I may be somewhat off topic but i've ran into a mountain i cannot seem to get over regarding the hardware design of a switching mode power supply.
I'm using an LM2585-ADJ simple switcher chip from National in a non-isolated flyback configuration to generate 3 output voltages: 3.3,
15, and -15. The monitored (feedback) output, 3.3V, has a lot (1.3 V pk-pk) of transient noise at and only at every transition of the primary switch (off to on, on to off). Elsewhere, it is very clean.
If anyone has any experience in this area and would be so kind to offer any help I would greatly appreciate it. If so, I'll be happy to post all the gory details. Thanks!
1: is the noise really there? you can't probe a SMPS using the alligator ground lead and the normal snout of you scope probe. Pull the snout off the probe. Now you can see the tip and the shield of the probe. If your probe comes with a low inductance ground clip (springy wire ~2cm long) then slide that on to the sheild. Otherwise, use a jewler's screwdriver to make the ground connection by holding the probe and the scewdriver in a little triangle on top of the output cap. It's tricky to make all 3 connections simultaneously - having someone else there to handle the scope helps a lot. The output cap is the only place to look for the noise - anything else is common-mode and can be managed at the system interconnect level (careful return path design).
2: If the noise is still there when properly probed, then the problem is poor circuit layout. The LM2585 data sheet should have a section on layout - follow the advice on keeping the output current loop small.
A quick glimpse at the schematic shows some possible anomalies. First, you should do your measuring, and sampling, after the HF filter. You have to expect these deviations at the filter input. It is possible that the regulator has its own hidden filter in the feedback loop.
Second, you should be taking great pains to avoid any ground loops. The surge current in the 3.3 regulated rectifier circuit should not have any impedence in common with anything else. The way the schematic is drawn indicates a casual attitude. Draw that ground area in as a big fat buss, with explicit connections.
--
"If you want to post a followup via groups.google.com, don't use
the broken "Reply" link at the bottom of the article. Click on
"show options" at the top of the article, then click on the
"Reply" at the bottom of the article headers." - Keith Thompson
I would expect some deviation but my gut instinct says it should be less magnitude (see screenshots below). Although I couldn't honestly say as I have no prior experience in this area.
Agreed, layout could use some work. I tried to avoid ground loops during the first pass layout but have learned a great deal since then and I now see areas for improvement.
I will try all the above suggestions this afternoon, thanks for the responses. I'll send another PCB with improved routing soon if the problems persist throughout the troubleshooting process.
I thought I'd post some waveform snapshots as well:
What kind of capacitor is C8? It's sometimes better to use 2 or 3 capacitors to get lower ESR (and spreading heat at higher power levels.). Further more, normally I'm using snubbers across the output shottkies.
C8 is a 25 V, 3300 uF capacitor, part# UPW1E332MHH
formatting link
The datasheet specs it to have an ESR of 0.018 at 20 degrees C (my nominal operating temp). I agree with you, I probably need to spread this out across a couple of capacitors but I don't think that it is the source of my problems.
While we're on the capacitor subject, does anyone have any idea how to measure the ripple current in a capacitor to make sure it falls within the capacitors operating ranges?
I have, since posting the schematic, placed a snubber (series RC in parallel w/ the diode) around my feedback output shottkie. Values of
0.1 uF and 20 ohms really cleaned up the ringing reflecting back to the switch. I failed to mention the hardware update on my july19 post with the waveform snapshots.
Bob mentioned earlier that you can't probe a smps using the alligator clip ground lead. I followed his suggestions and he was right, the noise was not of the magnitude that was first expected (0.5 Vp-p spikes instead of 2V). This makes sense; however, I do not understand why I can pull any off-the-shelf smps and probe it using the alligator clip ground lead with successful readings.
I apologize for not updating the thread but I figured no one was looking at it further. Thanks for any input and don't hesitate if you have questions, this thing is killing me.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.