Newtronics head close up

[Gyro Gearloose]

Check this out. Lighting the head from the bottom lets you see some amazing detail. I jammed the head on my old optical microscope (which was filthy and dusty) with lighting from the bottom, and my phone taking pictures through the eyepiece.

The thin light line in the horizontal black bar is the gap of the read/write head. The black bar is the ferrite head itself, embedded in the white ceramic wear face. The two larger gaps at 90 degrees to the thin gap are what's called the "tunnel erase" heads.

A floppy drive erase head doesn't erase the old data ahead of the r/w write head the way an audio cassette does. Writing digital 1s and 0s requires no erasing like analog audio does. You just whack the magnetic domains back and forth. The purpose of the "tunnel erase" is to erase the data the drive *just wrote* on either side of the r/w gap to create blank gaps on both sides.

This leaves a smaller channel of written data so that there's no interference between tracks. As far as I can tell, most floppies work on this principle. Except those "flopticals" but that's another story.

This is the famous alignment that needs to be done on 1541s, when another drive's head starts picking up data from two tracks at the same time. Why other drives don't have this problem I suppose is because the 1541 loves banging the head against the stopper because there's no track 0 sensor.

Floppy drives are "open loop", the motor steps the head and wherever it lands mechanically that's where it reads. There's no fine tuning system that aligns the head precisely like in a hard drive. (Although the first hard drives on personal computers were also open-loop designs.)

[Gyro Gearloose]

It really looks a lot clearer when I look right into the microscope. This is as good as it gets for now. This picture shows the r/w gap more clearly.

[Gyro Gearloose]

Mmmm, here's a nice application note about floppy disk heads
https://datasheet.datasheetarchive.com/ ... X55928.pdf

[Gyro Gearloose]

OK, so looks like that 1541 style of head is called a "straddle erase" head. I don't even know why they bothered calling that whole thing "erase", it's confusing. Creating those guard bands is called "trim erase", and since it has nothing to do with erasing data, it should have been called the "trim" coil, not the "erase" coil. The whole thing about tunnel vs straddle is more about the shape of the head than the actual purpose.

Also what's interesting is that a 48TPI disk with 35 tracks means that less than an inch of magnetic material is actually used on the radius of the disk. I wonder if they couldn't use all the surface because the disk material starts to wobble at the outer edge?

A textbook I found online says that the manufacturing process for these heads was "quite complicated". (Handbook of Micro/Nano Tribology, Second Edition, p 658) Take a look, I'm not cooking that up on my kitchen counter...

Why do I get sucked into these futile pursuits?

[Gyro Gearloose]

The Excelerator + uses a mechanism with a tunnel erase head.


A different approach. I have to clean the case and power the drive up first (another "as is" eBay find, this one looks like it was dropped) but at least the head measures intact.

[Gyro Gearloose]

Here ya go
https://tinyurl.com/stao5ew
If you click "images" you'll get a bit of a taste of what's involved here...

[Gyro Gearloose]

And here's the earliest patent I found for the straddle erase head
https://tinyurl.com/vwtlh2a

Why there are two ways to do this is not clear to me. Maybe one type of head permits much smaller gaps for higher density.
I mean before they abandoned inductive heads for pickup.

[Gyro Gearloose]

I'm slowly getting enough information together to contact manufacturers with more facts.
http://www.durascanheads.com/index.html

[Gyro Gearloose]

One of the facts is that it would cost US $5000 just for the upfront reverse engineering of this head. How much it would cost to manufacture them, and how to get them into the defective drives is another story.

[Gyro Gearloose]

It's even worse than I thought; the entire sled assembly with the metal strap and cable is basically one unit. It needs to be mechanically aligned very accurately so the head picks up enough signal.
I can't see a way to pop out the head "button" (the little metal can with the coil and ferrite assembly) from the sled, or how to get a new one in there.
The only solution would be to rework the existing head as-is.