Finally decided to build a transformer. I found an old transformer to disassemble for the E-I laminations and the bobbin, but it was so heavily varnished that there was no way to get the laminations out without boiling. Once I removed the laminations this way, they seem to be stripped of everything. There is absolutely no insulation.
Can I just put some transformer varnish on each of the laminations before stacking and expect reasonable insulation?
Thanks,
Mike
Can I just put some transformer varnish on each of the laminations before stacking and expect reasonable insulation?
Thanks,
Mike
You can try, in the worse escenario your transformer will only have "more" loses, but you can compensate with the windings and a aproximate saturation calculation, professionaly made transformers do soak the entire transformer in a varnish cube and do extract the air with a vaccum pump to ideally impregnate every space of air, but is also quality, dependant, at the end the transformers may have imperfections in your case lots of metal laminations may touch and that will certainly only reduce its performance, but it wont cause it to be totally unusable, there is people that uses Western Electric century old cores to do output transformers for vaccum tube amplifiers (an application of a transformer where the electric parameters are totaly dependant with the final product quality), and they have lots of acceptance, in the audio niche market, where is supposed to be expected the last quality and perfection ever.
Nowadays, laminations are almost always insulated by oxidation or a similar surface treatment, or sometimes by a mineral coating like Carlite.
This means that it cannot be removed or dissolved, it can just be mechanically or chemically stripped
Very old techniques were based on paper or varnish, but that's normally not the case anymore, except in special fabrications
This means that it cannot be removed or dissolved, it can just be mechanically or chemically stripped
Very old techniques were based on paper or varnish, but that's normally not the case anymore, except in special fabrications
Factory lams are often treated with steam. A thin almost-invisible layer of oxide (rust) is all it takes. It does not have to be "100% glass", it only has to be much lower conductance than Silicon Iron over most of the contact.
Since steam-treat needs precise control, I'd think shellac or varnish is the way to go DIY.
Since steam-treat needs precise control, I'd think shellac or varnish is the way to go DIY.
What I meant is that boiling water alone is unlikely to have removed the (invisible) insulation, unless you added some chemical
Agree with Elvee: boiling, solvents, even boiling in lye won´t remove sh*t (pardon the French).
While any acid treatment, including Phosphoric *may* so in principle avoid it.
Unless you sanded or blasted your laminations, you´ll be fine.
And surface treatment is not *that* invisible either; greenish / bluish / greyish surface *is* the actual insulation treatment.
Microns thick? ... thick enough,.
"If it weren´t there" surface would be "plain iron coloured" , like, say, what you find on a nail, or when you sand/grind iron sheet "down to the bare metal" and prof of that is that it quickly starts rusting, red rust that is.
While surface treatments *are* some kind of rust but far stronger and stable ones.
And rust spots here and there don´pt hurt.
While any acid treatment, including Phosphoric *may* so in principle avoid it.
Unless you sanded or blasted your laminations, you´ll be fine.
And surface treatment is not *that* invisible either; greenish / bluish / greyish surface *is* the actual insulation treatment.
Microns thick? ... thick enough,.
"If it weren´t there" surface would be "plain iron coloured" , like, say, what you find on a nail, or when you sand/grind iron sheet "down to the bare metal" and prof of that is that it quickly starts rusting, red rust that is.
While surface treatments *are* some kind of rust but far stronger and stable ones.
And rust spots here and there don´pt hurt.
Simple solution… put a stack of E (or I, but E is easier) between two sheets of plywood, in a vise. With a sheet of copper (just a bit will do) on each side, soldered to wires.
Then while it is compressed in a vise, hard enough to slightly indent the wood, measure the resistance.
If it is above 100 kΩ, you're golden. If below, then consider VERY light application of varnish to the lams.
If your varnish is acetone soluble (most are), then diluting 1 unit of varnish in 20 units of acetone is good. Dip the lams, don't spray. They'll dry off in less than a minute. It goes pretty quickly!
Old, old timer … the Goat … having done this.
Just saying,
GoatGuy ✓
Then while it is compressed in a vise, hard enough to slightly indent the wood, measure the resistance.
If it is above 100 kΩ, you're golden. If below, then consider VERY light application of varnish to the lams.
If your varnish is acetone soluble (most are), then diluting 1 unit of varnish in 20 units of acetone is good. Dip the lams, don't spray. They'll dry off in less than a minute. It goes pretty quickly!
Old, old timer … the Goat … having done this.
Just saying,
GoatGuy ✓
This is very confusing to me, as I have measured many transformers and the lams were "shorted" together. I have also seen some modern transformers with all of the lams welded together. I don't understand.... I have seen that the through bolts wee insulated though, maybe causing a turn with the end bells if not insulated?
Last edited:
Without delving into too many details, the "shorting" that actually counts is one that encompasses the field lines into a loop.
This means that a bolt tightening the laminations together does not qualify, since it encompasses ~nothing on its own (just its diameter) (but the pressure it exerts may promote shorts elsewhere), and a side-welding of the E to the I only closes ~<1/4 of the path.
The welding cord itself has a bit of area, and it wastes some power, but the welding also affects the permeability, meaning a smaller part of the flux will flow into it.
The result is slightly higher losses, and a slightly smaller effective area, but a much more economical construction compared to truly interleaved E-I's
This means that a bolt tightening the laminations together does not qualify, since it encompasses ~nothing on its own (just its diameter) (but the pressure it exerts may promote shorts elsewhere), and a side-welding of the E to the I only closes ~<1/4 of the path.
The welding cord itself has a bit of area, and it wastes some power, but the welding also affects the permeability, meaning a smaller part of the flux will flow into it.
The result is slightly higher losses, and a slightly smaller effective area, but a much more economical construction compared to truly interleaved E-I's
The insulation does not have to be very good. On thin lams it can be under one Ohm and still be much much higher than the resistance of a solid block of iron.
Yes, a weld-line is common where a transformer may be hit or needs stronger mounts. This is normally at the outside of the core where flux is weak, and a line rather than a big area.
Also a big hit of Silicon reduces the conductivity of iron a lot. Not so we don't need lams, but enough that a few line-welds don't matter much. (The weld metal may be super-Silicon iron.)
Eddy current in the iron was a big deal in 1910 but is much less of a problem with improved irons. Also less a problem for hand-size cores appliance cores than for garage-size distribution cores.
Yes, a weld-line is common where a transformer may be hit or needs stronger mounts. This is normally at the outside of the core where flux is weak, and a line rather than a big area.
Also a big hit of Silicon reduces the conductivity of iron a lot. Not so we don't need lams, but enough that a few line-welds don't matter much. (The weld metal may be super-Silicon iron.)
Eddy current in the iron was a big deal in 1910 but is much less of a problem with improved irons. Also less a problem for hand-size cores appliance cores than for garage-size distribution cores.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.