Electronics > Projects, Designs, and Technical Stuff

Design of SMPS - Not sure what these spikes are being caused by.

<< < (21/27) > >>

T3sl4co1l:
Looks like it's sitting pretty deep in cutoff, when unloaded.  Which makes it take a long time to throttle up.  Try putting some preload on it, like a resistor or CCS of 10% of nominal load.

Tim

Zog:
Thanks Tim,
Just mucking around with spice at the moment. Trying different ideas and different inputs to see how to drive spice, but I will do just that and post my results as usual.  :-+

edit:
I did just want to see just how hard I could hammer it too.
Nixie tubes are not that fussy about voltage spikes in the output. (to a point of course)

Zog:
Quick question guys.
I am getting in pretty deep with LTSpice now and was wondering what K coupling factor I should put in if I use two seperate inductors connected with a very short wide trace at the switch node ?
I kinda figure it won't make a lot of difference if they are really close right ?

Another one is putting in the values for the inductors ? At the moment they are just set at their nominated series resistances. Does putting peak current in make much of a difference or parallel ? I am guessing not much unless you are getting really nitty gritty and working with very complex inductor arrangements and not just my 2 effectively in series ?

MagicSmoker:

--- Quote from: Zog on September 17, 2019, 01:59:02 pm ---Quick question guys.
I am getting in pretty deep with LTSpice now and was wondering what K coupling factor I should put in if I use two seperate inductors connected with a very short wide trace at the switch node ?
--- End quote ---

Is this a dual boost or tapped boost? I kinda forgot which direction you ended up going here. At any rate, if a dual (probably best to say cascaded) boost then there is no coupling between the boost chokes; if a tapped boost then "it depends" noting that the leakage factor is approximately 1-(K2). So, for example, if K=0.9975 then leakage factor will be 0.005 (or 0.5%). This is typical for a single layer tapped-winding on a toroid.


--- Quote from: Zog on September 17, 2019, 01:59:02 pm ---Another one is putting in the values for the inductors ? At the moment they are just set at their nominated series resistances. Does putting peak current in make much of a difference or parallel ? I am guessing not much unless you are getting really nitty gritty and working with very complex inductor arrangements and not just my 2 effectively in series ?

--- End quote ---

??? Not sure what you are asking about here...

EDIT - I guess you are talking about the "secondary" parameters for the inductances? I advise specifying a realistic value for the series resistance but leave all the other stuff alone for now.

Zog:
Thanks for clearing that up Magic,
Tapped is what I am going for but only for the reason that they mention that - in general - flybacks are only good for 50W and forward converters are good for 150W. I read it somewhere can't remember where.
I have been getting good results from treez Dual Boost 12V to 220V at 100W_TAPPEDBUCK_CONST FREQ_1 though the part count is kinda high.

So me .. being me and and wanting to push the boundaries of just what I could do found this chip LT3757A https://www.analog.com/media/en/technical-documentation/data-sheets/3757afe.pdf.
What a little beast. One resistor for setting the clock. Wow.
I have made about 5 or so different versions of flyback and tapped boost but can come no where near the circuit treez gave to the group.

The problem I am having with the dang thing is that it will NOT keep the fet on long enough to get the inductor "saturated".
I have tried all sorts of things.

Frequency has no effect no matter how fast or slow you run the thing.
Different inductor ratios.
And on and on. Been at it for days and I don't think this thing will do it.

Pity, as is has a nice low input voltage of 2.9V up to 40V.
I have attached my latest version. Have not bothered with snubbers or gate control. Have tried all that and has no effect because the Fet just will not stay on long enough.

Is there a way I could MAKE is stay on longer. Is that even the problem ?

Happy snaps of my abomination in action.


Nice if you like rainbows. The Red trace is the Gate.


edit:
This part of the datasheet is kinda, maybe a clue ?
VC Current Mode Gain (∆VVC /∆VSENSE) 5.5 V/V
VC Source Current VFBX = 0V, VC = 1.5V –15 µA
VC Sink Current VFBX = 1.7V
VFBX = –0.85V

Or maybe a cap on the gate to lengthen the pulse somehow ? Zero clue on how I would do that though. Duty Cycle does not seem to be an option.

Edit: Again.
I think I have found my problem. Apart from the fact I can't read a data sheet properly.
Looks like it's already at almost 100% DC.


--- Quote ---For a boost converter operating in CCM, the duty cycle
of the main switch can be calculated based on the output
voltage (VOUT) and the input voltage (VIN). The maximum
duty cycle (DMAX) occurs when the converter has the
minimum input voltage:
DMAX =
VOUT − VIN(MIN)
VOUT

--- End quote ---

More editing.
I changed the inductor ratio to 1:3.
But would you be happy with this waveform ? I mean what on earth is causing this constant ringing in the inductors ?









Navigation

[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod