max power for a flyback drivers? / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

max power for a flyback drivers?

previous 1 2 3

Move Thread

LAN_403

Pinky's Brain

Sat May 21 2011, 07:07PM

Registered Member #2901 Joined: Thu Jun 03 2010, 01:25PM
Location:
Posts: 837

Antonio wrote ...

Nor the frequency nor the duty cycle must be fixed. You just have to leave a time interval after each switching off of the transistor for the pulse and the energy return. The maximum duty clycle is a bit larger than 50%, but there is no minimum.

It doesn't need to be fixed, it just generally is ... and it's all still irrelevant. If you keep the turn on time (and thus the pulse energy) constant but change the frequency you can't keep it in ZVS. Although if he is really getting 12.5 Watt output at 60 Watt input, ZVS is the least of his worries.

Antonio

Sat May 21 2011, 09:29PM

Registered Member #834 Joined: Tue Jun 12 2007, 10:57PM
Location: Brazil
Posts: 644

This circuit is not "ZVS" just if you consider the energy in the capacitor when the transistor turns on. This energy is dissipated on the transistor and wasted in the basic configuration. But it is insignificant in relation to the energy that goes to the load. It's perfectly possible to use this configuration for a fixed or variable power supply using the duty cycle or the frequency to control the output power. It's also possible to operate it af fixed frequency and duty cycle, varying the supply voltage using a buck converter, as frequently seen on monitors. I have HV power supplies built using all these techniques.

cedric

Tue May 24 2011, 09:45AM

Registered Member #2941 Joined: Fri Jun 25 2010, 08:08AM
Location:
Posts: 143

a lot of nice information ,thank every one!

Antonio wrote ...

Pinky's Brain wrote ...

It's not really relevant for the application though ... a QR converter needs to be frequency controlled with a fixed duty cycle to stay in ZVS mode, which isn't really an option for a variable frequency pulse generator with fixed energy per pulse.

Nor the frequency nor the duty cycle must be fixed. You just have to leave a time interval after each switching off of the transistor for the pulse and the energy return. The maximum duty clycle is a bit larger than 50%, but there is no minimum.

actually I like the concept of Antonio,in my design I can allow the fly-back to resonate ,providing every thing is finish after 50 us or so at witch point the rest of the system use the energy store in the output capacitor ,I can also replace the hv output rectifier by a hv bridge.
for the subject of efficiency I omitted an other load on the output so it's probably higher than 12.5w out of 60,my number where also approximative,but since we speak about it ,what are the optimization witch can be done on a fly-back to improve it's efficiency? since I am doing this one myself .
now I use a lathe to cut a groove in a Teflon rood,then I feel one grove on two with winding,
Back

Pinky's Brain

Tue May 24 2011, 12:12PM

Registered Member #2901 Joined: Thu Jun 03 2010, 01:25PM
Location:
Posts: 837

Are you using an air gap?

Dr. Slack

Tue May 24 2011, 12:17PM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

Having bought stock ferrite, do remember that it may be possible to improve it for flyback use by introducing an air-gap (do not put an air-gap in if you are using a forward topology).

This is only counter-intuitive if you don't have a good understanding of what a flyback is doing, and you'll need that to put in enough, but not too much.

I came across an analogy recently that helps me.

A forward converter is like a pully-block. Map force, distance and string stretchiness to flux, current and core permeability (maybe not in that order) to see the analogy.

A flyback converter is like a bow and arrow. During conduction you store energy in a spring, then it's released to the secondary on switch-off. The string in a bow isn't stretchy, it merely transmits the force and movement to the wooden or carbon-fibre spring that you're holding. In a pulley-block, non-stretchy string is great. But if you tried to make a bow with a stiff back, you couldn't store much energy by the time you'd reached the maximum tensile strength of the string (aka core saturation flux density). That is why a bow uses a springy back, and a flyback core uses air-gap. The airgap increases the H field needed to saturate the core, the springy bow increases the distance the arrow gets pulled back for a given string tension. Result for both, more energy is stored than otherwise.

If you have bought a flyback ferrite, it might already have a "dsitributed airgap", that is more epoxy holding the ferrite grains together than for a forward core. Such a core will be less than ideal for forward use. Like using a bow and arrow with a too stretchy string. However if you've bought a forward core, it will be improved by mounting the two halves across a few sheets of paper.

The difference in stored energy can be quite large, maybe a factor of 10, so do take it into account when doing the power math.

cedric

Tue May 24 2011, 02:43PM

Registered Member #2941 Joined: Fri Jun 25 2010, 08:08AM
Location:
Posts: 143

I just get the ferrite from the post today ,they don't mention distributed gap ,only soft ferrite,the company is a serious one ,they joint a lot of documentation with the core ,according to them the max power out put of those ferrite is 2672 w in fly-back mode 8000 in push pull and 4000 in forward converter.
so fare I made the test with 1 millimeter gap (double side sticky tape) ,I just finish machining the dielectric for the new core ,I wind the copper wire to it (0.224millimeter) 1600 turns ,I will make some test tomorrow if I have enough component .
for the mosfet I am planing to use one or two of those:

and to power the system in 48 volt.

Ash Small

Tue May 24 2011, 02:48PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Further to what Dr Slack has written, I read somewhere recently that one gap is better than two. This isn't always easy to achieve though (most TV and monitor flybacks cores have two gaps). It also makes a difference which side you put the windings, but I can't remember which side it said to put them on a core with one gap, but I do remember it said to put the primary under the secondary.

The gaps in E cores are usually ground, and surface finish is important. Using abrasive paper on a flat surface should be OK, using successively finer grades.

Pinky's Brain

Tue May 24 2011, 02:53PM

Registered Member #2901 Joined: Thu Jun 03 2010, 01:25PM
Location:
Posts: 837

wrote ...
actually I like the concept of Antonio,in my design I can allow the fly-back to resonate ,providing every thing is finish after 50 us or so at witch point the rest of the system use the energy store in the output capacitor ,I can also replace the hv output rectifier by a hv bridge.

If 50 uS charge time is acceptable and all you are doing is charging a capacitor at that stage of operation you should really consider the circuit from the DoD document I linked. It uses a thyristor as a switch, which is by far the cheapest switch per amp at those time scale.

cedric

Tue May 24 2011, 03:26PM

Registered Member #2941 Joined: Fri Jun 25 2010, 08:08AM
Location:
Posts: 143

Pinky's Brain wrote ...

it's interesting but I doubt I could use thyristor at 20 khz easily..

previous 1 2 3

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob