tinyswitch drain current spikes
Hi, I'm re-designing an ac-dc flyback converter because our previous design has a really annoying buzzing noise from the transformer, there are also some others features that we want to change/improve so I started measuring the old power supply.
When I checked a buzzing PSU sample I've measured the Vds voltage and the Ids current on the drain pin of the TNY279P and I've found some strange spikes on the Ids current.
The output voltages values are correct and regulated but I don't understand these spikes in the Ids current at each cycle (see attached files).
I measure the current with a LeCroy AP015 current Probe (DC to 50MHz probe) and a Waverunner6100A scope (with a differential voltage probe).
In the attached file there is a plot of PSU schematic circuit and some plots of the Vds and Ids with the strange current spikes (that are not present on the output diodes)
(there is also a plot of the Voltage at the catode of D5 diode @ +5V output).
Can you help me to understand what I'm measuring?
Note: the transformer has1,053mH for the primary inductance and 32μH of leakage inductance.
NSP=92, NS1=4 (+3v3), NS2=6 (+5v), NS3=16 (+15v), NS4=16 (-15v)
attached files:
psu_flyback_schematic.pdf: schematic file
wave1.pdf: Vds, Vcatode of D5, Ids with Vin=73VDC (undervoltage control disabled), light load
wave2.pdf: Vds, Vcatode of D5, Ids with Vin=125VDC
wave3.pdf: Vds, Vcatode of D5, Ids with Vin=317VDC
Comments
The spike you're seeing right at turn-off of the MOSFET is most likely the clamp current. This will show up on your waveform capture if you insert your current probe into the wrong part of the circuit. I've attached a copy of your schematic with some notes on where it would be best to insert your current probe.
Regarding the transformer noise: Have you tried varnishing the transformer core and windings? If you have and you're still getting the audible noise, you might try re-designing the transformer so that there is a lower peak flux density. Transformers can actually be somewhat noisey if not designed correctly.
-The Traveler
Thanks Traveler for your reply! The probe reading for the waveform capture is already in the position you suggested, the R6-0Ω is just a jumper for a current probe loop insertion near the drain pin of the tinyswitch... I didn't thought to the clamp current but I will check this too, maybe something is missing...
Regarding the transformer noise: yes I tried varnishing the transformer windings (not the core) but the noise is audible... I think I'll try to re-design it.
cyba -
If you're indeed locating the current probe at one of those locations, I'm a bit perplexed by by drain current waveform you've attached. It doesn't look like transformer saturation so the only typical issue that causes a waveform like this is putting the current loop in the wrong spot. One fool-proof wave of double checking is to lift the drain pin from the IC out of your PCB. Attach a small length of wire and use that as your current loop.
For the transformer, what varnish method are you using? Are you doing a simple varnish dip or actual vacuum impregnation?
If you can, please attach any PCB layout files (PDF) and PI Expert design files (including transformer documents). Having these would help me pin down where the noise issues might be coming from. Typically, the transformer can be a main source of noise in a power supply. Other sources, however, include, the clamp circuitry (ceramic capacitors are notorious for this) and your input EMI circuitry. A *really* easy way to see where the noise is coming from (if you're not 100% sure it's the transformer) is to see if you can get your hands on a stethoscope, or even just a lenth of some plastic tubing. Move one end around different areas of the power supply and see if certain componnets are making more noise than others. If you have a microphone and spectrum analyzer, you can use these together with the plastic tube to see what frequency the noise is at as well.You just want to make sure you keep the microphone away from the power supply so it's not picking up radiated EMI.
If you can't do any of that, try to get a drain current and drain voltage waveform plot at whatever condition you hear the noise when it is at its loudest.
-The Traveler
Hi Traveler, thanks a lot for your help, I confirm that also with a small wire directly from the drain pin to the transformer primary pin I can see the strange spikes. Our supplier make a simple varnish dip of the transformer. For the last requests, I have the pcb layout and other measurement too, but can we continue via private mail?
cyba -
I'm following up with a private message so you can email me.
-The Traveler
Hello? someone could help me, please? any suggestion?
thanks.