A 360K drive has a wider read head. In rare cases in which the track is damaged in the part that is read by a 80-track head, but is still readable in the part that is skipped by an 80-track head, reading the disk with a 40-track head might yield better results. Other than that, I think you are fine reading 360K floppies using a HD drive.

Thanks. This makes me curious: I wonder if it would be possible to temporarily offset an 80-track head in such a way that it would attempt to read problematic 40-track disks from a slightly different spot than it normally would... I can't be the first person to think of this, so I'm guessing that there's a clear "yes", "no", or "only if you're extremely competent at this" answer 😁

I've dumped quite a few floppies using Greaseweazle, here's some thoughts on the matter.

1) It's super rare that a DD disk would read in DD floppy drive but not HD one. It can happen but mostly with usual HW floppy controllers. You will be dumping raw magnetic flux in multi-pass mode and then using software to extract sector data from it.
2) HD drives have better heads and electronics, the resulting stream is way less noisy than when read with DD drive. I can actually tell by now what kind of drive dumped a DD disk just by looking at the track data in HxC.
3) PC HD drives spin at 360 rpm, not 300, and since the signal level is proportional to medium speed it's actually easier to read weak flux transitions, and let the software deal with the bit timing calculations afterwards.
4) The usual issue is dirt on the floppy surface, not weak media. Once you learn how to view the raw data in HxC you can actually pinpoint the dirty region and maybe try to clean it further with q-tip and some water or IPA. Just don't rub too hard or you will start damaging the surface.

I use both DD and HD drives (my HD drive for GW is capable of switching speeds as well) and frankly a properly used HD drive will give you cleaner result. So unless you somehow insist on having that DD noise in the dump, because it's more authentic, stick to HD drive with double-stepping. What you do care about is that the drive is clean, the heads above all since they can damage the media when dirty, and the mechanical operation is smooth and problem-free. You also want the drive to be rather quiet (no weird noised from the spindle or the clamping arm) so that you can quickly tell if something is happening with the floppy that should be stopped right away. Do note that old floppies tend to warp a bit when sitting unused in one particular position, this will result in specific noise from the drive when it's spinning - but that's not a problem.

Deunan wrote on 2023-07-18, 22:51:

I've dumped quite a few floppies using Greaseweazle, here's some thoughts on the matter. […]
Show full quote

I've dumped quite a few floppies using Greaseweazle, here's some thoughts on the matter.

1) It's super rare that a DD disk would read in DD floppy drive but not HD one. It can happen but mostly with usual HW floppy controllers. You will be dumping raw magnetic flux in multi-pass mode and then using software to extract sector data from it.
2) HD drives have better heads and electronics, the resulting stream is way less noisy than when read with DD drive. I can actually tell by now what kind of drive dumped a DD disk just by looking at the track data in HxC.
3) PC HD drives spin at 360 rpm, not 300, and since the signal level is proportional to medium speed it's actually easier to read weak flux transitions, and let the software deal with the bit timing calculations afterwards.
4) The usual issue is dirt on the floppy surface, not weak media. Once you learn how to view the raw data in HxC you can actually pinpoint the dirty region and maybe try to clean it further with q-tip and some water or IPA. Just don't rub too hard or you will start damaging the surface.

I use both DD and HD drives (my HD drive for GW is capable of switching speeds as well) and frankly a properly used HD drive will give you cleaner result. So unless you somehow insist on having that DD noise in the dump, because it's more authentic, stick to HD drive with double-stepping. What you do care about is that the drive is clean, the heads above all since they can damage the media when dirty, and the mechanical operation is smooth and problem-free. You also want the drive to be rather quiet (no weird noised from the spindle or the clamping arm) so that you can quickly tell if something is happening with the floppy that should be stopped right away. Do note that old floppies tend to warp a bit when sitting unused in one particular position, this will result in specific noise from the drive when it's spinning - but that's not a problem.

+1 to all this

Deunan wrote on 2023-07-18, 22:51:

I've dumped quite a few floppies using Greaseweazle, here's some thoughts on the matter.

I can absolutely corroborate your results.

I just finished dumping around 200 or so 5.25" floppies for archival with the Greaseweazle, and my HD drive gave far, far better results than my DD drive. Where the DD drive faltered, the HD drive gave clean reads (as long as there was nothing physically wrong with the disk). And we're talking about very similar 5.25" drives (both very modern, of a relatively late manufacture date. In fact, both drives are identical, except for the read head).

As I understand it, the only use case for a 5.25" 360k drive is where you specifically need to read these disks in old systems that can't use HD drives, so you write these disks with a bona fide DD drive, because a DD drive can't read a double-stepped 360kb-formatted disk...

Someone will correct me, but this does not apply to 3.5" drives, because the size of the head doesn't play a factor in the extra capacity of a 1.44mb floppy, but rather, it has to do more with the rotational speed (i.e. there is no double stepping on 3.5" drives).

And on this topic, the 8-bit guy claimed in one of his recent videos that the flux of HD 3.5" disks formatted as 720kb disks is not as strong as real 720kb disks, but I think he's doing something wrong... Maybe someone can elaborate on that.

jklaiho wrote on 2023-07-18, 21:01:

Thanks. This makes me curious: I wonder if it would be possible to temporarily offset an 80-track head in such a way that it would attempt to read problematic 40-track disks from a slightly different spot than it normally would... I can't be the first person to think of this, so I'm guessing that there's a clear "yes", "no", or "only if you're extremely competent at this" answer 😁

The idea works perfectly. You can also try to offset the head to read 80-track floppies that habe been written by slightly misadjusted drives, or which have local bumps in the media surface. The one isuue with this approach, which might shift the answer from "yes" to "only if" is that there is no physical "temporarily". To offset the heads, you need to intentionally musadjust the drive, and when you are done, you need to readjust the drive properly, which might be difficult or delicate, depending on the adjustment mechanism used by the drive.

mockingbird wrote on 2023-07-19, 05:51:

And on this topic, the 8-bit guy claimed in one of his recent videos that the flux of HD 3.5" disks formatted as 720kb disks is not as strong as real 720kb disks, but I think he's doing something wrong... Maybe someone can elaborate on that.

I share the experience that using 1.44MB media as 720KB disks produces subpar results. On undisputable fact is that 3.5" HD media use a different magnetic material than 3.5" DD disks, very much like there are ferric oxide (Fe2O3) type 1 cassette tapes and chrome dioxide (CrO2) type 2 cassette tapes. These materials have different magnetic properties. The primary one is that HD media can form smaller stable magnetic domains, enabling reliable HD recording.

There are other relevant differences, too

• You need a stronger magnetic field to reliably write to HD media than to DD media. DD drives are not specified to provide enough field to write HD media.
• The analog properties of the signal on DD and HD media are slightly different, as DD media act kind of like a low-pass filter. This seems to be compensated for in some drives at least by providing different kind of filtering (like pre-emphasis / de-emphadis on tape deks).
• The last bullet is in addition to the (digital) write precompensation performed by the floppy controller, which is 125us on certain bit patterns for both HD and DD media. The patterns requiring precompensation are specified in a sequence of bits, though, so if the controller runs at a different bit rate than the medium is intended for, it might miscategorize which patterns need precompensation by 125ns and write a suboptimal signal.

The situation for 3.5" drives is further complicated by the fact, that AT compatible 3.5" drives have two inputs that can select "HD mode" or "DD mode". Obviously, there is the "HD hole" in HD media, again not unlike an extra hole in type 2 compact cassettes, but as 5.25" drives don't have a mechanical indication, and the introduction of HD in the IBM AT started with 5.25 drives, there also is a density select signal on the floppy connector, which was required to switch the field strength of 5.25" HD drives.

The HD pin on the interface is to be driven low by the floppy interface of the AT (and all later compatible PC-type computers) if a DD signal is read or written, and it is high while an HD signal is read or written. I'm specifically not talking about IBM PS/2 computers in this paragraph - their floppy interface is slightly different. When the hole and the signal pin match, everything is fine, as long as the jole indicates the actual medium type, i.e. you didn't drill a hole into DD media or taped over the hole of an HD disk. The combination of "HD media inserted" and "computer requests DD" usually works good enough to successfully format a 1.44MB floppy to 720K, but you better don't rely on it. The last combination: "Controller indicates HD, drive sees DD medium" is the most funny one, because that mode switches some "3-mode drives" into a third mode besides IBM DD and PS/2-like HD, which is used by some non-IBM Japanese computers. This third mode is very similar to the IBM 5.25" HD mode by running the drive at 360rpm instead of 300rpm and thus recording only 1.2MB instead of 1.44MB at 500kbps. I experienced firsthand that connecting a HD drive to a non-IBM DD-only system (that was not driving the density select line low) and using it with DD media resulted in a drive that could read DD disks written on other systems, and other systems could read data written by that drive. But the drive could not read the data written by that drive itself. This symptom disappeared as soon as I grounded the density select signal, pointing to the fact that mix-and-match of HD and DD configurations can cause interesting issues.

As for the claim by the 8-bit guy that HD media provide a weaker signal than DD media: The requirement for a higher field for writing to HD media seems to indicate that the whole magnetic stuff is stronger on HD media, a higher coercitivity (required field for writing) doesn't always imply a higher remanence (signal strength while reading), although a correlation is likely. On the other hand, if a DD drive writes to to HD media (using insufficient write current, because DD drives generally don't write hard enough for HD media), a weak read signal is expected.

Adrian from Adrian's Digital Basement did a rather detailed video on this subject and did really well explaining a lot of various things that I hadn't even realized. His video was more with building a dos machine and not using a Kryo to greaseweazle board but the drive information and explanations he goes over is still packed with good information. https://www.youtube.com/watch?v=fRZVlsxSDw0

And for the record, I used a Sony HD 3.5 drive and my greaseweazel v4 to write a MAC 800k boot disk and it worked just fine. I am going to try hooking my greaseweasel up to a teac fd-505 drive and see how it handles it.

Veeb0rg wrote on 2023-07-19, 10:09:

And for the record, I used a Sony HD 3.5 drive and my greaseweazel v4 to write a MAC 800k boot disk and it worked just fine.

As long as you use DD media, this is expected to work with any kind of DD or HD drive. The more delicate question is whether you can write a MAC 800k image to an HD medium. As long as you don't want to investigate that as research topic, and you still have some DD media at hand, just use DD media for 720K-880K images (250kBit/s), and HD media for 1.2MB-2.0MB (500kBit/s).

mkarcher wrote on 2023-07-19, 07:28:

On undisputable fact is that 3.5" HD media use a different magnetic material than 3.5" DD disks, very much like there are ferric oxide (Fe2O3) type 1 cassette tapes and chrome dioxide (CrO2) type 2 cassette tapes. These materials have different magnetic properties. The primary one is that HD media can form smaller stable magnetic domains, enabling reliable HD recording.

I see. Is this also the case for 5.25" diskettes?

Fascinating info. Thanks

mockingbird wrote on 2023-07-20, 02:59:

mkarcher wrote on 2023-07-19, 07:28:

On undisputable fact is that 3.5" HD media use a different magnetic material than 3.5" DD disks, very much like there are ferric oxide (Fe2O3) type 1 cassette tapes and chrome dioxide (CrO2) type 2 cassette tapes. These materials have different magnetic properties. The primary one is that HD media can form smaller stable magnetic domains, enabling reliable HD recording.

I see. Is this also the case for 5.25" diskettes?

Well actually, I'm sure for 5.25" disks the material change between DD and HD was indeed ferric oxide to chrome dioxide, just as for audio tapes and most likely that also holds for 8" disks. I'm unsure whether 3.5" used classic Fe2O3 material, or whether 3.5" DD material already was more advanced than the 5.25" material. The tracks on 3.5" disks are shorter, so you need higher quality material to record at the same bit rate. Nevertheless, it's still true that 3.5" HD has different magnetic properties than 3.5" DD, like the transition from Fe2O3 to CrO2, as claimed in my post.