EDIT: (problem definition shortened into 2 sentences)

For anyone using EVAL_AD7984_PMDZ and having troubles to apply -2.5, 0, 7.5 [V] and signal to differential probes, because the voltage source and signal source shares ground (then there is 2.5 [V] difference between the two), check the last comment for solution.


I got differential ADC EVAL_AD7984_PMDZ (page 17 connection) on ZedBoard working just fine (from -5 to 5 [V]). It requires not only 3.3 [V] from PMOD connection (logic) but also -2.5, 0, 7.5 [V] from external power source (Laboratory DC Power Supply - 2 adjustable outputs, that I have connected in series, thus I have "ground" on the middle pin, while on the other 2 pins there is -2.5 and 7.5 - relative to my "ground").

And here it comes, when I introduce input from another Laboratory DC Power Supply, the readings are fine within whole range.

But when I introduce signal from AFG-2125 Arbitrary Function Generator, there is offset of 2.5 [V] from both of the ADC input pins and it kind of messes up whole measurement that is below 2.5 [V].

May I get a little bit of insight on that ?

The ground of the ADC probes is the "ground" mentioned above.

Technically I have 3 grounds:

  • ZedBoard (and thus PMOD connection)(powered from its own power source)
  • External power source for the ADC (with the "ground")
  • Signal Generator ground

I am not sure how to resolve that. Would it help if the Signal Generator was battery powered ?


simulate this circuit – Schematic created using CircuitLab

(I am sorry if some electrical engineer finds the above schematic offensive)

Schematics for EVAL_AD7984_PMDZ

enter image description here enter image description here enter image description here

enter image description here

  • \$\begingroup\$ It's not really clear what you're doing. You can add circuit diagrams to your post (click the 7th icon from the left, which looks like a pencil and a circuit). That would make it clearer what was connected to what. Also, as a courtesy to the people who are trying to help, you could provide datasheets for the ADC, eval board and other bits of kit. \$\endgroup\$
    – Jack B
    Commented Mar 25, 2016 at 22:02
  • \$\begingroup\$ Is your lab supply grounded to mains earth ground? Sounds like both the signal generator and lab supply might have the same ground (the signal generator is single ended, so connecting it to a diff ADC is not so great). \$\endgroup\$
    – uint128_t
    Commented Mar 26, 2016 at 1:12
  • \$\begingroup\$ Yeah, from my understanding and measurements, the Power Source and Signal Generator share the same ground, but that seems like a crappy Signal Generator then. Also, why is the differential ADC not so great? I would say it is advantage to have two probes independent of ground and measure the difference between them (not probe and ground). \$\endgroup\$
    – Martin G
    Commented Mar 26, 2016 at 8:38
  • \$\begingroup\$ Check the jumper configuration. The board can be configured for differential or single ended mode, jumper SL10 should be in position A. Do you have a 'scope? \$\endgroup\$
    – Jack B
    Commented Mar 26, 2016 at 9:58
  • \$\begingroup\$ I am not sure whether we are talking about the same product, the board I have doesn't contain any jumpers, just simple PMOD (12 pins) and 2 probe connections (small round golden connectors) ... yes, I do have a scope \$\endgroup\$
    – Martin G
    Commented Mar 26, 2016 at 10:55

2 Answers 2


The eval board consists of a unity gain buffer on each input, the output of which goes into the ADC. Barring some details about current and input impedance, it is as though the board inputs are connected to the ADC inputs. For the ADC to work, the input voltages need to be between the ADC's ground pin and the ADC's vref pin (see datasheet table 2 -> ANALOG INPUT -> Voltage Range).

If you are powering it from a DC power supply with the eval board's -2.5V input at mains ground, then the eval board ground and hence the ADC's ground pin are at mains ground +2.5V. The ADC's vref pin is held at eval board ground + 5V by the voltage reference IC.

So when either input drops below mains + 2.5V, the corresponding ADC input drops below ADC ground. That takes the ADC out of spec, so you can't expect it to work properly. Worse, it goes below ADC ground pin -0.3V, which is outside the absolute maximum rating for the chip (datasheet page 6) which could damage the chip. Fortunately, the board includes some 22 ohm resistors which will limit the current and probably protect the ADC from damage.

You seem to be expecting the fact that the amplifier is differential to mean it can take any differential input. That's not the case. Differential means the output of the ADC is the difference between the inputs, but for the ADC to work the inputs need to be within the input limits. Every ADC and differential amplifier has input limits - you wouldn't expect to connect 1000000V to one input and 1000001V to the other and read one volt without fireworks. For this ADC, the limits are ground to Vref (which is 5V on the eval board). Very few differential amps or ADCs accept input voltages outside their power supply voltage range. If you really need that, then you need to look at either an isolating amplifier or a transformer.


Just to finish this thread with short answer. Powering the EVAL_AD7984_PMDZ requires the PMOD connection 3.3[V] and GND (pins 5, 6 and 11, 12 on any 12 pin PMOD header) and -2.5, 0, +7,5 [V] on J3 connector. To prevent the above described problems (shared ground between source and signal generator) provide the voltages (-2.5, 0, +7,5 [V]) from a battery source, OR use Voltage Inverter to obtain -2.5 [V] (thus, the shared grounds are at the same level). The LT1054 Voltage Inverter is easy to use and does the job very well.

http://cds.linear.com/docs/en/datasheet/1054lfh.pdf page 10


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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