New to Coilguns - Coilgun Buildlog (Charger apparently fixed) / Electromagnetic Projectile Accelerators / Forums

Forums

4hv.org :: Forums :: Electromagnetic Projectile Accelerators

« Previous topic | Next topic »

New to Coilguns - Coilgun Buildlog (Charger apparently fixed)

first previous 11 12 13 14 next last

Move Thread

LAN_403

GhostNull

Fri Oct 29 2010, 12:26PM

Registered Member #2648 Joined: Sun Jan 24 2010, 12:45PM
Location: Australia
Posts: 291

Found some for you:

Be sure to check the power dissipation rating too, not just current

Quick guide to searching for components on Farnell:
1. Find out what you need, the more detail on the required specifications of your component the better

(Optional) On the homepage (at least in Aus) there is a link along the lines of "Last chance to buy! Reduced prices on selected product lines." These are clearance items. They are usually much cheaper than normal components and much fewer (less to search through)

3. Go the the Catagory you want,

4. Tick "In stock items only", this will get rid of the out of stock items which are useless and only mean more to search through. (optional) you might also want to tick exclude extended range items, these items usually take awhile to come in.

5. Click show results
a. If there are many items and most are irrelevant continue to step 6.
b. If there are only a few items then you may want to click to order by a rating you want higher, e.g. power then scroll through (to find what you want)
c. If there are many items that all seem to fit what you need, you should click sort by price then scroll through (to find what you want)

6. When there are too many items that are irrelevant to what you require, you want to get rid of them but still allow suitable items to appear so what you need to do is select some filters to allow as many as possible without getting useless items turning up in great numbers. This is where experience comes in, you have to play around a bit with the filters and often tick heaps of them to return the items you need.

BTW Farnell has changed to element 14 in Australia.

Artikbot

Fri Oct 29 2010, 01:22PM

Registered Member #3247 Joined: Mon Sept 27 2010, 09:42AM
Location: Spain
Posts: 137

Thanks for those references!
The IXYS is shipping from Germany, that's interesting for sure, with relatively low shipping costs I can pack more items with it :P

Those you specify are a pair of amps above mine, will that be enough? The IXYS one can dissipate as high as 300W, while mine stays around 190ish.

To say the truth, I find Farnell is quite chaotic to browse through

But it's got an amazingly wide catalog, and that's impressively useful. Thanks for the advice on how to make my browsing task easier ^^

Returning to topic:

Something here is not working as expected. I initially wanted to do a 30W charger, but the mosfet is driving over a hundred watts through itself.
Maybe the problem here is inappropiate current limiting, don't you think? It can't drive more current it is allowed to. If only three or four amps made it through the inductor, no 14 amps would be drained by the mosfet, eliminating the need for such a beast of mosfet and for such an enormous heatsink.

I know this is going to make the circuit inefficient, I know it should limit itself without any need for additional limiting components. But if it doesn't limit by itself, it's gonna be me who needs to limit it.

High wattage resistors right before the inductor maybe? Current limiting IC right after the battery maybe?

I've been looking in Farnell and I've found some high power resistors, such as this one:

25W seems reasonable to me, it's a good power to charge 3.6mF relatively fast. And I also avoid the MOSFET to be the current regulator and the main heating component on the system.
Also, that resistor (used lots of them on mechanical variators in rc model cars

) will work just fine with minimum heat ventilation, such as a small fan placed on the side of the charger.

What do you think?

GhostNull

Fri Oct 29 2010, 10:02PM

Registered Member #2648 Joined: Sun Jan 24 2010, 12:45PM
Location: Australia
Posts: 291

Look back, remember, in a boost converter current is limited by the frequency and inductance. Try upping the inductance to about 300uH or more

You should be able to figure that out =/

Artikbot

Sat Oct 30 2010, 12:43AM

Registered Member #3247 Joined: Mon Sept 27 2010, 09:42AM
Location: Spain
Posts: 137

I thought about that as well, but I had something rounding my head that discarded that option.

More inductance means a larger inductor (<-- obvious afirmation is obvious), doesn't that mean that I won't be able to reach circuit's saturation>more cycles to achieve the same output voltage?

Maybe what I need is just re-read the whole post again, rebuild the inductor using a larger toroid+MUCH more wire, replace the comparator, replace the resistors, and then try to switch it on again.

At this point, my head is like a boiling pressure pot, it's go so much stuff on it mixed up with university stuff, that it needs a fairly nice effort to take something clear out of it.

In conclusion: Bigger toroid+more wire. Okay, let's do that and see what in the bloody hell happens. I'm now at about 120mH, I'll try to plug 300-400ish and see. I got some beefy toroids and some spare wire around my desk, so there should be no problem to do it.

Oh god, thanks Ken for your continued support :)

ScotchTapeLord

Sat Oct 30 2010, 01:15AM

Registered Member #1875 Joined: Sun Dec 21 2008, 06:36PM
Location:
Posts: 635

Artikbot wrote ...

More inductance means a larger inductor (<-- obvious afirmation is obvious), doesn't that mean that I won't be able to reach circuit's saturation>more cycles to achieve the same output voltage?

Wut.
Anything that decreases the current will cause the capacitor to require "more cycles to achieve the same output voltage"

Less current at the same voltage and frequency means less energy out.
Energy of a capacitor = .5*C*V^2 ... so less energy into it means it's not charging as fast.

The more you put in, the more you get out for a given time, and vice-versa.

GhostNull

Sat Oct 30 2010, 04:58AM

Registered Member #2648 Joined: Sun Jan 24 2010, 12:45PM
Location: Australia
Posts: 291

What? 120mH? I hope you mean uH

When trying to find out how the resistances, inductance and all that effect the charge time, a proper sim comes in handy. I've made A sim of the boost converter for LTspice since I need a new charger too:
]boost_converter.zip[/file]
Inside this zip is the schematic file for LTspice. To use it:
Open it up with LTspice

Click run (icon of running person), a large black window should come up with a scale and stuff.

If in the (I forgot the proper name) bar thing at the bottom of window is says running dampened pseudo analysis press esc

It should say at the bottom something along the lines of "Simulation time = [some measure of time] Transient Analysis [some percentage] done...."

In the schematic window, when you mouse over a net (wire) a probe icon should appear, click and it will display the voltage waveform of this in the other window. If you mouse over a component a current probe icon will come up, click to show current.

Note: you do not have to wait for the sim to complete. Just charge up the capacitor or what ever you want the circuit to do. Then you should probably stop the sim

Once you got it working try changing the values of the inductor (right click it)

I made the output capacitor (C3) 5uF so what you don't have to wait FOREVER for the sim to finish charging the capacitor. I recommend you keep it at a low level, a 180mS is about 227MB of space

You can also make the sim stop itself sooner by editing the text at the bottom left ".tran..." right click it and the rest should explain it self

Artikbot

Sat Oct 30 2010, 11:50AM

Registered Member #3247 Joined: Mon Sept 27 2010, 09:42AM
Location: Spain
Posts: 137

Oh yes, 120uH, sorry for the unit misplacement :P

I've built a new inductor, it's about... 70 or 80 turns of .5mm gauge wire, haven't measured inductance (gotta buy an LCR meter), but it must be way above the older one, since the toroid is twice as big.

Looking at your desigin, I've seen that you've put a capacitor in the Voltage control pin, is that the decoupling cap you were talking about? Now I understand what's that pin made for!

So, I'm gonna install the +5V voltage regulator to feed the non-inverting input, recalculate resistors for the inverting, and install the capacitor in the Vcontrol pin. Does it need to be so small or it can be bigger? I've got ceramic caps and polyester caps of various capacities, salvaged from various discarded power supplies

I guess no electrolytics can be put there

The simulation on LTspice is, for sure, way better than the one in Yenka. It is also very useful to see the behaviour of the circuit (pulse durations, voltage rise and fall) which couldn't be seen on Yenka. Possibly because on LTSpice everything can be slowed down to the millisecond time unit. I definitely need to start making myself comfortable with LTspice.

So, hands to workshop and let's start remaking this. Oh, and the new comparator should be here for wednesday, it's an LM393 dual comparator. And the schottky diode aswell, specifically a C3D03060E (Toshiba or Fuji made I believe), rated at 600V 3 Amps. That should increase efficiency aswell right? It's got WAY better recovery times than 1N5408. The 1N5408's will be placed in reverse-biased disposition to increase pulse duration and absorb back EMF. 1kV 3Amp should be able to absorb that back EMF pulse, considering it's not just a single diode, but maybe 2 or 3 paralleled for more reliability.

@ScotchTapeLord: The demonstration of "obvious statement is obvious" xD

But yet, useful to fundament that idea

The good thing of limiting current on inductor and not with a resistor, is that I don't reduce efficiency that much. AND it allows me to make the 555 switch at a slower speed, reducing duty cycles and increasing circuit efficiency. And as a side effect, lower switching speed means lower losses on the MOSFET, which means less overall heat. If I increase the charging time from 20 to 25 secs but increase charger's efficiency a 10%, I'm happy.

Thanks a lot to both! Sincerely I don't know what would've happened with this without you guys

ScotchTapeLord

Sat Oct 30 2010, 05:52PM

Registered Member #1875 Joined: Sun Dec 21 2008, 06:36PM
Location:
Posts: 635

Bear in mind that lowering the switching speed will result in higher current. It basically undoes what the larger inductor does.

Raising the inductance value doesn't hurt efficiency at all. In fact, it helps. Efficiency falls exponentially as current rises. Higher voltage across a higher inductance equates to highest efficiency- it's more power across the system for the same given current.

You say lower switching times on the MOSFET will equate to less heat. This is true if the current is the same, but false for when the current rises as a result of slower switching times. Raising the inductance along with lowering the frequency will result in the same current for the lower switching speed, so it would in theory lower the losses in the MOSFET with the lower switching time. Negligibly. The MOSFET's switching losses are very low, even for 100's of kHz. I have only seen special conditions used to mitigate switching losses at 4MHz...

The point of slower speed is to increase the duty cycle... that's what puts power into the inductor.

And the control pin voltage determines at what level the threshold pin triggers, I believe. Putting a capacitor there just makes sure that little interferences don't alter its value. It can generally be omitted and left disconnected... Simulation programs often don't like disconnected pins, though.

Artikbot

Sat Oct 30 2010, 06:13PM

Registered Member #3247 Joined: Mon Sept 27 2010, 09:42AM
Location: Spain
Posts: 137

Aha, good to know, then no need to modify the timer pots

So, letting current drop by using a larger inductor is in fact helping the efficiency? Woah, that's one of the main goals of this circuit, increase the initial 15ish % efficiency of photoflash circuits.
So no need to worry for the MOSFET? Great! I was reading circuits all over the net where MOSFETs were actually heating up a lot, and it was caused by switching losses. I believe they were scrathing the MHz level.

Voltage control pin, disconnected? Is it any bad if I've plugged a 10nF capacitor to it? If it causes more harm that the wounds it heals, I'll remove it and case solved -it's just a cable and two solder spots.

By the way, so having a 97.something % duty cycle is in fact good because current will build up faster in the inductor? I thought that duty cycle is what killed circuit's efficiency o.O

Build log:

-Installed the new uber beefy inductor.

-Installed the voltage control module for the non-inverting input of the opamp.
Rised resistor values liek 40 times to ensure no high current builds up and burns resistors and op-amp. This time I've chosen 85.6 and 1.36k respecively for the divider. last time (my fault) was 2.4ish k and 47Ohm. High current burnt the potentiometer, and very possibly, the op-amp aswell.

-Changed the LED resistor. The previous one was small (they sold me 68Ohm instead of 680Ohm, dang it).

-Installed a filtering capacitor in pin 5 (not sure if it works)

Will keep updating, gonna try the circuit right now!

Edit (21:20, October 30, 2010):

Thanks a lot!

Mister GhostNull, mister ScotchTapeLord, mister Grenadier...

IT WORKED!!!!

The OpAmp isn't burnt!

Here it goes what happened:

1: Powered the circuit on, the led lit. That's good.
2: The inductor started to whine. Good signal, it dodn't with the old one.
3: The mosfet didn't heat up. Good signal aswell, that means current limiting is working.
4: Voltage rised VERY fast, it means it's powerful. In 3 secs a 1mF cap bank was charged to 160V. A-ma-zing. Photoflash ones took about a minute to charge to 150V the same bank.
5: A potentiometer started to make a strange smell in the 4th second more or less.
6: One second later, that same potentiometer spontaneously burnt with a flame of about 3cm high.
7: Then I cut the power

Now I gotta see if it's the potentiometer that burnt, or why did it burn, or if it's just a capillar of solder that was shorting two contacts, or what has burnt.

BUT IT WORKED!!!

I'm SO HAPPY!!

Edit: It's the small pot, the 10k one that burnt. It was fixed at 1.38k, it was the small one of the divider. The big one was a 100k set at 85.6k.

I gotta make a blog, post this and make you guys a monument

Veyr thanks for your help and patience!!

GhostNull

Sat Oct 30 2010, 09:25PM

Registered Member #2648 Joined: Sun Jan 24 2010, 12:45PM
Location: Australia
Posts: 291

In the voltage divider, I think the values should be greater. I think the larger resistor in the voltage divider should be 1meg ohm (fixed). As an added safety measure, you could put another resistor (10k) in series with the smaller pot to prevent from accidentally going too low.

The voltage control pin (CV) can be used to adjust the threshold voltage which is set internally to be 2/3 Vs. Usually this function is not required and the control input is connected to 0V with a 0.01ÂµF capacitor to eliminate electrical noise. It can be left unconnected if noise is not a problem. (source:

) C2 is for this, I don't have a decoupling resistor in that circuit.

But that's great! Finally got the charger working! Now to start working on the other parts.

first previous 11 12 13 14 next last

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