Forum Replies Created
- Posts
Hi John,
I do this all the time. Its a simple comparative test so antenna choice only needs to be something that covers the frequency range of interest. It simply has to remain in the same relative location/orientation.
Simply run the SA44B with a couple of traces – one in max hold and the other in average mode across the frequency range of interest. An EMC analyser peak mode is the same as max-hold, and quasi-peak will be less than peak but not less than average.
You just need to run each test for a little while to get a build up of most of the spikes or transients.
I then export the ambient (your DUT off) and DUT active respective traces out as CSV and post process with Excel/Matlab etc. This method is usually sufficient, however it can miss burst events due to the sweep times, and if your unit has any emissions in the region of persistent signals (FM radio for example) these may be masked. You can use the same method with the real time mode though to narrow in on any peculiarities. If this is going to be a common occurance for you and your product, then, if budget allows, the BB60C may be better (Spike includes the EMC Precompliance tools for BB series).I’d still like to see the EMC tools opened up for the SA series, just with a disclaimer around its shortcomings. I’d still have to run a compliance test anyway, it would just be more convenient.
Probably need to do this in something like Matlab (communications toolbox possibly, unless they have a new toolbox for LTE) and just feed it the raw I/Q data
I’ve been away from RF-land for a bit, but needed to use my SA44B today… I upgraded Spike, found more new features and thought I’d check with the forum for interesting news, tips, tricks and conversations.
Oh WOW. 1THz/s… WOW.
The BB60C is considerably better than the SA44B, but the new SM200A looks significantly better than the BB60C!Will be checking in more regularly to see how this product develops!
Well done Bruce, Andrew, Justin and whomever I have left out!! Keep up the good work!
I tried a bunch of things when it was disconnecting and nothing appeared to help. Yet it seems stable now… I don’t know… and I go to investigate the current behaviour for my last point, only to find its all crashed.
Okay that zero span fix sounds like it will do the trick. I’ll wait patiently…
In R/T mode especially, can a preferences setting be that the freq change up/down arrows increment centre frequency by the current span?? Seems to be 20MHz (in all modes) at the moment. Maybe a drop down with 20M, 10M, 1M, Span, 1/2Span, userval?
Hi AJ,
The connection issues are on the SA44B. Typically both difficulty in connecting and connection dropouts. Though today, they haven’t occurred yet… which is odd.
We are trying to reduce current consumption and one of the largest contributors is the products receiver open time.
If we can capture the trigger Tx (up to four randomised re-try events over ~8sec) and the receiving unit’s ACK we can ascertain the minimum and maximum window times. So for me, a 10 second capture would be sufficient.
Oh and the other useful thing would be a real-time button under the Full Span and Zero Span buttons.
Thanks
So the EMI precompliance module in Spike will functionally work with the SA44B? You’ve just assessed that it doesn’t have the same degree of confidence so have disabled it?
Can you not enable the module for SA44B, but when the user clicks on it, bring up a disclaimer dialog? Or could you not rename the feature for SA44B users to “EMC/EMI pre-screening”? Or both?
I, and I’m sure others, would use the feature if it was available for the SA44B, even with the caveats.
The software module (tables/markers) would just make it easier to document potential areas to look at. Log freq scale make it match to the compliance lab outputs, etc.
When I did pre-compliance testing on our product, I reduced the span to smaller chunks and left it sweeping with average and max hold on…
I then wrote down all the peaks I thought were problematic, and then drilled down and, in real time analyser mode, looked at the areas of concern.Yes that’s been the plan to date.
Though I saw the recent announcement of EMC/EMI precompliance support in Spike for the BB60C…
Looks great from the screen shot!! Well done to all involved!Any plans to roll this into the SA44B?
Soon?
Pretty please!!Hi AJ,
Ah yeah – that’s what is happening… simply a limitation in RTBW for SA44B?
How do you handle this in the BB60C with a 250kHz span setting?Hi Mike/Randy,
For EMC debugging I use average (set to 64 counts) and max hold traces.
I take an ambient capture, then a DUT on, then an ‘improved DUT on’ sweep. There are 6 available traces so this works out quite well.
As AJ has suggested…You’ll find EMC lab’s receivers, with their preselection filters, specific RBWs and defined QuasiPeak, Average and Peak detectors, that you wont reproduce EMC receiver results with a spectrum analyser.
However the BB60C is more than up to the task of EMC debugging.
Feel free to email me directly with some more specifics at
apollyon25 at hotmail dot comI’m more than happy to give you a few things to look at.
What other bits and bobs do you have for EMC?Jared
Cool – about to install the latest version.
BUG?
On the channel power setup, I have 3 counts, width set to 50k and Spacing set to 100k, however if I set the width to 51k I only see one shaded region…
count, width and spacing are displayed as
fc-spacing +/-width/2,
fc +/-width/2 and
fc+spacing +/-width/2
With these settings I should be able to increase the width right up to ~99k (and have a 2kHz gap between channels).
Do you have a sanity check in there causing this? Should I not be able to run the channel boundaries almost adjacent to each other??Hi Justin,
Thanks for the reply.
I was looking at making a 250kHz wide adjustable BPF but that is way too tricksie. And making a switched version doesnt help, requires a massive number of filters…So I’m just looking at breaking it up into the standard CISPR22 intervals with selectable HPF and LPF sections.
I looked at mini-circuits but that approach is going to add up pretty quickly.
I’d love a BB60C, however I cannot justify it for a personal purchase as it is compounded by the requirement for a new laptop too (to get USB3).
The office has shelled out for a NI VST but this is currently used in the factory for 3G radio compliance validation.Cheers guys.
My ebay one turned up and once I repositioned the GPS antenna to the window sill it locked in pretty quickly.
I’ve got a couple of NDK oscillators as samples (DOXCO/OXCO) so I’ll make up a board to run them and then see how it looks compared to a 3ppb OCXO.
I’ll keep an eye on ebay for a rubidium standard as well.Cheers Justin, I had a look for a Rb standard, but nothing available at the time. I found a nice LT part to convert sine to square and not add too much phase noise (-103@100hz). You may want to add some detail info in the documentation for this input…
Thanks for clarifying!
JaredUpdate:
So the retail price of a entry-level commercial GPSDO starts at around 600USD, which with currency conversion, freight and local taxes is way more than I want to spend on it.
The ebay ones, upon a lot of internet reading, seem to actually be quite good, so I have one of those on its way.It outputs ~+13dBm and has a sinewave output.
I expect that is a little too much for the input on the SA44B (it’s at home so I cannot look at its back)?
Also does this prefer a sine or square signal?I use a combination of the mRS VNA, the SA44B as a receiver, and the VSG25A as a transmitter.
Works well so far…Jared
