RIGOL DSA-815 | Help With EMI Pre-certification Test

[Hyphen-B]

Equipment: Rigol DSA815, EMI precertification test software, a bicon, a log, a preamp, and a comb generator.

Hello, everyone. I considered posting this topic in the Newbie forum, but my situation seems to be more specifically related to what's discussed here. I apologize if this is inappropriate, and will close the topic and move it if needed. Please bear with me if anything I say is either extremely obvious or just flat out wrong.

I work as an electrical assembler for a small company building light-industrial laser automation systems. Recently the owner of the company tasked me with EMI pre-certification of systems that will be subject to CE testing. I'm already in way over my head and have received very little training on how to actually do this.

Here's what I know:
I need to use a spectrum analyzer to determine that the Qpeaks of the radiated emissions fall under acceptable limits for EN55011. Our systems are considered to be Class A Group 1, and I'll be testing at a distance of 3m. This gives me a limit of 50dBuV/m for 30-230 Mhz, and 57dBuV/m for 230Mhz-1Ghz.

Here's my test setup:

I'll be testing the system in situ in an empty warehouse across the street. Not the best environment, I'm sure, but it is what it is. I'll be using a 48v PSU to drive an inverter in order to step the AC frequency down to 50hz, as this system is going to Europe. Inside the system is a 650w PSU and a 40w CO2 laser with an internal RF driver. I am not using the preamp right now because I don't fully understand what it's for or how it works.






Here are my analyzer settings:






This is what really has me scratching my head:



This is my baseline ambient reading with all the equipment under test unpowered. No matter what I do, the reading is always the same. Regardless of whether I'm using the bicon or log, whether or not all equipment under test is powered on, or whether I'm firing the laser at full power. Unless this is the quietest equipment on the planet, I feel like I must be doing something wrong. Not only that, but it appears that the ambient noise is already pushing me over the 50dBuV/m limit in the 30-230Mhz range.

I feel like I should have correction files set up for things like the antennas, the cable, a preamp, etc.

Thanks very much in advance.

[srb1954]

EMC test results are notoriously variable depending on the test setup. The aim of your pre-qualification testing should be to replicate as closely as possible the conditions of the open-air test sites (OATS) used by commercial test laboratories so that you can have greater confidence that your unit will pass the final test. Failing a test and having to go back for a re-test can get quite expensive and time-consuming.

I would recommend that you:
(1) Move the UUT away from reflecting objects such as the metal racking in the photo. You should attempt to move it as far away as possible but certainly no less than 3m. If you look up the testing standards there are typically diagrams showing the area around the UUT and antenna that have to be kept clear.
(2) Place a large ground plane on the floor under the UUT. The OATS test procedure assumes that there will strong reflections from the ground adding to the received signal levels. Commercial laboratories typically use perforated metal sheeting covering the whole test site but as a cheaper alternative you could use wire mesh provided the mesh is fully electrically bonded.  The ground plane should extend all the way out to the measuring antenna and a meter or two either side of the axis. Again, the testing standards usually have diagrams specifying this.
(3) Place the UUT on a wooden table with as little as possible in the way of metal fasteners and brackets in that table.
(4) Place the table on some sort of turntable so you can rotate the UUT under test to find the maximum emissions. If you have a motorised turntable finding the maximum reading can be a lot easier.
(5) If you can't move the auxiliary equipment such as your power supplies well away from the UUT then place them underneath the table in contact with the ground plane to minimise the effect of any emissions from that equipment contaminating the readings.
(6) Check that your auxiliary equipment is properly EMC qualified so that you don't get noise from that equipment confusing the readings. This particularly applies to the laptop computer and its power brick if you need to keep these running during the test.
(7) Run the mains power cables to the equipment under the ground plane so that the power cable doesn't act as an antenna coupling external mains-borne noise into the readings. Feed the mains supply through a mains filter to further reject any mains-borne noise.

Lastly, you will definitely need to include correction factors for both the antenna and its cable. The correction factors for the antenna should be obtainable from the antenna manufacturer but you could characterise the cable yourself using the tracking generator in the spectrum analyser.

[JohnnyMalaria]

Inside the system is a 650w PSU and a 40w CO2 laser with an internal RF driver. I am not using the preamp right now because I don't fully understand what it's for or how it works.

I think that could be important. Do you mean there is usually a signal that goes in to the RF driver but you aren't providing it? Is it possible that without such a signal, nothing is being amplified and so there's little emission? For example, the input could be a TTL pulse to excite the laser with PWM RF. Not providing a signal that will be present during formal testing seems like a risk.

[ted572]

Hello Hyphen-B:  I just sent you a PM with my name, location, and telephone number. I have experience with EMI measurements and would be happy to talk to you about what you are doing there. It may be helpful to simply talk about this project. Although of course you have received some excellent replies on the subject, and it may not even be necessary now. Its your choice.
I'm available if you would like to discuss anything about the task     Regards and good luck, Ted 

[Hyphen-B]

Wow, thanks for the response, this was incredibly helpful! Having now read up on OATS procedures and diagrams, I'm beginning to understand things a bit more clearly. I'll make room as far away from the shelving as possible, and I'm pretty sure we have some material laying around that would make a suitable ground plane. I was able to find the antenna factors listed on the datasheets for the bicon and log I was using, and I'm reading up on how to find the cable factor as well. Ancillary equipment is all chassis bonded to ground and looks good. Don't have a good turntable solution as of yet, but I've got some near-field probes that I can use to find any hotspots.

Thanks again, man. I'll be much more confident in the accuracy going forward with all this in mind.

[Hyphen-B]

Inside the system is a 650w PSU and a 40w CO2 laser with an internal RF driver. I am not using the preamp right now because I don't fully understand what it's for or how it works.

I think that could be important. Do you mean there is usually a signal that goes in to the RF driver but you aren't providing it? Is it possible that without such a signal, nothing is being amplified and so there's little emission? For example, the input could be a TTL pulse to excite the laser with PWM RF. Not providing a signal that will be present during formal testing seems like a risk.

Ah, I should've been more clear: the preamp I was referring to is a Com-Power PAM-103. It's a discrete preamp that goes inline between the antenna and the analyzer. You're right, we use a PWM signal to fire the laser. I'll be testing the system in three states: Idle (E-stopped), System Enabled (E-stop released, interlocks made and shutter open), and Firing (cycling the laser through a powercheck - ie. incrementally increasing duty cycles). Thanks for the reply!

[Hyphen-B]

Hello Hyphen-B:  I just sent you a PM with my name, location, and telephone number. I have experience with EMI measurements and would be happy to talk to you about what you are doing there. It may be helpful to simply talk about this project. Although of course you have received some excellent replies on the subject, and it may not even be necessary now. Its your choice.
I'm available if you would like to discuss anything about the task     Regards and good luck, Ted

Reply sent. Greatly appreciated!

[srb1954]

Wow, thanks for the response, this was incredibly helpful! Having now read up on OATS procedures and diagrams, I'm beginning to understand things a bit more clearly. I'll make room as far away from the shelving as possible, and I'm pretty sure we have some material laying around that would make a suitable ground plane. I was able to find the antenna factors listed on the datasheets for the bicon and log I was using, and I'm reading up on how to find the cable factor as well. Ancillary equipment is all chassis bonded to ground and looks good. Don't have a good turntable solution as of yet, but I've got some near-field probes that I can use to find any hotspots.

Thanks again, man. I'll be much more confident in the accuracy going forward with all this in mind.

Near-field probes can be very useful for identifying potential hot areas in your equipment but improvements made after near-field investigations don't always translate to improvements in the far-field OATS measurements.

What happens is that the cables out of the UUT act as antennas and resonance effects and orientation of those cables and reflections from the ground plane can result in considerable gain at some frequencies magnifying what might been insignificant emissions in the near-field measurements to being a major problem. This is the reason for using the turntable as it allows you to orient the UUT to obtain the maximum readings as required by the testing standard.

You should note that different orientations will generally be required to maximise the readings at different frequencies. The usual approach to find these maxima is to slowly sweep the spectrum analyser with the peak hold function on whilst the turntable is rotating. The problem frequencies identified as exceeding the specification limits in this procedure can then be examined individually.

[ted572]

Hyphen-B:

If the Power Supply you use with your 'Laser Product' is unique, designed specifically for, and required for use with your product, then of course you will have to include it in your tests with the assemblies cabled to together as normally configured for operation.

You should (and generally must, and if not - is risky) provide a detailed Test Procedure to your EMI/EMC Test House to make the measurements by. Make sure that they provide a complete list of the test and support equipment used along with manufacturer's name, model, serial number, current calibration date (very important), etc.

You will then also have to provide your clear and concise Test Procedure with any edits that may have come up from the Pre-Compliance Testing (I wouldn't simply pass on the Test Report from  your EMI/EMC house) with the resultant compliant test data to the country of use - EMI/EMC approval  authority. Be it FCC, CE, PTT, (Postal Telephone and Telegraph), Mil. Std. 461/2, etc.

I always provided detailed pictures of our equipment that was tested, along with close-ups of the it's Model and Serial Number name plate labels, and any other approval labels, etc. You may want to go to your EMI/EMC test house with the equipment being tested to take pictures of the the actual tests setups so that you  have the capability of providing them if they are needed.

I haven't done 'Pre-Compliance' testing myself (although I assisted others with it often), instead I always did my own complete Tests and Reports for direct submission to the appropriate certification authority of several countries.

Your most difficult job will be in preparing a concise Test Procedure for your EMC House for doing the measurements, and then later {with edits A/R) for submission to your customer's country certification authority.

It is possible that your 'customer or in-country dealer' will offer to take your report for submission for approval for you. Although I have never seen this happen. But you want make sure that your company is listed for certification approval, and that its not listed under the name of your customer/dealer! Just a precaution, although unlikely as it is, could potentially be a problem in the future if it occurred. I have seen some foreign dealers come up with too many shady deals.