can't see my target

RE: Segger Debug Out on DK can't see my target

Jay Tyzzer — Tue, 27 Oct 2015 23:40:32 GMT

If you use nRFGo studio you will see the part number and Build Code for the device you are programming. The part on the nRF51-DK is a nRF51422. If you are using a nRF51822 you should see that listed instead of the one on the DK.

RE: Segger Debug Out on DK can't see my target

Martijn — Tue, 27 Oct 2015 20:59:42 GMT

Pro tip: cut the measurement pad on the dk(SB9). To program the dk itself place a jumper on DK measurement pins. To program an external board remove the jumper to be sure the dk itself is not programmed..

RE: Segger Debug Out on DK can't see my target

Bret Foreman — Tue, 27 Oct 2015 20:46:56 GMT

It always turns out to be the simplest thing. I'm connecting to my target using pogo pins to landing pads. The whole works is pretty small and my fixture seemed precise. It turns out that the board was shifting in the fixture and the pogos were not lining up reliably.

I've now got the board programmed with an S110 SD and with my test app. And it's connecting to my phone.

One pro-tip: It's very easy to accidentally program the DK rather than the target if you forget to power up the target. You really can't tell which is being programmed. So I loaded the "blinky" app onto the DK. If the LEDs are blinking, I know I programmed my target, not the DK.

RE: Segger Debug Out on DK can't see my target

MartinBL — Tue, 27 Oct 2015 08:46:40 GMT

Also, packet CFAC means that you have a 3rd revision IC so a pull down should not be necessary although some customers use one anyway to avoid ESD messing with the line.

Did you check the wiring?

RE: Segger Debug Out on DK can't see my target

MartinBL — Tue, 27 Oct 2015 08:45:40 GMT

Hi.

So I understand it like we have isolated the problem to your custom board?

The 16MHz clock should not be necessary. But for you information: The CL mentioned in the white paper is given in the crystal's datasheet. Then you can calculate the values of your capacitors as described on page 9 and 10 in the white paper. EDIT: Remember the parasitic capacitance inside the NRF51 of about ~4pF (~1pF on nRF24LE1). These will in effect be connected in parallel to C1 and C2.

These are the data sheets for the nRF51: Reference Manual & Product Specification.

RE: Segger Debug Out on DK can't see my target

Jay Tyzzer — Tue, 27 Oct 2015 08:27:10 GMT

To address some of this items listed above.

A target board running at 3.0vdc will program fine. (The Beacon Reference design for example runs on a button cell)
The Unit will flash without the High speed crystal operating.
The CL value can calculated : (CL*2)-4 (for the nRF51) The 4 is the stray capacitance in pf on the pins from the ESD protection and wire bonding. Pick the nearest standard value cap. CL=8pf then load cap SB 12pf (each side!)
The high speed crystal clock comes on when needed.
Rev 2 and Rev 3 silicon have pull ups and pull downs on the SWD lines. No external ones are needed but will not hamper it from working either.
Pin 1 on P19 is not a voltage source, it is a reference input to the Segger part telling it that there is something connected and correct Vcc..

RE: Segger Debug Out on DK can't see my target

Bret Foreman — Mon, 26 Oct 2015 23:16:37 GMT

I've been looking for CL in the product specification document but I can't find it. What CL is the crystal oscillator white paper talking about?

RE: Segger Debug Out on DK can't see my target

Bret Foreman — Mon, 26 Oct 2015 23:15:23 GMT

Hmm. Figuring out the correct value for the load caps turns out to be harder than it should be. I looked at the white paper entitled "Crystal Oscillator Design Considerations" and the calculations it suggest are simple enough. However, it contains the following sentence:

When buying crystals you can usually choose from a range of capacitive loads (ex: 8-20pF). When ordering a crystal you must choose a CL (ex: 12 pF) that is within the range of the CL specified for the nRF device. The CL range that nRF devices can operate in is only limited by the maximum CL (ex: CL < 16 pF). This maximum CL depends on the on-chip crystal oscillator design and is specified in the device data sheet.

I can't find a data sheet listed here.

RE: Segger Debug Out on DK can't see my target

Bret Foreman — Mon, 26 Oct 2015 22:27:20 GMT

It looks as though the 16MHz crystal is not starting. I'm not sure if that will prevent the JTAG from working but this thread seems to indicate so. I used 12pF caps on the crystal, copied from the reference design. But I see the values are rather critical as specified in section 3.6.1 of the Product Specification. I'll do some measurements and calculations and see if changing the caps will get the crystal to start. Does anyone know if that would prevent the JTAG from working?

RE: Segger Debug Out on DK can't see my target

Bret Foreman — Mon, 26 Oct 2015 22:13:43 GMT

I added a 100K pull-down to the SWDCLK pin as per this thread. But that did not fix the problem. I still think it's a good idea to have this pull-down in place, though, since my target does not have it and the DK I'm using to program doesn't either. And I'm not sure if I have the latest version of the silicon where there is an internal pull-down.

RE: Segger Debug Out on DK can't see my target

Bret Foreman — Mon, 26 Oct 2015 17:46:33 GMT

I broke out a second nRF51 DK and connected a cable from P19/Debug Out to P18/Debug In. I did not connect the second DK to USB. Rather, I powered the second DK from a benchtop supply. Just to make sure of which nRF51 I was programming, I used the "blinky" example. I was able to see the second DK's nRF51 from the Segger on the first DK and verified that I was programming the second DK when the blinky lights on the second DK flashed as expected. So my Debug Out is working with an external target.

I then connected my new board in the same way and gave it another try, powering it from 3.3V just be be sure. No action - the Segger can't see my target.

That said, my target is extremely small and I connect to it with a custom-made programming fixture with pogo pins. There's a chance that fixture is mis-wired or the pogos are not making contact. I'm going to verify all the wiring and try to verify the pogo contacts. If all that checks out, then I'll upload the gerbers for examination by the group.

RE: Segger Debug Out on DK can't see my target

MartinBL — Sat, 24 Oct 2015 12:58:17 GMT

Hm... That's unfortunate. In the first link Asbjørn is actually referring to the same user guide as I did and it clearly states 3.0V. In the second link Asbjørn says "or close to", but I agree that "close to" is pretty ambiguous. Anyway I did a test with two nRF51 DKs, one acted as the debugger and the other one as the external board powered by the first one (as OP describes). The voltage on both boards were 2.88V and it worked just fine.

RE: Segger Debug Out on DK can't see my target

RK — Sat, 24 Oct 2015 11:24:49 GMT

I got 3.3V from, devzone.nordicsemi.com/.../ and devzone.nordicsemi.com/.../

RE: Segger Debug Out on DK can't see my target

MartinBL — Sat, 24 Oct 2015 11:19:05 GMT

Where does it say that it needs 3.3V? In the nRF51 DK user guide pp. 20 it says that "The voltage supported by external debugging/programmins is 3.0V". There is actually nothing on the nRF51 DK that runs on 3.3V (by default anyway). The USB power gets regulated down to 3.3V, but the current is running through a reverse voltage protection diode with a ~0.3V voltage drop so the actual VIO (supplying the Segger chip) and VDD_nRF are usually at ~3.0V or even lower.

So I suspect the problem lies elsewhere. It should be plug and play. Do you have other debuggers to test if the problem is with your custom board or your nRF51 DK? Are you able to program other boards/kits with your nRF51 DK? Do you know for sure that your custom board is working? No shorts? All traces are connected? Do you mind posting your layout?

RE: Segger Debug Out on DK can't see my target

RK — Sat, 24 Oct 2015 02:53:49 GMT

The debugger on the DK needs 3.3V to work (it's in the manual), you say you're using 3.0V, so I'd start by increasing that to at least 3.3V to see if it starts working.