5
\$\begingroup\$

I've got an ENC28J60 Ethernet Controller in a circuit I'm building. I was testing it out but when I got to the point of talking to it over SPI with my micro-controller I got utter silence back from the device.

I got a 'scope out and sure enough, the the SCK and MOSI pins on the device are wiggling the way I expect them to while the /CS pin is driven low, but the MISO signal is flat-lined at GND (all measured at the pins of the ENC28J60). So I got to looking at the solder joints on the device more closely (using an SMD package). I saw that I had a subtle bridge between two of the pins: VCAP and VSS!

So I went back to the soldering station and remedied the bridge and headed back to the 'scope. Still not working, but doing something different (progress!?). Now when the SCK signal fires away, the MISO signal does "react," just not as I'd expect. The signal has a "wiggling characteristic" that is correlated with SCK, but it's certainly not what a digital output should look like, and voltage levels go negative to about -100mV, and never exceed 0V.

On to the question. What is the impact of having grounded VCAP? Is the chip fried? What is the explanation for the negative quasi-inverted quasi-digital output on the MISO line?


Edit I measured (with a 'scope) every pin of the ENC28J60. Why would the crystal not be ringing? Anything else look out of the ordinary?

+-------+----------+-------------------------------+
| Pin # | Name     |Measurement Result             |
+-------+----------+-------------------------------+
|   1   | VCAP     | 2.66VDC (760mv P-P @ 4.15MHz) | 
|   2   | VSS      | GND                           |
|   3   | CLKOUT   | GND                           | 
|   4   | /INT     | 3.3V                          |
|   5   | NC       | 3.3V                          | 
|   6   | SO       | Noisy/Negative Digital Signal |
|   7   | SI       | 0V/3.3V Digital MOSI Signal   |
|   8   | SCK      | 0V/3.3V Digital SCK Signal    | 
|   9   | /CS      | 0V/3.3V Chip Select Signal    |
|  10   | /RESET   | 3.3V                          | 
|  11   | VSSRX    | GND                           |
|  12   | TPIN-    | 1.08V (floating)              | 
|  13   | TPIN+    | 1.08V (floating)              |
|  14   | RBIAS    | 1.28V                         |
|  15   | VDDTX    | 3.3V                          | 
|  16   | TPOUT-   | 3.3V                          |
|  17   | TPOUT+   | 3.3V                          | 
|  18   | VSSTX    | GND                           |
|  19   | VDDRX    | 3.3V                          | 
|  20   | VDDPLL   | 3.3V                          |
|  21   | VSSPLL   | GND                           |
|  22   | VSSOSC   | GND                           | 
|  23   | OSC1     | GND (hm...)                   |
|  24   | OSC2     | GND (hm...)                   | 
|  25   | VDDOSC   | 3.3v                          |
|  26   | LEDB     | GND                           | 
|  27   | LEDA     | GND                           |
|  28   | VDD      | 3.3V                          |
+-------+----------+-------------------------------+
\$\endgroup\$
  • 1
    \$\begingroup\$ Just replace the chip and move on. You are wasting way more time guessing whether the existing chip is useable than what a new chip costs. Since you violated a absolute maximum spec, the chip could be broken in subtle ways even if it does appear to work. \$\endgroup\$ – Olin Lathrop Apr 18 '12 at 12:23
7
\$\begingroup\$

VCAP is the 2.5V reference of the IC. It's internally generated from the 3.3V supply.

There's a telling footnote under the "Absolute Maximum Ratings" table:

VCAP is not designed to supply an external load. No external voltage should be applied to this pin.

The most important part of that is: "not designed to supply an external load". Your short circuit was a significant external load.

If VCAP hasn't recovered after the external short was removed, replace the IC. Even if VCAP is OK, if you can't get correct behaviour on MISO and you have any doubts, replace the IC.

\$\endgroup\$
3
\$\begingroup\$

Short answer : Probably Yes

Long answer: Page 83 of spec says...MAX! VCAP with respect to VSS -0.6V to 3.0V But you shorted Vcap to Vss..i.e. shorted internal 2.5V reference on Vcap to ground. Vss so Check pin 2 for 2.5V. and go from there.

If ok then pin 3 CLK out... if ok then other pins.

Pushing the envelope or stationary running? ha. Score as fit as you may be.

OSC1 is input and OSC2 is the inverted output. If both pins are low (gnd) then the output is either shorted to ground or blown. Normally high resistance feedback between pins internally will make input Vdd/2 and output must oscillate. Is crystal connected? if so which?

\$\endgroup\$
  • \$\begingroup\$ See edit for measurements of every pin. \$\endgroup\$ – vicatcu Apr 17 '12 at 23:33
  • \$\begingroup\$ Fix Vcap pin 1 with 10uF or so to get 2.5Vdc with < 50mV ripple using a differential method with common mode shield ground. & short ground probe to scopes probes. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Apr 17 '12 at 23:55
  • \$\begingroup\$ Vcap does have a 10uF capacitor to GND on it... This crystal (search.digikey.com/scripts/DkSearch/…) connected with 22pF capacitors to GND on either leg. \$\endgroup\$ – vicatcu Apr 18 '12 at 1:34
1
\$\begingroup\$

I am not sure. You can have no response from chips at all when the SPI is not clocked in correctly. This chip also has an oscillator, so if you can check that do that as well (if there is anything sine-wave like in the frequency of the crystal it's OK).

I've experience with microchip parts working fine when VCAP was shorted or tied to 3.3V in a few initial test attempts. This was, however, a PIC32 or PIC24.. so if the microcontroller doesn't want to load a program you're checking voltages and ICD connections straight away and fix the issues.

Chapter 16.0 says that Vcap should stay between -0.3V and 2.75V in respect to VSS. So, 0V in respect (that is VSS) is also included. It doesn't say anything about the short circuit capability though. If the Vcap still has got 2.5V, it's probably fine. As long there is a decent 10uF capacitor on there it should work.

In that case the chip may still work. The negative ground signal may be something to do with the measurement set-up. Are you probing GND directly at your chip, or very externally? This may cause voltage spikes or shifts.

If your Vcap is now way off (I can imagine you have been running this for some time figuring what's going on), then I'd replace the chip.

\$\endgroup\$
  • \$\begingroup\$ I was connecting probe GND to a distant GND pin the board, not at VSS on the chip... I'll report back on the status of VCAP and the crystal pins. \$\endgroup\$ – vicatcu Apr 17 '12 at 21:18
  • \$\begingroup\$ See edit for measurements of every pin. \$\endgroup\$ – vicatcu Apr 17 '12 at 23:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.