RE: ADS1120 Configuration registers

Hi David,

Welcome to the forum!  I split your post from the previous thread as that thread was a slightly different topic and 3 years old.  Regarding register configuration, you only need to write to the register when you want to change a setting.  Furthermore, you do not need to write all the registers, just the register you wish to change.

The ADS1120 does not have self offset calibration, so if you are seeing offset issues I would suggest setting the mux of the ADS1120 to the setting of "AINP and AINN shorted to (AVDD + AVSS) / 2" and use that conversion result to correct for offset by subtracting the result from the conversion result of the analog input.

If you are not changing the configuration and you are using single conversion mode, then all you need to do is send the START command to start a new conversion.

Best regards,

Bob B

Thank you.

Ok. I will change the form of pick up the measure.

It is strange, at the start of measure all samples are correct, 0 or ±1, the offset is ok. When the samples are more or less 500 or 1000, some ADC beggin to failure but in general the most of ADC are working correct, without failure.
When the ADCs is measuring bad, if we do a reset or power down of proto. The wrong ADCs return to work ok.    

Ok. We change the measure block, we will do one first measure of offset and the second measure of the input.

Ok. Thank you.

David.

Hi.

I have other question about ADS1120. It is about the marked.

I have seen that th components have differents codes, for example.

The componente with more failure rate have the follow codes.
In the top of component: ADS1120 / 71KG4 / A82V
in the botton: Malasya / M27 / Axx  or  M28 / Cxx

This type of component have many failure during the measure and only in one of the inputs.


The other component  without failure rate have this follow codes.
In the top of component: ADS1120 / 44KG4 / ACKV
in the botton: M82 / E1E (it doesn't apear the malasya word)

This componente doesn't have failure of offset.

Can you say me which is the difference between this codes? and if is it possible that the first component had had problems with production?

Thank you.

Hi David,

The numbers on the top of the package give information as to when the devices were packaged along with a code regarding assembly/test site and lot trace information.  The first 2 numbers give the year and month of manufacture.  The 71 indicate 2017, and first month (January).  The 44 would be April, 2014. 

Regarding the device code in question, we have not had any reports of issues.  It is possible that your design in marginal.  If you would send me your schematic I can attempt to evaluate why you might be seeing some issues or possible areas of concern.  Also, if you could send me the configuration setting you are using that would also be helpful.

Best regards,

Bob B

Hi Bob.

Thank you.

I send you a schematic of one analogical cell of equipment. This equipment have 8 cells with two inputs for cell. And all cells are controlled with a CPLD. If you need all schematic I send you, no problem.

The circuit work correctly if I connect de resistors R35 and R43, the detection is ok with this configuration. These resistor fix the negative input to GND of the cell, erase or match the potential.

With this configuration the circuit work correct. But I need that the two inputs of ADC is working with true differential configuration.

When I have done a test without the resistors many inputs work correctly, the offset is 0 or ±1, but some inputs begin to work bad, and this inputs measure one offset that it can be between 12 and 180 points.

The problem, I suppose, it is generated for noise or differences between potential of input and the analog GND of ADC. The inputs only have the capacitors (162, 163, 165, 167) for interconnect or match the input with the analog GND of ADC.

I think that it is the problem but I don't understand where I have a problem and how I can erase the problem with the potential.

The circuit have one change. AVss is 2V7 and AVdd is -2.7. At the initial this voltage was 3V3 and -3V3, but during the test I changed the voltage. I think that it was not necessary but i have fixed it to 2V7.

The GND of the cell is the same for the digital and analog part but i have rutted the digital and the analog for different parts of the cell and only I put together near the GND of ADUM or negative pin of C90.

Thank you.

David.

2705.Analog.20m_V1.00.sch.pdf

Hi David,

Let's start out with the analog power supply.  Initially you were using +/-3.3V or a 6.6V span.  This is coming very close to the absolute maximum values shown in table 6.1 on page 5 of the ADS1120 datasheet.  Second, -3.3V exceeds the absolute maximum for AVSS to DGND.  It is quite possible the devices were damaged at this point.  Even with +/-2.7 V there could be a problem as this voltage is approaching the absolute maximum and depending on regulation could exceed it.  For the analog supply using a bipolar supply, I would suggest using +/-2.5 V with a well regulated supply.  Using the direct output of the charge pump may be a bit noisy.  From the charge pump I would suggest using a low noise LDO linear regulator to help lower the noise and better regulate the supply.  I also see that you are using a 2.5V reference, and you might consider using the ADR431 reference to also supply AVDD.

Unfortunately I do not see R35 or R43 in the schematic, so I cannot tell what those devices are doing.  Perhaps you mean R14 and R205?  If that is the case, then you must make sure that your inputs do not float out of the common-mode range. When the resistor is installed, then the common-mode is established around mid-analog supply.  How are your sensors powered?  How are the sensors maintaining proper common-mode? 

With bipolar supplies and when AINN is connected to AGND, you still have a differential measurement.  However, based on your design I think I understand why this may be a problem.  I think you understand that the input is possibly outside the common-mode input range based on what you have told me, but without a better understanding on how the sensor is powered it is difficult to say how you should connect the sensor to your board.  Is the sensor powered by the bipolar supply?  You say the grounds are all connected together, but where is the ADS1120 inputs relative to AGND (what is the voltage from AINP to AGND and AINN to AGND at each input pair?)  Also I do not see C90 on the schematic.

Best regards,

Bob B

Hi, Bob.

I do not see R35 or R43 in the schematic. Perhaps you mean R14 and R205?
Yes, i have had a mistake with the scheme. It's the same but the components have other names. You have the truth, resistors are R14 and R205.

I don't put the AVdd and AVss to 2V5 because the signal input have a FS to 2V5. And in the beginning  to the design we prove with this voltage and we had a problem. It is possible that the problem was generated for other factor.

The inputs do not float out of the common mode range.

The test has two parts, the first, the input is only a short-circuit between INp and INn. The second part, I put a measure with a current injector (portable), this use batteries.

The reference of injector and the ADCgnd is different.

With this type of injector I don't control the common-mode, the equipment work with batteries and i thought that the CMRR could erased this problem.
Normally this equipment work in extreme environment with different potential and magnetic fields, the sensors that it measure many times they are connected to different potentials. Normally, the work zone have the same earth but it is possible than no. 

Is the sensor powered by the bipolar supply?  

Normally this equipment work with sensors with unipolar supply. This supply is generated with batteries (connected to earth of installation).
The sensor can be, with 3 wires and use a common GND and unipolar output and the other type of sensor use 4 wires with output floating.

R14 and R205 should not be mounted in the pcb with a normal use.
The inputs doesn't have a relative connection with ADCgnd, except through to the capacitors C167, C165, C162 and C163. But this is not a really.

I think that my problem is the attached of noyse or potential. Is it possible? And how can I erase?
Do I have change the idea and conect the INn to ADCgnd?

What is the voltage from AINP to AGND and AINN to AGND at each input pair?

If i don't put the resistor R14 and R205, the external signal don't have reference with the ADCgnd and it is possible that the measure is displaced or it have attached a potential.

Thank you, Bob.

David.

Hi David,

Earth ground is reasonably good for safety, but is not very good for signal ground.  What we are really talking about with 'ground' is a common point.  Your input signals must be within a region as specified by the datasheet relative to the common point. 

A capacitive coupled connection to ground will have some measurement issues relative to point where the capacitor may be high impedance or low impedance.  You may end up measuring both a DC as well as an AC component.  If the AC component is truly common to both inputs, then this should be rejected.  However, if there is any filter mismatch due to tolerance of the resistors or capacitors, a difference voltage can be created.  This difference voltage will directly impact your measurement.

Best regards,

Bob B