X5-RX calibration

mrodby · Joined: 09 Feb 2010 Posts: 36

In the Snap sample application for the X5-RX, I changed Direct Mode to true, in order to allow us to collect data continuously and avoid any extra memory copies. When data is received by ApplicationIo::HandleDirectDataAvailable, are values already calibrated, or do we need to apply a scale factor and DC offset? Also, are values from multiple channels obtained simultaneously?

During some preliminary tests, we fed the same signal into both of the first two channels, and ran them at 130 MSPS. The resulting magnitudes were different by about 4.5%, there was a DC offset of about 190 counts between channels, and there was about a half sample time phase offset between the channels. We did not do very extensive testing, such as swapping the cables from the splitter to the X5-RX to see if there were any resistive or capacitive effects that could explain any of these differences. I would just like to know what I should expect before spending a lot of time on further tests.

Thank you.

jhenderson · Posted: Thu Mar 18, 2010 5:34 am Post subject:

The stock X5-RX FPGA firmware applies a linear y=mx+b correction to all incoming A/D samples. The values of m and b are derived during factory testing, and are stored into an onboard Flash ROM. These values can be seen or changed via the EEProm tab on the Snap example.

The standard testbed uses a 5% as the threshold for acceptable gain error. However, as the cards exit the test stand, they will generally exhibit inter-channel gain errors < 1% and offset errors of less than 10 counts.

Check the EEProm tab of the Snap example: Is calibration enabled? The SW uses unity gain and 0 offset if the calibration bit is cleared, defeating factory cal settings.

mrodby · Joined: 09 Feb 2010 Posts: 36

I had not been able to look at the Snap application again until today. When I press the Load button on the EEProm tab, all of the Gain and Offset values are 0. I assume this means that the unit is uncalibrated. Do you keep a copy of the calibration values that were obtained at manufacture time for each board? If so, how can I get a copy of them?

jhenderson · Posted: Thu Apr 22, 2010 4:12 am Post subject:

The most likely explaination for this is that the unit was shipped before the production testbed was complete. Since we were late in the shipment of the X5-RX, the unit may have passed manual testing, and not have been subjected to the automated tests.

Supply the serial number of your board and we'll look in the database. If present, we'll supply the factory cal data. If not, you can return the board for factory test and cal.

mrodby · Joined: 09 Feb 2010 Posts: 36

The serial number is A101.

jhatcher · Posted: Fri Apr 23, 2010 8:03 am Post subject:

I have reviewed the database information on X5-RX, A101 and it appears that the board did not get machine calibrated. As Jim mentioned, early boards sometimes get shipped prior to having a machine calibration.

If you wish to return the board for calibration, call Peg Todd at 805.578.4260 and she will issue to you an RMA#.

Thank you.