Transformer Core Alloy Question

Thread Starter

graybeard

Joined Apr 10, 2012
31
I repair a lot of vintage Smiths tachometers for owners of classic British and Volvo cars of the 60s. The Gen 1 tachometers use a pulse transformer to couple the ignition signal to the tachometer circuit inside of the case. There is a single-turn loop of the ignition wire screwed to the back of the tachometer and a multi-turn winding on the inside of the case. There is a U-shaped clip inside of the internal windings and a mating clip inside of the external loop. Those clips act as the core of the transformer. Without both clips, the ignition pulse will not couple into the tachometer. The frequencies of these pulses are audio frequencies.

I have attached photos of the clip for your reference.

1646246375719.jpeg
1646246408012.jpeg

Unfortunately, many of the external clips are lost during the car restoration process. Nisonger Instruments had an NOS supply of these clips, but it was apparently wiped out by a recent flood.

I have been trying to find a steel or iron alloy that I can use to reproduce these clips, but nothing I have tried will work. I have tried pretty much everything I can find in my scrap metal bins and have also taken apart a small power transformer to use a piece of core laminate.

I thought that there may be a coating on the metal that prevents good magnetic coupling, so I used a Dremel tool with a cutoff wheel to remove any coatings on the metal bits, but that didn’t help either.

If there are any magnetic core experts here who can help me source some sheet metal made from an alloy that will work, I would really appreciate the help.
 
Last edited:

MisterBill2

Joined Jan 23, 2018
11,888
What I see in the second photo is an assembly that holds the wire in a very specific position relative to that clip. And that clip is an interesting arrangement that is thicker than most sheet metal. It seems that the plastic block, if that is what it is, has a great deal to do with holding the wires in exactly the correct position.
What I see in the first photo looks like an effort of soldering with an eight-hundred watt soldering iron.

It seems that the arrangement is half of a transformer, and that the dimensions of the portions coupling to the internal portion must hold things in very close proximity, probably touching, so as to transfer enough magnetic flux. So what is required is a very precisely made part, if accuracy is required.

Is it possible to see the inside portion of one of those tachometers so that all may understand what is involved.
It may be that there is not even any electronics inside, but rather a scheme to use the string of magnetic pulses to deflect the meter pointer. Like some of the early SUN brand tachometers, the early Corvette tachometer, and the Stewart Warner hand-held tachometer.
 

MisterBill2

Joined Jan 23, 2018
11,888
It may be that what will work out best is to replace the whole assembly with a ferrite core transformer. Some timing lights used a clamp-on ferrite core pickup for the trigger pulses. That sort of core should be available as a current product, and probably work as well or better. How much electronics is in those tachometers? Probably not very much, A replacement transformer glued to the back of the tachometer housing could be an adequate replacement. for the missing pulse transformer half.
The issue with the part is far more the magnetic coupling by close contact, rather than the particular alloy. Any good transformer steel of the right thickness should work. The "big deal" is in the magnetic flux coupling.
If you can create a design for a replacement part, by copying the original, you can name your own price.
 

Ian0

Joined Aug 7, 2020
5,147
I used to rebuild these using an LM2917. The original circuit is indeed a current transformer, so transformer steel should work. I think that the original circuit was a two-transistor monostable, with germanium devices.
My circuit used a capacitive pickup off the HT lead between coil and distributor (a piece of insulated wire wound around the ht lead). It’s a long time ago, but presumably I developed the capacitive circuit because I couldn’t get the inductive circuit to work because the clips were missing.
 

Thread Starter

graybeard

Joined Apr 10, 2012
31
Guys, thanks for the replies.

Relative to the plastic block, all it does is prevent the wire insulation from chafing on the transformer core clip. When I am testing the tachometers on the bench, I don't bother with the plastic block, I just put a 1 turn loop around the clip. I made a 1A amplifier that simulates the coil current, driven by a function generator. The position of the loop around the core clip does not matter with an OEM clip. FYI, I have reproduced the plastic block using 3D printing.

The first photo is of a tachometer that a customer sent to me. It was only intended to show how the loop was attached to the back of the tachometer. You wouldn't believe the condition of some of the tachometers and wiring that are sent to me to work on.

The inside portion of the pulse transformer looks like this:

1646330888907.jpeg
I put a spare internal winding and clip on top so you can see them better. The internal core clip is identical to the eternal core clip and can be harvested from donor tachometers. But these tachometers are getting too expensive to buy for cannibalization unless they are in really bad shape. The clip is about 1/2" x 1/2" x 3/8".

Here is a photo of a rough clip I made from the transformer core laminate. It also shows how my amplifier wire loop goes around the clip. It works very well with the EOEM clip, but the signal will not couple through the transformer with my clip:
1646331315591.jpeg
I took any coating off the metal with a cutoff wheel and you can see it is making good contact with the mating core clip.

Most of my customers are sticklers for restoration and do not want to replace the innards with modern circuits, nor will they want to modify the case to replace the pulse transformer. I agree a more modern design would work a lot better. But they want the original unit as it was shipped from the factory.

For those of you interested in the circuit, here are the schematics I drew for the typical Gen 1 Smiths tachometer:
1646331579927.png
Most of these tachometers actually work pretty well with a new timing cap and sometimes with other components replaced as needed.

I would have expected that the transformer core laminate would have worked, but, for some reason, it doesn't. If anyone has any creative ideas, I am all ears...

Thanks again for the replies.
 

MisterBill2

Joined Jan 23, 2018
11,888
You mean the ones that were mechanical, driven by a cable like a speedometer uses? Or not so early? One of my friends has a 1957 Corvette and it has a mechanical cable tach.
Yes, I do mean the mechanical tach!

Certainly the LM2917 will provide an excellent tachometer, especially if the scale is linear on the meter. I built several of them at one job and they were all very satisfactory
 

Ian0

Joined Aug 7, 2020
5,147
Most of my customers are sticklers for restoration and do not want to replace the innards with modern circuits, nor will they want to modify the case to replace the pulse transformer. I agree a more modern design would work a lot better. But they want the original unit as it was shipped from the factory.
I know that sort of customer - who can see inside the tacho? I'd rather know what speed the engine was running!
I think your problem is the gap. You have a made a transformer with a gap in the core. I think you need a very good fit between the two parts of the clip for it to work.
 

michael8

Joined Jan 11, 2015
287
What does C1 do in the circuit? It's accross the 2ndary so perhaps it resonates at some frequency? Does your replacemet
transformer/core hit the same frequency? I'd put a scope across C1 and see what's there...
 

MisterBill2

Joined Jan 23, 2018
11,888
After seeing the scheme, it looks like an epoxied torroid could replace that transformer scheme with a much more durable setup. And given that the main ignition wire from the coil comes right into the area near the radio it seems like an invitation to radio noise. But perhaps a little ignition noise is OK for folks driving those models.
Certainy grinding off the oxide coating does not improve the magnetic coupling at all, as the surface is no longer even.
 

Thread Starter

graybeard

Joined Apr 10, 2012
31
What does C1 do in the circuit? It's accross the 2ndary so perhaps it resonates at some frequency? Does your replacemet
transformer/core hit the same frequency? I'd put a scope across C1 and see what's there...
I am not sure what C1 does. I assume it is there for a little filtering since ignition signals are so noisy. I am not seeing anything noticeable across C1 with my homemade core clip in place, but I see a nice pulse across it with each ignition pulse with the OEM clip.
 

Thread Starter

graybeard

Joined Apr 10, 2012
31
After seeing the scheme, it looks like an epoxied torroid could replace that transformer scheme with a much more durable setup. And given that the main ignition wire from the coil comes right into the area near the radio it seems like an invitation to radio noise. But perhaps a little ignition noise is OK for folks driving those models.
Certainy grinding off the oxide coating does not improve the magnetic coupling at all, as the surface is no longer even.
As I said, my customers are funny about modifications. The best solution will be for me to reproduce the core clips. I did beat up the clip I fabricated. I guess I should try to fabricate another one from the transformer core material, taking a lot more care this time.
 

MisterBill2

Joined Jan 23, 2018
11,888
I am not sure what C1 does. I assume it is there for a little filtering since ignition signals are so noisy. I am not seeing anything noticeable across C1 with my homemade core clip in place, but I see a nice pulse across it with each ignition pulse with the OEM clip.
Two things about the clip you made. First, cod working alters the magnetic properties of steel quite a bit, and second, grinding off the surface made the contact much less so another clip, just bent around a mandre of the proper diameter sothat the egs ie fat against the legs of the other half should serve much better. AND the thickness probably matters a bit as well. Thicker is better in this instance. The soe reason for having the cross part of the clip flat is for that bot to hod it in place.
 

Thread Starter

graybeard

Joined Apr 10, 2012
31
MisterBill2, You called it perfectly. I pulled a much flatter piece of the transformer core out and was careful fabricating the next clip. I didn't try removing any coating and was careful to retain the original dimensions. It worked perfectly with the tachometer on my bench.

The thickness of the metal from the transformer is about the same, so that wasn't an issue either.

I didn't expect that it would be so sensitive to the flatness of the sides. Now I need to finish deconstructing the transformer and then find access to an NC milling machine so I can make the reproductions look nice. Should be easy to find here in Silicon Valley.

Thanks so much to all of you who looked at this and helped me get to a working solution, especially MisterBill2!

Mark
 

MisterBill2

Joined Jan 23, 2018
11,888
You can probably mill out a tool and a die to form the clips just right each time. Not sure where milling will be useful except for putting a nice edge on the transformer strip.
Did you square the corners on that new clip, or round the top ? Rounding the top may be sighty more efficient but much harder to put that bot hole into, unless you use a diamond bit in the mill and a very sow vertical feed. That would be impressive but not needed, although you can list it as your "improved original" version.

I earned quite a bit about transformers and what does not work so very well, which is how you got the good advice. Surface contact matters.
 

Thread Starter

graybeard

Joined Apr 10, 2012
31
The metal bit I used today already had a hole the right size in it from the transformer, so I just marked on the metal where to bend it. I bent it with 90 degree angles, just like the OEM unit and then cut the lengths of the two legs with my Dremel cutoff wheel. I then had to narrow the legs to fit into the holes in the back of the tachometer. I used the Dremel to do that too, so it looks very rough from the side. But it worked as a proof of concept.

The whole transformer was covered by a coat of shellac. The first metal bit I used was from the outside layer of metal, so I had to remove the shellac. Today, I used a bit from the middle f the core.

The volume of these clips won't be super high, so it may not make sense to make a die to stamp them. I suspect an NC program to bang out however many the metal from this donor transformer will make will be a cheaper solution. But I am sure the shop I use will tell me which is a better way to go. I do want it to look like the original and like a professionally made part.

I am grateful that you shared your experience.
 
Top