Next Step, Power Control

[QUAZ HOLT]

On to the next project. Problem posed, can a wind mill power supply be made from old parts and junked equipment. Here is a lot of old equipment.

And still some more of these computer power supplies.

"Concept Drawing of How to Control High Voltage Power"
The idea seems simple and comes after the wind-mill is producing power.

The next step is to start collecting parts that will do the job. There are lots of these things in the supplies and monitors. All different sizes and types and who knows the wires in them? These are the good ones shown below.

Can you tell the difference? The first ones are high frequency type material and the second ones are low frequency material.The second ones will not work for high frequency transformers.

Now to make some useful parts out of the old transformers. First I had to break the core in half. This can be a lot tricky. I lucked out this time and picked one that cracked like an egg. I collect these things and then add them to my collection.

Now time for a circuit drawing.

This is the basic inverter circuit and should give me a simple understanding of the parts required. There is a FET driver power supply in the next circuit diagram.

Winding the transformer is one of those things that I have done many times before. This time it will be wound differently. Guess work I guess. Wire size, let me see what size of wire is in stock. Have some 18 gauge and 20 gauge. The larger number is smaller wire and the smaller number is larger wire. Looks like I will use the thinner wire for the primary windings and the thicker for the secondary. Wire comes in a large roll if you want to get it cheaper.

Start winding or I will be here all day.

62 wraps of # 20 gauge 31 wraps per row. This is done twice to give me two complete coils of 62 turns # 20 gauge. Yellow mylar tape is added to insulate between the primary and secondary coils. There are 30 Turns #18 gauge in the secondary and two of these coils are added to finish the secondary.

This white tape is also use to hold and insulate the coils.

The core on this one has a small air gap and you can see it in this picture

A finished core, hand wound and ready to add the the circuit.
Hand wired inverters, I was told, will not work. Funny thing this one works fine.

Finished Proto Type

Just a bit of testing now.

And some more testing.

The project works but has a few draw backs. This type of work is akin to witchcraft. I am looking forward to making the next one. 45~120 VAC in 15~40 VDC out. Any questions? Bill

[Nando]

Bill:

Questions ? MANY !!

I work with SMPS and presently from 20 watts to 5 KW ( 20 to 500 KHZ +) depending on the project.

Yes hand wire transformers DO work, as I have been doing for about 4 decades and as the same way that prototypes are made and high level productions of the transformers you have in one of the photos

The transformers you took apart are machine winded for speed but manually finished ( lead soldering, tape wrapping, etc ).

The converter using the MC34063, why using a 2N3906 ?. It is just a 300 ma device ?.

Were you experimenting ? Specially driving the PWM controller from an AC supply.

Photo number five, showing some common mode chokes and cores can under certain conditions be used for power supplies.

I use the common mode chokes for a push-pull voltage doublers with good output power

The one on top of the photo seems good for about 10 or so watts at around 30 - 40 KHZ, best way is to use an oscillator and divider driving 2 Mosfets at a frequency that does not allow core saturation prior switching.

A CD4013 and a CD4050 or CD4049 driving some medium power MosFets

What are the drawbacks you have ?.

Nando

[willib]

timely topic Bill

i found this gem just today , it takes a lot of the mystery out of SMPS design.

http://schmidt-walter.fbe.fh-darmstadt.de/smps_e/aww_smps_e.html

if you click on the home button it brings you to a whole page of , boost , buck boost  ect.. converters

[QUAZ HOLT]

Nando, The main drawback is the range of voltage that this switcher will operate at. The 20 volt stepup/down converter runs the pwm and cuts in at 6v and has a range to 40v

the step up/down circuit operates at 3 ma 12 volt and total circuit power is 10 ma. This was an attempt at making something and not having to leave and get any new parts. Another drawback: One of a kind project. A switcher that I used from an old monitor required no modification and preformed better than this circuit. That switcher was a flyback type and cuts in at 20vac and puts 20 volts out over the range of 20vac to 140vac. I was seeing if this type of switcher would work. Million projects and switchers are not my strong point.

Problem posed. Good voltage and frequency range, low operating power requirements, high power output and short circuit current limit. One other thing low cost. Bill

[oztules]

Hi Bill,

Happy to see your at it again. What kind of power are you aiming at. Are we talking tens of watts or hundreds at this stage.

A bootstrap from the windmill dc and an extra winding on the tranny for the supply voltage would make it independant of the ac.

The transformer ferrite you used.... was it the one with the air-gap on the inner leg? Is this necessary to prevent saturation in push pull of this type.... I felt it to be only necessary on the buck and boost types.

I haven't tried at these frequencies, but a voltage doubled input as per the normal computer supply, may give more useful dc voltage to play with out of the mill.

That tl494 i built could be used as a buck with a few mods. It has current reg, voltage reg etc.

I can  research it further and see what i can come up with if you wish. The transformers I wound for the booster were fine up to about 600w. (300w each...twin)

In other halfbridge types I have built (900w)  I wind the primary inside and outside the secondary. This gives tighter coulping, (less leakage), and so better efficiency.

There's always computer power supplies with a few changes, but this would not be instructive or universal for others to try. Perhaps  a common 250w AT power supply and create a board that will use the tl494, the transistors, tranny's and recs, and from this , a  universally buildable device may be done, that others could emulate with no other components.

Oh, and what kind of power was the waveforms taken at, and from where (switcher out, or across the primary, or is it the secondary.... at what load.

I like seeing this kind of thing happening

.......... oztules

[Stynus]

http://www.otherpower.com/images/scimages/5797/pscon5.jpg

I think you need a feed-back loop.

otherwise your battery's could get overloaded (to much voltage) or you can have not enough voltage.

[Nando]

The proper design procedure has to be "DESIGNED" to be able to produce a PWM controller with the parameters one wants.

Input Voltage range

Output Voltage range

Voltage ratio output/input

Output Current

Power conversion range

Then one needs to treat each section carefully and develop the circuit to cover the desired parameter.

Like 100 input volts --

MosFet selection (Voltage and current) one or more units

PWM controller and voltage range

Secondary power supply for the PWM controller

Voltage isolation MosFet driving circuit,

And so on -- so here are some of the steps that may be necessary .

Develop something with very limited number of available parts is generally a case of futility, for this reason alone, I have developed a set of minimum parts in my personal "inventory" to be able to do a workable study of a project at hand..

Inventory like a complete set of resistors values ( 100 Resistors per value), a set of MosFets with certain characteristics to cover ample range of operation.

A family of Integrated Circuits .

Cores and bobbins for evaluation purposes

Nando

[QUAZ HOLT]

Stynus, Thanks for posting your feedback. This problem I have posed is most likely an impossible task. I think that most things done were said by some one,"that will not work". Could you give more info on how you would do the task.
1. Wide voltage range input 20 to 130 volts AC/DC.
2. Low circuit power usage.
5. Cut in output 14 volts
4. High current output
5. Current limiting.
6. Low cost DIY
Bill

[QUAZ HOLT]

Oztules, Bill here
Question 1. lots of watts
Question 2. AC on circuit drawing is from wind mill.
Question 3. Air gap is FYI and don't know exactly why.
Question 4. Fly-back, Buck, Boost, Switched capacitor, Doubler and Computer controlled.
I will try some or all of the above and call the finished unit "A Windmill Voltage Manager".
Tall order number 10875.

[QUAZ HOLT]

Nando, I have piles of stuff and the pile keeps getting deeper. I was not going to post this circuit because it only worked. It showed me what is required to manage the voltage from RE sources. The unit on question was windmill self powered and did put amps into the batteries. Have you worked with switched capacitor pumps to do this job. I will keep posting simple circuits that work. Bill

[QUAZ HOLT]

Willib, Cool calculator. Thanks :>) Bill

[Nando]

Switched pumped capacitors are QUITE impractical for POWER conversion at levels higher than a few watts, efficiency problems and capacitor ESR, as well as, capacitance levels.

Magnetics are the real practical solutions for high power conversion as you can be able to observe with the transformers from computer power supplies that you have dis-assembled for your projects.

Nando

[Flux]

Just a couple of comments that may help.

Forward converters really need a choke (inductor) in the dc output to the battery, especially for larger outputs.

The other thing is the 1k gate resistors, these are far too high, the gate capacitance will slow the drive pulse and your waveforms show it. High impedance gate drive also means that the gates are exposed to transient spikes from the miller capacitance resulting in spikes from the rapid voltage change on the drain being induced on the gate. If they exceed the gate breakdown voltage the fet dies.

Try reducing these to about 22 ohms.

I would keep with the air gap in the core although in theory it is not strictly necessary. It will give a margin of safety against core saturation.

I didn't read your transformer construction but you should bifilar wind the primary and the secondary and even then you may need a RC snubber across the primary.

flux

[willib]

I've run a simulation on the circuit

i've chosen to use 24V as the battery voltage with an internal resistance of 10 mOhms , because after the hookups between two 12V batteries this seemed a reasonable value.but i could change it if somone knows better...

V(n002) is the voltage from the pwm

V(n004) voltage on the battery

V(n003) is the voltage at the output of the inductor , as you can see there is a tad of oscillation, not exactly sure why ...but i'll try to find out.

I(v2) current into battery

 V(n004)is the voltage at the battery.

i'm still learning this simulator ( it is a neverending process i'm afraid), so i'll try to find out how to average the power

questions?

[commanda]

Bill,

That type of converter circuit will not work over a large input voltage range.

The ideal approach, as I see it, is to cascade two converters. First is a boost converter to take the windmill voltage to some high dc voltage. Then a push-pull down converter like you've built to the battery voltage. The second converter should have an mppt derived control voltage.

I have built a prototype mppt controller, which I briefly describe in this thread

http://www.fieldlines.com/story/2006/3/12/14840/1315

I haven't published the circuit here yet because it's just a prototype and not fully tested with a real world application. Please contact me directly if you'd like a copy.

If your generator was similar to a Fisher & Paykel, where the output power curve was linear with respect to rpm, you can bypass the fancy mppt controller and use a tacho output voltage as the control voltage.

Amanda

[QUAZ HOLT]

Flux, I have used these type of switchers mostly in the step up mode to run power for H switches. They have been bifilar wound and have up to 14 secondaries. The problem is these ones I have made are for control voltage and very low current. Higher power seems more difficult to get to work properly. The wire is harder to wind. Willib post is good but can a person get these transformers and at what cost. Will add the 22 ohm resistors and the choke to the circuit and do more testing. I think one of the chokes from a old power supply will work. Thank for information on the air gap. What I need is a high power variable frequency and voltage output generator to simulate all types of voltage conditions.

"How many volts do you want and how many amps." Bill

[QUAZ HOLT]

Nando, sort of thought that. If it would work good it would have been used in high power stuff. I had made a nicad charger that could charge 4 cells at 5 amps each cell. This was at max 2 volts and 5 amps. Capacitors would heat up at a high frequency. For that low power it worked. 2 volts at 5 amps = 10 watts. I was able to change the charge current by adjusting the frequency and capacitance. Bill

[QUAZ HOLT]

Nando, I have read about nano-capacitors that are being developed. They do not suffer from total discharge as a battery does. Have you heard about these things. Bill

[willib]

"We have fabricated high-density arrays of individually isolated semispherical nanocapacitors consisting of porous anodic aluminum oxide (AAO) layers as dielectric materials and carbon nanotubes (CNTs) as electrodes. It is shown that the nanocapacitors made with the CNT electrodes exhibit much better C-V behaviors than those without the CNT electrodes. The improved electrical behavior is explained in terms of the use of the CNT electrodes deposited within the porous AAO layers. The capacitance calculated using semispherical capacitor formula is in agreement with the experimental value. ©2005 American Institute of Physics "

if you want to know more you gotta pay , bastards!!

carbon nanotubes are allready used in super capacitors .

[QUAZ HOLT]

Amanda, never used word or power point ever. Must be MS virus. Bill

[commanda]

I had to think about this for a few minutes. I sent Bill a pdf of the circuit of my prototype mppt controller. My email inserts a signature in all outgoing emails, which is as follows;

> Please avoid sending me Word or PowerPoint attachments.

> See http://www.gnu.org/philosophy/no-word-attachments.html

It wasn't directed at Bill specifically, everyone gets it.

Amanda

[dinges]

Amanda,

I was wondering, for the downconverter there are plenty of switchmode power supplies around (old PC PSUs, or their bigger brothers). I know these things are voltage controlled and not current controlled, but still, possible to use when used together with a dumpload?

If the above is correct, all we would really need is the upconverter to the high voltage. (like 310VDC here in Europe)

I've seen plenty schematics of power converters, but none of the last kind.

Peter.

[commanda]

Basically Yes.

This chap in Germany shows how to re-build an old pc power supply.

http://www.qrp4u.de/docs/en/smps_new/index.htm

This is yet another project I have half-completed.

By using the output voltage from my mppt controller to modulate the output voltage from this power supply; basically we use the line resistance going to the batteries to turn the constant voltage output into a controlled charge current. You probably need an understanding of Thevenin-Norton equivalent circuits in order for that to make a whole lot of sense.

The circuit whipped up by quaz holt

http://www.fieldlines.com/story/2006/7/9/23635/47768

is also an ideal candidate.

For the up-converter, almost any power supply chip, a single inductor, and single fet will work. Oztules built one http://www.fieldlines.com/story/2006/5/4/31543/52124 following on from work done by Flux http://www.fieldlines.com/story/2006/3/17/185646/194

I've been playing with a PFC (power factor correction) boost circuit using an IR1150, but haven't figured out the correct amount of invective to use when swearing at it to get it to actually work :-)

Amanda

[willib]

a while ago i found a real nice coil calculator on the web , but misplaced it .

well last night i found it again , on the other box

http://www.otherpower.com/images/scimages/2965/coil_dims.JPG

it just does air core coils , but at least it takes some of the guessing out of making a 137uH coil.

in the photo are three runs all of 137uH

the first two are with 20 gauge wire the last is with 10 gauge