Inductors for ZLT+

[robk]

Has anyone had any experience using J.W. Miller PM105S 100uH (or less) inductors with the ZLT+ circuit? I have made dozens of these with the hand wound inductor, but I want it smaller with a commercially available inductor. I tried a few fron Coilcraft and they didn't work too well (saturated , delayed turn-on, overheated transistor)
Thanks for any help,
Rob

 

[Reboot]

I have a 47uH on order from Digi-Key that should be here by mid week. It's a TYCO DU1330 Series with resonably low DCR and a saturation current of 500ma. It's about half the size of the hand wound version. /ubbthreads/images/graemlins/blush.gif I'll let you know if it works.

 

[Reboot]

Rob,

The inductor I got turned out to be useless for this particular application. But, I did manage to shrink the capacitor. I replaced the 150uf 10V cap with a 47uh 6V cap from the same Kemet T520 series. (Digikey # 399-3061-1-ND)

Much smaller /ubbthreads/images/graemlins/smile.gif ... and half the price as well.

Here is what the ripple looks like with a load current of 350ma. (Scale is 10mv/div.) This is basically the same waveform you get with the 150uf cap.

Using this should give more space for the giant inductor.

 

[robk]

Good information. It's tough to find a suitable inductor. I cut down the ferrite bead to 1/2 it's original size and used 3 turns of wire, and it seems to work pretty well. I've used caps as low as 33uF without any problem.
What's with the negative going pulses on the scope? What load was on the output when you got this waveform?
Thanks,
Rob

 

[Reboot]

I don't know what the negative spikes are. Maybe some reverse current sneaking through the diode when the transistor turns on?

The load is ~350ma through the Luxeon simulator (3 1N4001 diodes & a 1ohm 1W resistor). I don't have any Luxeons yet... should be ariving Monday or Tuesday.

That's very good info about being able to shorten the length of the inductor. I'm trying to fit everything inside a Brinkman Legend AA, and there is little room to spare.

Shrinking the capacitor allows me to scoot the inductor over to the middle of the board which helps. But cutting the length of the inductor would be even better.

 

[robk]

Looks good. I should have taken a picture of the one with the shortened ferrite bead, but now it's buried in epoxy. I just took it to the side of a grinding wheel and removed half, then touched up the edge on the wheel. I had better results with 3 turns (used #24 wire), which is lucky cause I can't get 4 turns of #24 thru the hole! I do want to look at the ripple on my scope to see if there is a difference. I don't seem to remember those negative spikes, but I haven't tested one in months.
Rob

 

[robk]

You got me curious, so I went out to the garage and took some measurements. This is on a giant ZLT, made to glue to the back of a Mag heatsink. It has a 100uH choke and a 33uF tant cap. Odd waveform, no spikes like yours, but still a strong negative component. And, this is at 100mv/div, quite a bit of ripple! Looks like I'll have to increase that cap a bit. Frequency is just under 8kHz, should be easy to filter out the ripple. I guess that negative pulse is just the cap discharging between cycles, a smaller choke (higher frequency) might reduce it, but it's never presented a problem! The load was an unknown bin LS white. Photos below.
Rob

 

[MrAl]

Hello there robk,

That kind of waveform comes from super cycle oscillation.
The cause of that is usually the wrong inductor choice,
but if it doesnt cause excessive heating of the
transistor there is a chance it could still work.

It would be a good idea to check the efficiency however,
to make sure you are up around 80 percent or better.
If you find it is 70 percent, change the inductor.
You also want to make sure the circuit can still start up
at low input voltages, as when the battery starts to run
down, because there is a chance that lower input voltages
might prevent the circuit from starting to oscillate,
which could burn up the transistor. This might only
occur once out of every 20 turn on's so it takes a bit
of patience to test for this condition.

In general, the toroid type cores show less loss then
the open frame inductors. You can of course experiment
with the number of turns from maybe 2 to 10.
As you probably already know, always check the transistor
for overheating when trying a new inductor and when testing
for proper start up (with lowered input voltage).

Very nice pics by the way :-)

Take care for now,
Al

 

[dat2zip]

Reduce the load ripple. Use a Ceramic Capacitor.

Good candidates are:

PCC2267CT-ND 100UF 6.3V and other similar to this. Check out the multilayer capacitors on Digikey page 734 catalog TO32.

NOTE: The Switching frequency is not that much different than the LT part I use on the Badboy and I use a 22uF capacitor. To get similar Ripple on the ZLT, I think I came up with 10uF ceramic to get similar results as the 150uF tantalum. The ESR is much better with the ceramic capacitors and a much lower value is required to get similar results.

robk,

Your scope waveforms appear to show a problem with loop stability. It appears that the converter is not controlling cycle by cycle. This could be because the inductor is going into saturation. The ZLT is not an easy circuit to stabilize. Careful component selection makes a big difference.

The inductor should have a saturation rating of around 2A to be on the safe side. What worries me is the inductor shown in the picture appears to be more of an RF type device and will saturate will very little current.

Just an obvervation...

BTW: Some of my scope pictures can be seen at:
http://home.comcast.net/~wayney8888/hobby/max167x_/converter_comparison/dcdc_comparison.htm

This page is rather dated and old. Created over a year ago.

Wayne

 

[robk]

MrAl,
I'll try different inductors, looks like it may be time to buy a Coilcraft assortment. I have some small toroidal ferrite cores, I'll wind a few and try them also. Thanks much for your input!
Wayne,
That inductor is some unknown part, rated 100uH but unknown specs. I'll play with other values and cores. Thanks for the info on ceramic caps, I'll change it out to see if it helps the ripple. Actually, is ripple that bad as long as the peak voltage isn't too high for the LS? I've looked at some circuit designs that don't use a schottky diode or filter cap.
Rob

 

[ARC mania]

Hello robk,


I just sent you an e-mail.


ARC mania /ubbthreads/images/graemlins/grin.gif

 

[robk]

ARC,
email answered! Thanks.
Rob