Electric Fence Zapper

[madlabs]

Oz, think a core from a CRT flyback would work? The older style, two "C" shaped, air (paper) gapped? Cuz' I gots one of them.

Jonathan

[madlabs]

So, I wound the secondary, using AWG 24 magnet wire. I couldn't locate my little coil winder, so I made this sophisticated setup:



The secondary coil bobbin was made from some karropak gasket material rolled into a tube with a couple of insulating washers epoxied on the end. All went well until about the 7th layer, when a combination of cat, coffee and sneezing caused a little uneveness in the coil. Still, it came out OK. By my calculations I should have close to the 600 turn target.



A question about the primary, I assume we are talking about two side-by-side coils, not layered. Also I see 10 turns and 24 turns on top, but in the schematic see two 24 turn coils. I'm going to go ahead and wind 24 turns and leave room for the rest, just so I can test the coil ratio and see if I'm in the ballpark for 600V.

Jonathan

[madlabs]

Wound a primary with 10 turns and 20 turns of 18 awg wire. I couldn't fit the full 24 turns, so hopefully that will be ok. Must have got the secondary turns in the ball park because 97 mVAC in yields 920mVAC out. I prolly have the stuff laying around to lash up the input side and see if I get around 600VDC out.

[oztules]

Ok, CRT transformer is fine. If the starts and finishes are correct, the transistor will see positive feedback in the base, and it WILL oscillate.
The turns... I seem to have created confusion with my slack winding practices.

It has two primaries. Both the same turns number. I use 18-24 depends on what wire I have on hand... why?

Simple, I like to wind the first layer so that it runs from one end of the former to the other, close wound... just to keep the layer neat and clean...... but it would take 40 odd turns to achieve a full layer..... so I wind two in hand, and so wind both primaries together. This gives us close coupling, a single flat layer, and I only need to count once, and the starts and finishes are obvious.... so thats why I do it....... I'm a lazy slob.

The secondary is as many turns as you can fit on, but less than 600, or the reactance may well current limit the output, and is a waste of wire to go more.

David.
The free wheel diodes need to be as big as you can get in there.
EDIT: That should be 1000v not 100v.........6A10 are 1000v diodes
I use 6a10 100v 10A diodes X2. They are only 10A but surge is about 800A combined... and on a larger system, should use more or bigger.

Primary of transformer is 24 and 24  ( or 18 and 18 etc) as on the circuit, not as in the inset (24, and 10)

R4,R5,R6 1/2W
R1 2k2 R2 640   1w  for lower power systems,  with bigger transistors for more drive 2W.

Both resistors are very flexible. Want faster charging for larger UF caps, then , go to lower values and bigger transistor..........to lower the power in the circuit and keep the charge rate at 1 second or more for low UF systems, then higher values or smaller transistors.

Toroid..... any toroid from the outputs of the computer power supplies would do. They are materials suited to this purpose, and are freely available.
Size... don't really know, about 32mm outer, 12mm width,and 20mm hole... just a guess from here. They are out in the shed, but just grab em from computer supplies. They are not critical in any way.
The etd49/39 (maybe etd29 may be a bit small for 30uf systems and above), can be any thing you can scavenge from tv sets power pwm cores, or computer monitors etc, or even as Madlabs has done, the old eht cores from older tellies.
 There is nothing critical in any of this except to get the starts and finishes of the little oscillator correct, if so,it should run.... amplifier+ pos feedback=oscillator

The diacs.. no markings, but from memory, about 30 odd volts step off/ avalanche. I bought 100 for $2 inc shipping from your country.
Step off voltage is of no concern if yours are different, as we just need to change the sense circuit (the 3 1m resistors) to suit the the diac... they are the divider after all.

D1 is just a diode to protect from reverse leads.. a 1A diode of anything over 30v will do.

R4-8 and C4 are your sort of timing components. They give you a delayed reaction of what is actually happening with the capacitor V, as well as determining when the step off of the diac occurs.  C4 provides a load for the divider to work into, and stops spikes triggering prematurely, R4,5,6 and 7 provide the divider, and R8 some fine control.


....................oztules

[David HK]

Oztules,

Many thanks for the update. All points noted. I will assimilate everything into the drawing, including supplementary notes if I can get them in.

I have ordered an ETD 49 from local HK company.

I may (will definitely) come back with an updated drawing. Check back in a day or so to see if I have more questions.

Thank you so far.

David in HK

[David HK]

p.m. Saturday 6th April 2013       

Hello Oztules,

Drawing updated as follows:-

All resistors rated for 1 watt, or, ½ Watt.

All diode type numbers are 6A106A       1,200v x 100A

C2 and C3 specified as 2uf 400v AC (will change spec if wrong)

Method of winding coils 1 and 2 primaries, and 3 as secondary, noted.

All triacs are:-   BTA26-600B   600v  x 25A

A question:-

In Page 1 of this thread, there are some colour photographs at Reply #10. On the 12 volt PCB I can see some light blue items – what are they? Do they feature in the circuit now being drawn?

There are also two brown “Caps” are these the 2uf @ ~ 400v items?

Capacitor C4 @ 0.7uf. Can you offer a voltage for this, or, a product number if known. It looks like a Mylar of some value.

That’s about it for today.

Dave

[oztules]

http://www.ebay.com.au/itm/100-BT137-800E-8A-800V-Philips-Sensitive-Gate-TRIACs-/200530873298?pt=LH_DefaultDomain_0&hash=item2eb0924bd2
I use these, 18 bucks a hundred, 800v 8A for the "scr" ( triacs are cheaper...), but technically it needs only a scr not a bidirectional switch.

The 1200v 150A diodes are way over kill...... 2 or 3 parallel 6a10 would be enough for most folks for the free wheel, they have high surge, and the surge is very very short.

 The rectifier diodes need only be 1n4007 eg http://www.ebay.com.au/itm/100x-IN4007-DO-41-Rectifier-Diode-1A-1000V-1000-Voltage-/120712606542?pt=AU_B_I_Electrical_Test_Equipment&hash=item1c1b08334e  3 dollars per hundred and free shipping.
and use them for the input diode

The blue things were an attempt at stopping spikes>600v MOV (varistor).... sometimes I use them, mostly don't. These things have been remarkably stable out in the field.

The caps for the storage (50uf 450vac) are motor run AC caps  We can get away with the 450vac, as the peak for this is over 600vdc (450X1.414), and we use DC.

The 2uf ones should be the same, but the poly ones @ 400v seem to do alright,as the super low impedance of the primary of the transformer keeps the voltage clamped down low. Never actually measured it with the scope.

The C4 is any cap running over 100v. It really only sees less than 40v before the trigger goes, but a bit of clearance is good.



................oztules

[David HK]

Thank you Oztules.

Diodes and Triacs changed as per your comments. All available locally. Just one or two items (Cap C4) to check out for local supply and I think when that is updated the schematic will be complete. Will probably post that next week.

Next step see how big the ETD 49 is, and then see what I can do for PCB design.

So far so good.

Dave

[David HK]

Oztules,

At the age of 65 I am still learning electronics. Some things I easily understand and others I have to learn as I go.

ETD 49 transformer. I have studied this in the Internet and the bobbin width is 32.7 mm wide. This means that a winding of 2 in hand by 1mm diameter wires will only achieve some 16 (32mm) turns rather than the desired 24 turns (48mm). About a 30% shortfall.

It then occurred to me that the objective of 24 turns could be achieved by finding an old  (or newish) laminated steel transformer with a 50 mm wide bobbin to accommodate the required turns. Thereafter the bobbin can be wound for whatever one can get on it but not exceeding size constraint of the laminates.

No doubt there is a performance difference between ferrite and laminated steel, but in this case would it really matter. Your advice would be welcome.

Dave

[David HK]

Advancing the above logic on a cool Sunday morning and with the help of Excel and a small laminated type transformer I have some new arithmetic.

Rectangular bobbin = (L x B x H) 31.5 mm x 20 mm x 23. Total circumference (2 x 31.5) + (2 x 20) = 103 mm. 

The ETD 49 bobbin is 16.7 mm dia and 32.7 mm wide. At 24 turns of 1 mm wire (one in hand) the length of wire  = 1,259 mm.

12 turns of 1-in hand (1 mm dia) on the rectangular bobbin = 1,236 mm (12 x 103 mm) which is not far off that for the ETD bobbin. 12 turns @ 1 mm is a width of 12 mm. Since the rectangular bobbin is 23 mm wide (or as near as damn it) it will accept 2-in-hand and leave a shortfall of 1 mm.

The only problem is working out how many turns of 0.5 mm wire can be placed over the two primaries.

What a way to spend Sunday morning.

Dave

[oztules]

Golly David,
Your over thinking this. I just get a microwave tranny, and use the primary coil off that for winding the primary of mine, and the secondary wire from the micro for winding the secondary of mine. 1 microwave transformer will wind lots of these little oscillator transformers.

However, you have goaded me into actually measuring one of my formers, and they are in fact 33mm long (internal) by 19mm diameter.
There is a LOT of fexibility here... and it is obvious that my eyeballing the wire and guessing it was about 1mm, was not as close as I had suspected. It must have been somewhere nearer  <.9mm. ( after you unwind it, stretch it to straighten it, I don't know what it started out at, but usually 18 (36) fits the former)

Wind whatever you can get hold of for the layer, and it will do fine.
The ferrite is much better than the iron, as the frequency will vary from low to very high as the charge voltage changes. the iron may interfere with the efficiency of the transformer as it tries to get to the >15khz part of the cycle, and it is in this region it struggles the most.

So, just get the etd49 former, and wind 2 in hand of whatever wire you can get hold of that is close to the .9-1mm size ( I know I have used <.8mm before to just as good a result. The small turns number, and modest current would allow for .5mm wire, and still be fine, but like I said, I just want to fill the first layer from end to end, and be tidy. 18-24 turns gives you some idea of what wire I must have been using out of the old microwaves, and whatever else I can scavenge on this island. ( startermotor relays, normal relays, wire out of the fans in microwave ovens..... you name it I will have tried it)

Your tables don't take into account the stretch that can be applied when you wind it on a 4 ton lathe or big handled coil winder, practical will tell you what really fits.

Remember when making the big transformer, each layer needs to be insulated from the last with transformer paper or equivalent.. to stop arcing.... 10000+ volts like to find a way where it can, but separate layers seems to work fine, then epoxy the transformer when finished.... don't need moisture in it. Over here night dew is almost like a mm of rain.... ocean next door.


...................oztules

[David HK]

Thanks Oztules,

I have finished, for the time being, the schematic and will put it up for your eagle eye sometime next week.

Your last note was very useful as a guideline. I have a home made coil winder with counter left over from my Hugh Piggot turbine coils so I should be able to adapt that for winding.

Point taken on the ETD 49. Mine should arrive this week so I will let you know how that goes.

I have also picked up an old 220/110 transformer from the local scrapyard so that should keep me occupied for a day or so. It weighs 10.93 lbs.

If you could cast your mind to it, is there a recommended way of testing any part of the control circuit up to the main transformer?

Dave

[oztules]

Hmmm.... testing....
Tricky..... Without the transformer, you need to be very careful and quick.

Suppose it is all ok, what can we expect to happen?

1. If we connect the 12v we will hear a low pitched squeal rapidly growing to a high pitch squeal... but by the time it gets to high pitch, we may have well over 600v in the big cap.... so...

2. We connect up the 12v by touching the terminal, if we hear the oscillator start and climb, we quickly pull off the 12v wire. Now measure the voltage stored in the capacitor, if you held for 1/2 second, it will probably be over 400volts.. so the oscillator is working..... so how do we test the next bit... the trigger... ( use a resistor to bleed the cap back down to zero)

Well without a load of some type ( low value resistor perhaps) we don't dare try.

With a low impedance load of some type across the output where the transformer would be, we can tickle the input and hear the squeal run up with each tickle.... with a meter across the capacitor, It should get to about 500-600 and then discharge into the resistor.

We can do this by tickling the oscillator, and watching the meter.. if we get up to near 600v, and nothing has happened, then there is a fair chance it is not going to trigger.... almost impossible to see why, as it is so simple... but it gets scary driving it much higher, as the cap is nearing it's limit..... if you want you can put 2 caps in series, and now have 1300v head room to test with, but the oscillator is capable of over 1000v in all likelihood, and can drive the diodes out of business too. The whole thing relies on the trigger using up the voltage, if it is not connected to a load ( transformer primary) then all bets are off unless you have a load.

Things that have gone wrong:

If the freewheel diodes die, they ( mostly) die short, and the oscillator just does a low squeal, and does not rise, it is clamped by the freewheel diode/s  being shorted.

If the rectifier diode/s fail, the thing may develop low voltage 100-200v, and the diodes may fail short, but no always.. fist time I have seen this, usually they always fail short.

If the oscillator fails, then dead.
..
If the main cap fails, then odd behavior, these wound caps can fail all strange. Sometimes ok up to a certain voltage, sometimes no capacitance, sometimes short... So the oscillator will scream to match the fault.... ie rise normally for a bit then stop at that point... cap charges half way, then shorts etc etc.

Never seen a fault in the trigger circuits.

It the triacs have failed, then drop off the transformer, and that removes the load it would have presented to the cap, and so tickle up the voltage, if all well, triac/s failed... usually only 1... closest to the cap.


Hope that helps a bit.


................oztules

[David HK]

Oztules or any one else with coil winding experience.

I now have the ETD 49 plastic coil former. After doing the mathematics I shall be using 0.8 mm wire for the primaries which will permit 20 turns of 2-in-hand across the width of the coil. This is a halfway house between 18 and 24 tuns and 1 mm dia wire which only permits 16 turns across the bobbin width.

Having done that I intend to wrap that coil(primaries) with thin transformer paper.

The next bit forms my question. This involves the 1 mm dia wire for the secondary. After I have swept a row of 1 mm across the width of the coil do I need to add another transformer paper liner over the wire before commencing with the second row and so on.

Or, do I (say for example), wind a 100 turns and then cover that with transformer paper and each successive 100 turns thereafter until I have completed ~ 500 turns or thereabouts?

Advice would be welcome.

Dave in HK

[oztules]

David, there won't be room for transformer paper with 1mm wire... secondary should be spider web wire.. .5m etc

I don't use transformer paper on this tranny, as there are plenty of layers separating the start layer from the end layer. Keep the start and end fly leads apart at least 1/4" (Don't cross them in the transformer).

The loaded voltage will always be less then 600v, so it is no real concern... even with recycled wire from the microwave transformers.

Now for some Weird stuff:
For anyone else watching this, I have used a transformer secondary from a micro, only to find that it would not work. There was something strange with the wire... I never found out what. I rewound the same transformer 3 times with the same batch of wire from the same micro, and none of them worked...odd, and can't explain why, next batch of wire from another transformer worked fine as always....... never seen it again, but there must be some voodoo wire out there somewhere.

Just keep it tidy, without letting the next layer slip too deep at the former ends into the last layers, and all should be fine. Never seen one of these transformers fail from anything now I think about it, and there are quite a few out in the elements for the last 4 years or so.

The main transformer is a different story.  High voltage, and very high currents exist here, and each layer needs to be carefully wound, with transformer paper exceeding the layer width by about 5mm each side. This will make it  very hard for the arcing which would occur instantly without this protection.

Commercial formers for this voltage have lots of individual 1/8th" wide sections all side by side, and you fill each "slot" before going to the next, and this separates them for factory winding.

The paper works a treat like this, and does not arc.... but won't last 5 mins without this treatment.


..................oztules