Speaker magnetizer repair

[Rhodesplyr]

A local speaker company has asked for help in getting their speaker magnetizer working again. It reportedly stopped working a couple of years ago after a lightning storm, though that could simply be coincidence. It was custom built for them years ago, so all they have is this hand-drawn schematic.



It's obviously a brute-force three-phase rectifier with a timing circuit.

For switching, it uses an octal plug-in relay inside the box to control the SCRs.

Based on what I can make out, the timing circuit's 20uF and 0.1uF capacitors are charged to a positive voltage when power is turned on to the unit via a large breaker box. Then, I'm guessing that when the switch (a big red momentary button on the front) is pushed, the discharge of the 0.1uF capacitor across the relay's coil is enough to trip it, at which point, the relay stays on until the 20uF capacitor discharges through the relay's 10k coil.

The other pole of the relay apparently switches to allow gate current to flow in the SCRs, turning them on. I've been told by a friend that the devices at the input are "Thyrectors," though I'd never heard of these before. He suggested that they were there to prevent spikes from getting fed back into the line.

If I help them fix it, I get free use of it for the foreseeable future. I'm thinking that my strategy should be to check the timing circuit first and them proceed to the heavier-duty components. The 20uF electrolytic capacitor in it looks pretty old, so it might have shorted.

Any advice/feedback is appreciated. With high-current power circuits like this, I understand how they work, but not necessarily the finer points of whether it's a good or a marginal design.

[Enzo]

First, what does "stopped working" mean? Is it just dead? Does the relay pulse but no magnetics? What are we trying to fix? Hard to give advice if we don;t know that.

Marginal/schmarginal, if the design has worked for them a long time, why try to re-engineer it now?

Isolate the problem, whatever it is. I'd disconnect the three phase power and just connect 120v to the control circuit. Does that function? Does it pulse the relay? Now that the three phase is unpowered, stick an ohm meter on the relay contacts to see if they close. You can pull the relay and connect 120v (I assume) to its coil and see that all contacts work in both states. That would allow measurements without dealing with a short pulse.

The button appears to have make and break contacts, so... do they? Certainly a bad button switch would be no mystery.

Dried out electrolytic? Sure, why not. But shorted? I bet you'd know if that 20uf cap were shorted, after all it is charging whenever 120v is on, and that means it if were shorted, you'd be lowing fuses or breakers.

On to the high power stuff. First, has this thing been blowing breakers? If so, then something is shorted, but if the breakers are happy, then your thyrectors must not be shorted. You can google thyrector for more information on them. So then, what is the high powered circuit. It is a coil/electromagnet and some diodes. They may be huge diodes, but they are diodes nonetheless. So thest them as diodes. They conduct one way and not the other. You can test the SCRs by applying a small voltage to the gates. While not powered of course.

Is the magnet coil open? Is it shorted to frame? With its leads disconnected, I bet RG's transformer short detector from his geofex site ought to work as a test.

And looking at it now, I imagine I could connect a low voltage supply in place of the 240 threephase, and maybe a light bulb in place of the magnet coil and so check the functionality of it all.

[Rhodesplyr]

My apologies for my post rubbing you the wrong way, Enzo. I will refrain from posting things like it in the future. Actually, I thought of deleting it before you replied, but this forum doesn't seem to allow complete deletion of posts. Perhaps it should.

I guess we can get back to debugging Behringer powered mixers.

[Enzo]

Hey wait. I am not sure how I gave that impression. I don't feel rubbed at all. I have no beef with you in the slightest.


I just asked for clarification as to what you needed, and pointed out my thoughts on design. And then I made as best suggestion as I could as to a systematic approach to it, whatever might be wrong. It was my intent to give an honest assistance.

I was not sure you even knew what it was retired for, as in they might have said , "It don't work." and left it in a corner.

As to not posting this stuff, well, as esoteric as it is and though it may pertain to very few of us, it still is part of music electronics. But even if it were a garage door opener, it is still electronic troubleshooting, and that is the main thrust of all my posting here. I don;t try to teach electronics, I try to teach troubleshooting. And to have some unusual or off the wall thing to fix, serves to show the less experienced that troubleshooting is a universal skill, not specific knowledge of specific amps.


Believe me, sitting here in my shop, I;'d be a lot more interested to see this thing walk through my door than the 1000th Fender Hot Rod Deluxe needing tube sockets resoldered.


SO I think I am sorry I rubbed you the wrong way, because I sure meant no hostility.

[jrfrond]

David???

[J M Fahey]

Well, believe it or not, there's somebody here who is very interested in helping you repair this machine.
Who might I be?

To begin with, I already have one very similar, 3 phase, with basically the same schematic (there's not that many ways to skin a cat).

And I'm building a larger one, so I'll ask you some info on that one , curious as always and it's not an electric device that is much seen, nor sold over the counter.

1) all what Enzo said is correct, of course.

2) start with the main and most dificult component, the magnetizing yoke:

a) measure its resistance, we want to check that there's continuity.
Post its value, to check it's reasonable.
I'd expect something between 5 and 10 ohms DCR
If not, check connections, pigtail wire soldering, is it hardwired or it has some plug?
The thick wire of the coil itself will rarely have a problem, unless it was used continuously for too much time, it overheated and burnt insulation ... but that would be a gross abuse, not consistent with a storm.

b) some yoke pictures and a simple diagram with dimensions (internal/external diameter, coil/core height, etc.) will certainly help.

3) check that D3 is still a diode, not open or shorted.

If shorted, it will bypass the magnetizer coil and blow fuses big way.
It may also kill some of the bridge diodes or thyristors, if fuses don't blow fast enough.

If open, it will not catch the EMF pulse and the resulting unquenched arc will *destroy* whatever it catches.

4) the trigger/timing circuit is crude but effective.
And quite foolproof.

a) For one, it's antibouncing: the first pulse triggers it for good and eventual later sparks (if any) have no effect.

b) even if the operator stays with the button pushed, once the 20uF capacitor is discharged, the trigger relay opens, and can't be retriggered until the operator stops pushing the button and lets the .1uF capacitor recharge.
So if the 20uF is open, the trigger relay won't be on for long enough (a few seconds) but a couple milli seconds.
And if the .1uF or R2 is open, the relay will not trigger.

Measure the 20uF cap voltage, it should be around 140/150V.
I'd replace it with a fresh one, by the way.

As Enzo said, you can test the trigger with the 3 phase power disconnected, using just the 120V one.

Of course, the relay may be bad: burnt coil, pitted contacts, dirty/worn socket contacts, etc.

5) Diodes "D1" and "D2" are testable, of course.

6) to functionally check Thyristors/SCR I'd first disconnect fully the 240V 3 phase input, disconnect (one at a time) the anode and the end of R1 conected to the trigger relay and connect a small 12VAC transformer , one wire to cathode, the other to anode through a 12V car type lamp.
The lamp should not light, the untriggered SCR is an open circuit.
Then get a fresh 9V battery, connect negative to cathode, and "brush" the positive end to the free end of R1, the lamp should blink but only when the 9V reach the gate (through the resistor), and self turn off when not, because it's receiving AC.
When the lamp turns on, it will do so with less light and some "vibration"because it's receiving 1/2 wave AC .

You have your hands full for now. test and post results, want to be sure before hooking the 3 phase 240F line.

I *hope* there's some beefy fuses there (I expect 25 to 50A each) and that they plug into a *good* industrial power line.

I have my own "big" magnetizer at a friend's factory, 2 or 3 hours away, because my shop power line can't feed it, go figure.

[Rhodesplyr]

Hi Juan,

I went back for some preliminary diagnostics. The control/timing circuit still works, but components are very old, so it should probably be rebuilt. The 20uF capacitor is a Sprague Atom in a yellow cardboard tube. Early 60s??

The three 60A fuses that supply the high-current circuits are all missing, for good reason, since I measure dead shorts all around the D1 diodes and SCRs, including among all three supply lines. I need to disconnect things for more investigation, but the control circuit is run from a completely different breaker and is still hot. The breaker wasn't marked, so they were going to have to wait till after business hours, turn off all the computers, and flip breakers to find the right one.

Also, in contrast to the schematic, there are actually two SCRs in parallel for each D1 diode, probably due to current limits of individual SCRs at the time. There are no spike-protecting Thyrectors.

Steve Conner suggested to me that there are now high-current Diode/SCR modules that could be used to rebuild it. These turn out to be cheaper than buying six 50A SCRs. It can also be rebuilt with higher PIV semiconductors.

I'll post more after they figure out how to cut the 120VAC supply to the control circuit.

[J M Fahey]

Hi Juan,

I went back for some preliminary diagnostics. The control/timing circuit still works, but components are very old, so it should probably be rebuilt. The 20uF capacitor is a Sprague Atom in a yellow cardboard tube. Early 60s??

Well, that's something.

The three 60A fuses that supply the high-current circuits are all missing, for good reason, since I measure dead shorts all around the D1 diodes and SCRs, including among all three supply lines.

Ouch !!!
Perhaps it was permanently connected to the power line (since it's "self switching") when a lightning hit.

Also, in contrast to the schematic, there are actually two SCRs in parallel for each D1 diode, probably due to current limits of individual SCRs at the time. There are no spike-protecting Thyrectors.

OK.

Steve Conner suggested to me that there are now high-current Diode/SCR modules that could be used to rebuild it. These turn out to be cheaper than buying six 50A SCRs. It can also be rebuilt with higher PIV semiconductors.

Did he suggest any model number?
I'm also interested.

Although anyway high current classic SCR aren't that expensive anyway.
I'm delaying the construction of a larger magnetizer, because I can't get the truckload of monster caps to make it a Capacitive Discharge one ... and a brute force one like yours is no big (technological) deal ... but requires a monster current line.
What size magnets do they charge?
Can you post some yoke picture?
Is it a "cup" type (where you fit only the speaker's "ass" in), an oldstyle "C" type?
Mine is a heavy iron rectangle, weighing almost 1000 Lbs, with 2 6" polepieces, coils wrapped around them and you put the already built speaker in the middle.
Can magnetize up to G12H/V30 size magnets mounted on 15" speakers, but now must build a larger one for EVM15/E130 type speakers ... as you can imagine I'm delaying construction of such a monster until I can find a more modern and practical solution, but if not, a monster it will be.

Unfortunately if I want to make JBL/EV/RCF class speakers, I need the same size magnetizer as them, even if I will make only some 10 speakers a month.
Can't beat Physics

So any info on actual machines used in real Factories is very useful.

Plus there's little or no data about this in the Net.

Are you allowed to name your Friend's Factory or Speaker Brand?

And yes, you will be able to custom make some hairy speakers.

[Rhodesplyr]

This is a C-frame style magnetizer with coils above and below the speaker. I've never actually seen it in use.

There is a paper online, "Methods of Magnetizing Permanent Magnet Speakers," in which is it suggested that while simple rectified DC magnetizers are sufficient for charging Alnico magnets, they may not produce the necessary peak magnetic pulses to charge ferrite magnets to saturation. Capacitive-discharge may be necessary for ferrites.

oersted.com/magnetizing.PDF

I'm mainly interested in recharging Alnico magnets in vintage recones.

[J M Fahey]

Ok, thanks.

It must be a *real* old machine, maybe from the 50's or early 60's.
In fact it may have used mercury rectifiers or something, and SCRs were "upgrades".

Mine, in fact, which I built in the 70's, didn't use SCR but regular (huge) bolt-on diodes and switched on the AC side (so arc self extinguishes whem crossing zero) with a three-phase "contactor", which is a monster relay used to start subway electric motors and such.

If your friends are not a factory but reconers, they may have bought it in an auction, from a defunct speaker factory.

And yes, it will be more than enough for Alnicos, which are "soft".

I read the Oersted paper, it's quite accurate, although they tell only "just enough", they want to sell you their services or equipment .

PK, test diodes and SCR one by one, you'll soon have it working.

Do your friends have another machine?

I mean, since it sounds like this one has been out of order from some time.

Good luck.

[Rhodesplyr]

Juan,

I only know what I've been told about this magnetizer, but here's what I have. It was custom-built in the 1970s for this speaker company (which now focuses much more on pro audio installations) by a retired Electrical Engineer who was fairly old at that time. They used it for years at their original location and reinstalled it when they moved to a new building in the 1990s. It worked there until a few years ago when it may have been damaged in a storm, but no one is 100% sure about that.

The timing circuit is crudely tacked together on perfboard with parts that look like they may have been old when they were installed--from someone's surplus parts collection. The high-current diodes and SCRs are stud-mounted on aluminum busbars, and all those connections are screwed together with large connectors. The power circuits are mounted in a metal box on the wall and connect to the free-standing C-frame via armored conduit.

They are now interested in getting it working again because they got a bunch of nice cast aluminum frames and Alnico slugs out of storage and are thinking of building and selling some of their own new drivers for the first time in years. Coincidentally, I've gotten into some reconing at the same time and would like to be able to recharge Alnico magnets that may have been subjected to abuse.

That's all I know. Photos will follow.

[J M Fahey]

Well, you know a lot, and have cleared many doubts.

Specially because I recognize the design as a very robust but old one, your description of the maker explains the apparent date incongruency. Cool.

And your friends' idea is cool, Alnico is fashionable again, they will turn their dusty stock into fresh cash, what's wrong about that?

I'd do the same.

In fact I have tons of old speaker frames, left here when musicians replaced theirs with mine.

They are taking a lot of space, so I guess I'll spend a couple days reconing all and selling them in our version of EBay or mounted in some simple generic cabinet.
Back then it was not worth reconing them singly, but doing them as a batch will save a lot of time.

I will certainly have more use for any cash they bring than for them as they are now.
I don't need that many doorstops

[Rhodesplyr]

Juan, I got some pictures of the magnetizer for you. I also determined, with power off, that all D1 diodes are shorted. The SCRs seem, on the basis of a quick check, to be OK, but I won't say I'm sure yet. Getting the diodes out of the circuit involves a good bit of disassembly. D3 measures good.

[J M Fahey]

Hi Rhodesplayer, THANKS a LOT !!!
Actual magnetizer pictures are as rare a hen's teeth.

Good to know the SCR look good.
You can test them without unmounting them, just disconnect anode and gate connections (I guess you'll need a BIG iron there) use a small 12V + lamp in series and trigger gate with an external 9V battery through 470 ohms or use original 100 ohm resistors connecting them to anode.
The SCR should switch on with a few mA into the gate, check datasheet .

As a side comment: I LOVE the way it was built, very Industrial/Pro.

Wiring hidden inside flexible ,metal tubing, main circuit insulated inside a good Electrical box, the works.
It must have cost quite a few $$$$$

If the coils are fine (should be, look very well built), all that remains is the SCR.

Some questions (sorry :

What's the distance from the center of a coil to the column?
Can you fit a 15" frame there?
Looks like you barely could; a 12" should have no problem.

What's the external diameter and height of each coil?

The DC resistance?

I can reasonably estimate nš of turns and wire diameter from that; of course if some bare end is visible somewhere and you can actually measure diameter, much better.

Although sometimes these big coils are wound with doubled or tripled thinner wire, just to make it bearable.

What's the section of the column iron?

As a rule of thumb, you can easily magnetize the same section of Ceramic magnet, so if you have, say, 120 x 120 mm, it's good for an up to 155 mm diameter magnet (Celestion V30/G12H etc.) , same as mine.

For Alnicos, it's overkill, also for the drivers or Bullet tweeters I saw in the first picture.

So, in a nutshell, it's worth repairing it.

Good luck.