GBT IF System Non-linearity

1. Scope

The purpose of this project is to locate and correct the devices which exhibit non-linear gain in the GBT IF system. We also will need to identify "zero offsets" in various places in the IF system that can act like non-linear gains. Correction of non-linearities in the system have also been shown to improve the baseline shapes in spectroscopic observations.

2. Goal

To have the IF-system be linear to 0.5% over 12 to 13 dB changes in the input power (from 0.5 Volts to 9 Volts as measured in the IF Rack V/F power monitors). This is derived from a desire to have any systematic errors from non-linearities be no more than 10% of the error in a measurement with 5% calibration errors.

3. Personnel

Project Manager: Toney Minter
Project Engineer: Steve White
Project Scientist: Toney Minter
Project Team: Roger Norrod

4. Work Packages

Identify non-linear devices in the IF system.
Correction and/or replacement of non-linear devices (mainly amplifiers).
Identification and correction of zero offsets in V/F detectors, square law detectors and backends.
Data reduction handling of non-zero offsets and non-linearities

5. Memos

Number	Version	Title	Authors	Date
I	1	GBT01A_004 Observation Checkout Report - March 31, 2004	Toney Minter	March 31, 2004
II	1	GBT IF System Non-Linearity: April 4 and 8, 2004 Tests Results	Toney Minter, Ron Maddalena and Brian Mason	April, 19, 2004
1	1	May 2004 Measurements Of IF Non-linearities In The IF Rack And The DCR	Toney Minter, Steve White, & Roger Norrod	Nov 8, 2004
2	1	August 2004 Lab Measurements of the Optical Drive Module Linearities	Toney Minter, Steve White, & Roger Norrod	Nov 8, 2004
3	1	Backend DC Offsets and Calibration Errors as Sources of Non-linearity	Toney Minter	May 2, 2006
4	1	Lab Measurements of the Optical Drive Module Linearities with the New NRAO Built Amplifiersa	Toney Minter & Steve White	May 19, 2006
5	1	Linearity of the GBT IF Rack to DCR Square Law Detectors, March 2006	Toney Minter & Steve White	June 16, 2006
6	0	Ghost Images Of Strong Signals Arising At +/-100 MHz And From Aliased Power In GBT Spectrometer Spectra	Toney Minter	June 4, 2007
7	0	A VXCO Upgrade of LO3 To Reduce ±100MHz Sideband Image Ghosts	Toney Minter and Mike Stennes	August 16, 2007
8	0	Spectrometer Sampler Behaviour	Toney Minter	TBD
9	0	Continuum Level Comparison Between The DCR and Spectrometer	Toney Minter	TBD
10	0	Linearity Response Of The DCR and Spectrometer With Changing Input Power To The IF System	Toney Minter	TBD
11	0	Antenna Pointing Issues	Toney Minter	TBD
12	0	Observing Techniques To Minimize Baselines And Non-linear Effects	Toney Minter	TBD
13	0	Receiver Resonances	Toney Minter	TBD

6. Three-Month Goals

Cycle 6 2004 goals (August 16 to September 30)

IF Rack
- Identify non-linear amplifiers in all Optical Drivers
- Perform astronomical checks to confirm linear behaviour of Optical Drivers

DCR/IF Rack
- Have all Optical Drivers with linear gains and no zero offsets
- Investigate DCR Square-Law detector response (Van Vleck type correction needed?)

Data Reduction
- Investigate Re-calibration techniques for non-linear data.
- Investigate zero offset correction to data.
- Create document on how observers should measure zero offsets

IF Rack Zero Offset Monitoring
- create plan for monitoring offsets and when to correct in data or hardware
- begin monitoring zero offsets

Other devices
- Astronomical measurements to search for non-linearities in other IF components

Cycle 7 2004 goals (October 1 to November 12)

IF Rack
- Begin fabrication of "linear" amplifiers for Optical Drivers
- Begin replacement of non-linear amplifiers in all Optical Drivers (depends on IF Gain Balance project proceeding)
- Perform astronomical checks to confirm linear behaviour of Optical Drivers with new amplifiers

DCR/IF Rack
- N/A

Data Reduction
- Implement zero offset corrections to data.

IF Rack Zero Offset Monitoring
- continue to monitoring zero offsets

Other devices
- Astronomical measurements to search for non-linearities in other IF components

Cycle 8 2004 goals (November 15 to December 30)

IF Rack
- Complete fabrication of "linear" amplifiers for Optical Drivers
- Complete replacement of non-linear amplifiers in all Optical Drivers (depends on IF Gain Balance project proceeding)
- Perform astronomical checks to confirm linear behaviour of Optical Drivers with new amplifiers

DCR/IF Rack
- N/A

Data Reduction
- N/A

IF Rack Zero Offset Monitoring
- continue to monitoring zero offsets

Other devices
- Astronomical measurements to search for non-linearities in other IF components

Goals from previous cycles

Cycle 5 2004 Goals

7. Long Term Goals

Convertor Rack
- Search for and correct non-linearities
Analog Filter Rack
- Search for and correct non-linearities
Spectrometer
- Search for and correct non-linearities
- Identify zero offsets (in conjunction with Spectrometer Upgrades Project).
Spectral Processor
- Search for and correct non-linearities
- Identify zero offsets
Software
- Impliment zero offset correction methods

8. Resource Requests for Cycle 5 (1st July through August 15th)

Electronics Division Request
- Steve White (50% in conjunction with IF Gain Balancing Project)
- Roger Norrod (5%)
- Technician: (50% in conjunction with IF Gain Balancing Project)

Mechanical Division Request?

Scientific Services Division Request

- Toney Minter (25%)

Software Development Division Request?

Antenna Access Request?

Commissioning Time Request?

Additional Request Info?

Resource Requests for Cycle 4 ( May 16, 2005 - June 30, 2005)

These effort requests are reduced due to lack of funds to complete this project at this point in time.

Electronics Division Request
- Steve White (10% in conjunction with IF Gain Balancing Project)
- Roger Norrod (5%)
- Technician: (10% in conjunction with IF Gain Balancing Project)

Mechanical Division Request?

Scientific Services Division Request

- Toney Minter (25%)

Software Development Division Request?

Antenna Access Request?

Commissioning Time Request?

Additional Request Info?

Resource Requests from previous cycles

9. Status Reports

August 2004

Initially, the emphasis was on Optical Driver Modules which exhibited non-linear gain effects in the microwave regime and remove them from the system. Careful measurements in the receiver room identified the units in question. ODM sn007 and sn002 exhibited gain expansion while sn003 was the only unit showing compression. The decision was made to sacrifice sn007 for one complete working unit with the desired linearity. The unit, sn002, upon rebuild was thoroughly tested for linearity with the calibrated attenuators and spare GBT noise source. Thorough testing revealed a very linear response with a 1% compression point 12 dB below the 1 dB compression point as expected.

Having removed the units with nonlinear response from the system, emphasis shifted to the RF detection circuit, which exhibit errors from linear greater than 1% over a greater than 10 dB range. Inconsistent DC offsets over time and temperature accounted for the majority of errors indicating problems with the post detection amplifiers. Given the linearity in the microwave regime, the following modifications to correct the DC offsets lead to an error less than %+/-1 over a 20 dB range nearing the measurement accuracy of the laboratory environment:

1) Better grounding of the power supply ground to the chassis at the terminal strip 2) mounting the RF detector next to the video amp as apposed to mounting on the opposite side of the ODM with a coax cable interconnection 3) increased thermal mass of the RF detector housing reducing the temperature drifts of the DC offsets. 4) replacing a wire with a coax form the input connector to the PCB on the inside of the video amplifier. 5) adding a resistor between the non-inverting input and ground of the first DC amplifier. 6) drilling a hole in the lid so the potentiometer could be adjusted with the lid on.

The ODM sn002 was placed in the IF rack in channel number 3 on June 30. ODM sn008 was removed for testing. The modifications listed above were performed and the V/F circuits were adjusted for linearity. Improvements in testing methods allowed the V/F offsets to be removed ( I learned how to use a statistical feature digital scope properly). This resulted in an error less than +/- 1 % over a 20 DB range including the analog to digital conversion of the V/F circuits. ODM sn002 was in good agreement with the voltage reading from the square law detector, but may need check again in the future. ODM sn008 was inserted in channel # 6 and sn004 removed for testing.

Current Channel -> ODM SN map:

IF Channel	ODM SN	Comments
1	5*	Needs adjustment in RF Detection Circuits
2	6*	Needs adjustment in RF Detection Circuits
3	2	Adjusted for linearity in RF Detection Circuits
4	removed	Currently being tested in the lab
5	removed	Showed gain compression
6	8	Adjusted for linearity in RF Detection Circuits
7	1	Needs adjustment in RF Detection Circuits
8	removed	Showed gain compression

sn004 is currently being modified.

September, 21 2004

Current Channel -> ODM SN map:

IF Channel	ODM SN	Comments
1	1	Microwave response exhibits gain expansion of 1% beginning at 2 DB above operating point.
2	5	Linear with expected 1% compression at 13 DB above nominal operating point.
3	4	Linear with expected 1% compression at 13 DB above nominal operating point.
4	6	Linear with expected 1% compression at 13 DB above nominal operating point.
5	removed
6	8	Linear with expected 1% compression at 13 DB above nominal operating point.
7	2	Linear with 1% compression 6 DB above nominal operating point.
8	removed

Each ODM is modified to reduce DC offset drifts. The DC offsets of the video amp were zeroed as well as the voltage-to-frequency converters. ODM's sn001 and sn002 will need to have amplifiers replaced once they are built. Prototype amplifiers are currently being fabricated with testing expected to begin September 27.

Device Status Notes

Date	Device	Note
December 1, 2004	IF channel 5,6	A new and improved Optical Driver Module is installed in channel # 5 and is ready for use. This unit has two of the preproduction IF amplifiers, which have better linearity than the ODM in channel # 6. This link has been tested for stability.
December 8, 2004	IF channel 1	During testing on 8Dec04, intermittant instability in IF channel 1, on the order of 1% p-p was seen. See file: Dec8s16-23_ch1357.ps compare with: Dec8s1-15_ch246.ps These plots show total power from IF Rack detectors; the Chan number in the legends corresponds to the ODM number. Note the instability in the first few minutes of ODM 1 data. Until this is corrected, users probably should use ODM2 instead.
December 22, 2004	IF channel 7	All amplifiers in the system operate at least 20 dB below the 1 dB compression point for links without gain added to the ORM. In Link #7 with 15 dB added to the ORM, two amplifiers are operating close to the 1% compression point in the converter module. One is the second amplifier after the variable attenuator, where the bandwidth is ~500 MHZ, is operating in compression of the 1% point; the other is the final amplifier in the system for the 50-100 MHZ bandwidth mode, operating at the 1% point. This analysis for -5 dBm ODM output power. The sampler/filter module has enough margin.
December 22, 2004	IF channel 3,4 - S10	PTCS Q-Band tests (see PTCS Project Note 41) show that "ODM 3/4" behaves non-linearly. It is suspected that S10 in the IF Rack is the culprit.
anuary 4, 2005	IF channel 3,4 - S10	The S10 transfer switch in the IF Rack, suspected of producing non-linearities and gain stability problems particularly poignant in the Q band reciever temperature measurements,was replaced in the IF router.
anuary 21, 2005	IF Channel 1	ODM1 was replaced with a new module using the new amplifiers. This version should have better linearity and stability.
February 1, 2005	IF Channel 7 - CM 16	Non-linear behaviour was reported for ODM7 to CM16 to the ACS. This was investigated and it appears that the variable attenuator in CM16 is not working correctly. It seems that the fine steps are not being set correctly and that the gross one are working.
February 4, 2005	CM 16	CM16 has had its variable attenuator fixed and is not working correctly.
February 16, 2005	IF Channel 8	Another new ODM will be installed in link #8 during maintenance on February 17. Once checked out, ODM8 should be given preference over ODM7 since ODM8 has about 2 dB more range before the 1 dB saturation point is reached in the Converter Rack.
February 16, 2005	IF Channel 3,4	For narrow bandwidths separated by large amounts, ODM4 has about 4 dB more gain than ODM3. Preference should be given to ODM4 when observing under these conditions.
February 17, 2005	IF Channel 8	An ODM was installed in channel #8 and tested through to the DCR for signal. Another quick check revealed that the ALC of the optic link needs to be adjusted and stability tests ran, so I had Galen remove this channel from the cabling file. I'll adjust this link next maintenance day. I swapped the ODM of channel #3 so I could remove an older model amplifier; this link is ready for use. Steve
February 19, 2005	IF Channel 3,4	There was a report from Karen O'Neil that for project 3B18 there was insufficient power on ODM 3. ODM 4 was used instead. More information has been requested.
February 24, 2005	IF Channel 3,4	Report from Steve White:Investigating Fiber Optic Link #3 revealed 7 DB excess loss compared to Link #4. We have yet to determine the cause, but will investigate further tomorrow. A malfunctioning component of the ALC of Link #8 was replaced; however, the stability is still not adequate enough for use. This link will be removed from the cabling file until the cause of the instability can be determined and corrected.
February 25, 2005	IF Channel 2,3,8	Report from Steve White: Link #8 instability problem has been corrected and is ready for use. During stability tests Link #2 exhibited instabilities and should be avoided until further notice. The cause of the excess loss of Link #3 was not found; however the link is stable and can be used if the low output power is not a problem.
March 2, 2005	IF Channel 3	Report from Steve White: All fiber optic links were tested today and are now ready for use. The cause of the excess loss of Link #3 is not yet determined; however the link is stable and can be used if the low output power is not a problem.
March 3, 2005	IF Channel 3	Report from Steve White: The Optical parameters of Link #3 were adjusted today increasing the gain by 6 DB; however, the link gain is still 4 DB below link # 4 and could not be adjusted further. If this excess loss is a problem, we may have to do something radical like replacing the modulator. There appears to be 2 DB of excess optical loss somewhere between the laser diode and the photo-detector. All links were tested and remain stable.
March 8, 2005	IF Channel 3	The Optical performance of Link #3 continues to degrade to the point where the automatic level control is effected. Avoid using this link until further notice.
March 9, 2005	IF Channel 3	The laser diode modulator pair of Link #3 was replaced and the link tested. This channel is stable and the gain is back to normal. Also, the ODM of channel #4 was replaced improving the stability slightly.
March 23, 2005	IF Channel 2	During routine tests, Optical Link #2 presented instabilities. This link should not be used until further notice.
April 1, 2005	IF Channel 2	Optical Link #2 is now back in service.
May 5, 2005	IF Channel 6	During routine testing today, it was found that the total power measured at the output of Optical Receiver 6 shows intermittant instabilities of a few percent. Until Steve has a chance to further investigate the problem, I've asked Galen to add OR6 to the config tool bad device file. Please use IF channel 5 instead. Optical Drive 6 is still OK for continuum detection to DCR channel A6.
May 11, 2005	IF Channel 6	This link passed stability tests today and is ready for use.
May 13, 2005	IF Channel 5?	Jeff Wagg noticed intermittent ripples in his spectrometer data with a wavelength of ~53 MHz throughout the K-band receiver. (See the observing report on the web page for details.) After some tests it appears that the problem may be related to optical driver module 5. This is not conclusive. Nevertheless, if you see such features in any data please inform Roger Norrod. Also, you might try to add ODM 5 to a local dev_health.conf file to avoid this path. Dana We'll try to find a problem in IF 5, but the symptom is so intermittant and subtle it may take a while. And, please pass along any experience observers have with a 53MHz baseline ripple (the frequency seems to be very specific) as it might help pin down the cause. Thanks,Roger
May 17, 2005	IF Channel 7	IF Channel 7 no longer provides enough power to the convertor rack to balance the spectrometer
May 19, 2005	IF Channel 7	A defective laser driver module was replaced today. This link is cleared for use.

9. Other Information

This project is closely related to the IF Gain Balance project. Any amplifiers that are found to be non-linear in the IF Rack will need to be replaced. They should be replaced by amplifiers that are developed as part of the IF Gain Balance project. Currently there are no spare amplifiers for the IF Rack Optical Drivers.

-- ToneyMinter - 24 May 2004

Search the wiki with Google!

Topic IFNonLinearity . { Edit | Attach | Ref-By | Printable | Diffs | r1.52 | > | r1.51 | > | r1.50 | More }

Revision r1.52 - 22 Aug 2007 - 15:57 GMT - ToneyMinter Parents: WebHome	Content copyright © 1999-2007 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.


NRAO Home > Green Bank \| Wiki Topic: GB > Projects > IFNonLinearity

Changes \| Index \| Contents \| Search \| Statistics \| Jump to Topic: