Re: inductor sizing


"Turbo" <tech@xxxxxxxxxxxxxxxxx> wrote in message news:af6f96a9-30d5-4355-83af-8b2926333c9d@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
On Mar 23, 5:42 pm, "Harry Dellamano" <har...@xxxxxxxxxxxxx> wrote:
"Turbo" <t...@xxxxxxxxxxxxxxxxx> wrote in message

news:11a984c8-1b25-4ab8-ac64-1db4b55fd3ad@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

> Ummm... Well I had thought of that approach, however... I was
> looking at something considerably less complex, I hope. What I am
> looking at doing for the moment, I do not care about input ripple or
> noise. I was planning on using simple 60hz input rfi filter to a
> bridge rectifier, capacitors and maybe ainductorto help with ripple
> on the dc bus.
> Here is where my question starts.
> Similar to the original question, he had a DC supply to a mosfet
> bridge then aninductorout. My output needs to be very clean with
> <2% THD+N. I however am not looking to recreate AC. Just clean
> filtered DC between 20-50A. The mosfets will be PWM controlled
> 200-250khz and act as current control \ limiting. I will try to post
> an elementary schematic this evening. I was wondering if multiple
> lower current LC filters and mosfets in parallel would be better than
> a bank of mosfets feeding 1 LC or LCLC Circuit and what values for the
> LC should I look for.

> Thank You in advance
> Cory

I still have a problem with your requirements, so this is a one of a kind
with no AC input EMI requirements. When we state THD, we normally mean input
current harmonics. The Vdc output normally has very low AC input harmonics,
reduced by the loop gain, which are ignored but does have harmonics of the
switching frequency (200KHz) which we call noise. So <2% output noise means
at 160Vdc we can have 3.2Vrms or 9.0Vpp noise, not very clean but workable.
There are many solutions, depending on the parts availability, size, cost
and circuits that you are comfortable with. Please post your circuit and we
will critique.
Cheers,
Harry

I Apologize for my lack of proper terminology. I am trying to replace
a large water cooled pass bank in a laser power supply with the pwm
circuit described. The pass bank consists of 35 transistors acting
like one giant variable resistor. The dc voltage in the tube can very
with tube pressure and with temperature. Hence the range of voltage,
however, the current must be regulated and as clean as possible. <1v
p-p ripple is ideal if possible.

http://www.twincityturbo.com/plasma.bmp

The resistor in the diagram is a current shunt and will be for current
measurement and monitoring.
There will probably be a bank of 5 or 6 mosfets as the switch. And the
diode and inductor can either be one per mosfet in parallel to reduce
the size and current of the individual inductors and diodes, or one
big diode and inductor....
My other option is to use a large IGBT I have from work that has a
very large current capability but the switching frequency would have
to be ~10-12kHz. 1 Diode and 1 LC or LCLC circuit.
This is where my question comes to play.
If you could please recommend a value and a style for the inductor and
capacitor combo I would be greatfull. I am shooting for the multiple
mosfet and multiple inductor combo because I think it would provide
more options for inductors and the parts are much cheaper. Not to
mention the size...
As far as the drive circuits and support electronics, I work for a VFD
(Variable Frequency motor Drive) Repair company. The drives range
from fractional Horse Power to several hundred Kilowatts. I am
comfortable with the design and implementation of this kind of high
powered circuits, but I lack the engineering background to spec the
proper filter components to keep the output as clean as possible. I
would ultimately like to use PFC and a HF transformer or 2 as a DC-DC
converter, But I need a basic design now to get rolling for testing
and experimenting.

Cory

Ah, the design is getting simpler. Let's see if I can restate the
requirements.
Given a 300VDC input supply, Design a Buck converter with the optimum
topology to control a 20A to 50A output with 130VDC compliance. The output
must have <1Vpp (<0.35Vrms) ripple.
I have no simple way that I can look at a requirement and select the
optimum topology or even one that I can live with. What I would do is wack
out maybe three designs of the power section only and compare them for cost,
size, parts availability and friendliness. I would look at a single Buck as
your drawing shows, then a 2 phase (2xQ,D,L) (180d interlaced) and 4 phase
(90d interlaced). In SPICE, this can be done at maybe four hours per design
then another 4 hours to percolate the designs into my favorite. At this
point I then could design the output inductor knowing all the AC and DC
currents and voltages needed. Asking someone to immediately optimize is very
difficult. But looking at you output ripple requirement, I would bet on the
4 phase puppy but maybe that is not friendly to you so you my choose the 2
phase.
It is important to note that the input ripple current in C1 when operating
at 150V and 50A out is 35A for the single Buck and 0A for the 2 and 4 phase
topologies. Your design must calculate all ripple current and some you may
not want to live with.
Sorry I cannot give you a simple answer.
Harry

.