Yes I am aware of that but thank you for the reminder. I have the keysight N9010A it’s 10Hz-3.6GHz I use for RF design and pre-compliance tests but for this application I use the old HP8590D that goes up to 1.8GHz which is already more than enough as I am looking at max 4-500MHz. Both analyzer tolerate up to 25-30VDC but I am measuring a gainstage output for oscillations so it should only have few mV.
BTW you had not further commented on the analyzer issue, but the problem is related to the CH2 input. This points out to most likely the 1uF AC coupling cap most likely needs replacement or a bypass as it probably got lossy. If you or anyone else has any other tips please let me know I will probably open the analyzer soon to see how easy it is to get access to the coupling cap to run a test
BTW you had not further commented on the analyzer issue, but the problem is related to the CH2 input. This points out to most likely the 1uF AC coupling cap most likely needs replacement or a bypass as it probably got lossy. If you or anyone else has any other tips please let me know I will probably open the analyzer soon to see how easy it is to get access to the coupling cap to run a test
Thanks Duke and Tom. The Range is fixed in this test to 2.5V. If I set it to 5V Range, I get both channels to fail similarly (max limit is -80dBa)
If I choose 2.5V range as the original test now it fails only the right channel and the -75dB figure remains unaltered. Now if I set autorange intead, as Duke suggested, at 2.5V noise could now be more of a problem, below it clearly shows one channel is significantly worse than the other by a large margin especially above 1KHz.
Originally, I thought it could have been a leaky coupling cap, but I have just realized the test runs with DC coupled, which is basically bypassing the input caps, so I would assume the cap is fine.
Duke, do you have any suggestions as to what it could potentially be? it seems one channel is noisier than the other and if I swap generators the problem remains on the same side, which would point to the analyzer's input, that is why I was erroneously thinking the coupling cap.
To answer Tom, how much do I care? it's not affecting the measurements I need to make at the moment, but I would still prefer to have the instrument if full compliance if possible.
If I choose 2.5V range as the original test now it fails only the right channel and the -75dB figure remains unaltered. Now if I set autorange intead, as Duke suggested, at 2.5V noise could now be more of a problem, below it clearly shows one channel is significantly worse than the other by a large margin especially above 1KHz.
Originally, I thought it could have been a leaky coupling cap, but I have just realized the test runs with DC coupled, which is basically bypassing the input caps, so I would assume the cap is fine.
Duke, do you have any suggestions as to what it could potentially be? it seems one channel is noisier than the other and if I swap generators the problem remains on the same side, which would point to the analyzer's input, that is why I was erroneously thinking the coupling cap.
To answer Tom, how much do I care? it's not affecting the measurements I need to make at the moment, but I would still prefer to have the instrument if full compliance if possible.
increasing detection time and moving input voltage to 0dBu the difference is more marked between the two channels.
I understand and support that. I didn't realize the discrepancy was this large. Duke is a better person to ask for repair advice. I leave the repairs to AP (and cry myself to sleep from the expense). Thankfully, I have not needed a repair.
Tom
Tom/Duke....what is this test trying to measure the CMMR of the Analyzer's input? What could cause this discrepancy?
By selecting DC on the Analyzer's input, am I bypassing the input AC coupling cap or is it always engaged?
By selecting DC on the Analyzer's input, am I bypassing the input AC coupling cap or is it always engaged?
- The DC is bypassing the Analyzer cap. This analyzer (SYS2 2522) does not measure DCV.
- The Analyzer is OPEN 100K ohms, select 600 inputs (Check the box), and see the BANDWIDTH NOISE plots.
- Try running a series of tests by selecting different fixed ranges.
- The Analyzer is measuring FIXED BANDWIDTH NOISE, not CMMR
- I would always know both channels will be slightly different.
First Plot in Gen Mode, Yellow Trace is Channel 2 while Cyan is Channel 1. It can be see as CH2 gen is noisier. Note selecting DC or not, makes no difference.
BUT if instead of selecting GenMod I go with external XLR Gen CH2 into ANL CH2, it gets a little bit better
This is noise with Analyzer OPEN with 100Kohm input
600ohm input
300ohm input
Unfortunately, my knowledge compared to your is extremely low, so I am not understanding what other tests you would like me to run. When I run the CMMR test at different ranges, Channel two remains still the same at -70db @20KHz while on CH1 I the CMMR varies. Plot at post #245 is with auto range for example, while at @244 I show same measurement at different ranges.
Is this what you were asking me to run or something else? Again apologies I certainly don't know how this works.
BUT if instead of selecting GenMod I go with external XLR Gen CH2 into ANL CH2, it gets a little bit better
This is noise with Analyzer OPEN with 100Kohm input
600ohm input
300ohm input
Unfortunately, my knowledge compared to your is extremely low, so I am not understanding what other tests you would like me to run. When I run the CMMR test at different ranges, Channel two remains still the same at -70db @20KHz while on CH1 I the CMMR varies. Plot at post #245 is with auto range for example, while at @244 I show same measurement at different ranges.
Is this what you were asking me to run or something else? Again apologies I certainly don't know how this works.
Last edited:
What you're seeing in the yellow trace is mains hum. Try swapping the cables. I.e. the cable that connects OUT1 to IN1 should now connect OUT2 to IN2 and vice versa. Does the problem move with the cable? If so, you have a bad cable.
I don't know the innards of the AP well enough to say for sure, but I suppose this could also be caused by a dead supply reservoir cap or bypass cap. I would be a bit surprised if the two channels have separate supplies, but like I said, I don't know the internals well.
I wonder if a blown output impedance or input load could do this as well.
The CM test tests the (should be) balanced output impedance of the source and the CMRR of the analyzer input.
This may seem a bit random: Do you have the AP sitting below or on top of another piece of equipment? If so, try turning that equipment off and/or moving the AP so it sits alone on the desk/workbench.
I have a pair of HP 34401A multimeters sitting atop my APx555. If I turn on the one that's above the analyzer source the hum increases by 20-30 dB. It's still at the gnat fart level, but in a 128k point FFT with the 44.1 kHz ADC it shows -130 dBV at 60 Hz with a noise floor around -160 dBV. That drops to a few dB above the noise floor when I turn that DMM off. I saw the same issue on the APx525. I brought it up with AP and they just shrugged and said it met spec (which it does). But still... I think equipment should play nicely with others. I now have some mu metal sheets between those DMMs an the AP.
Tom
I don't know the innards of the AP well enough to say for sure, but I suppose this could also be caused by a dead supply reservoir cap or bypass cap. I would be a bit surprised if the two channels have separate supplies, but like I said, I don't know the internals well.
I wonder if a blown output impedance or input load could do this as well.
The CM test tests the (should be) balanced output impedance of the source and the CMRR of the analyzer input.
This may seem a bit random: Do you have the AP sitting below or on top of another piece of equipment? If so, try turning that equipment off and/or moving the AP so it sits alone on the desk/workbench.
I have a pair of HP 34401A multimeters sitting atop my APx555. If I turn on the one that's above the analyzer source the hum increases by 20-30 dB. It's still at the gnat fart level, but in a 128k point FFT with the 44.1 kHz ADC it shows -130 dBV at 60 Hz with a noise floor around -160 dBV. That drops to a few dB above the noise floor when I turn that DMM off. I saw the same issue on the APx525. I brought it up with AP and they just shrugged and said it met spec (which it does). But still... I think equipment should play nicely with others. I now have some mu metal sheets between those DMMs an the AP.
Tom
Tom, I do have the oscilloscope sitting on top of it I did turn it off for the test as I had the same idea you had, so I am not sure it has to do with magnetic field picked up from a nearby instrument. However, out of curiosity I will just put the instrument on another table and repeat the test.
The CMRR test which is band Noise as I was kindly explained, can fail because the symmetry of the amplifier is not perfect and one channel being a bit noisier than the other cause the trigger in a different spot thus causing the misleading result. I know I am reporting the explanation most likely incorrectly, but overall I was reinsured the analyzer performs correctly unless I do government work 😂😊
The CMRR test which is band Noise as I was kindly explained, can fail because the symmetry of the amplifier is not perfect and one channel being a bit noisier than the other cause the trigger in a different spot thus causing the misleading result. I know I am reporting the explanation most likely incorrectly, but overall I was reinsured the analyzer performs correctly unless I do government work 😂😊
Similar issue here. On the bench that I use to take specs, I have SMPS PSUs. I found that the all-linear PSUs did a very similar thing with my analysers, whether AP or R+S. You see a small spike in the 200K-ish region from the SMPS that the linears don't exhibit, but you get a much better THD+N figure in the all-important 20-20KHz area.
As I have mentioned previously, I use Sam Groner's 60dB ultralow noise amplifier from Linear Audio -- some of the diodes are MELF 1N4148 in clear package for the voltage reference -- if I don't put the amplifier in its biscuit tin you can see the effect of fluorescent lights modulating the diodes.
MuShield sample packs are available from the company https://www.mushield.com/material-sales/magnetic-shields-design-kits-msk-kits/
MuShield sample packs are available from the company https://www.mushield.com/material-sales/magnetic-shields-design-kits-msk-kits/