These are some cool disc drives but we have some issues (Atari 810)

I was an Atari authorized warranty technician for 10 years (1985-1995) in Federal Way WA. The local school districts would send me all their Atari disk drives for service every summer, hundreds of drives, plus many hundreds off the street. Some things came to mind as watched this video. The early 810 drive (MPI mechs) came with no separate FDC daughter board, but it was found that the built in 1771 data separator was defective and so that outboard data separator daughter board was created. Spindle bearings can get grabby and this can cause issues. The PLL can't stay locked to the data clock when the disk speed momentarily dives. The disk head needs to be thoroughly cleaned. You can't just look at the head and say, "oh it looks clean" as the interfering filth is microscopic. As you saw, the analog board has no shielding and so strong magnetic fields from monitor yokes did cause problems, even though the head signals are differential. Your problem drive probably needs a new pressure (compliance) pad, after another thorough head cleaning with a cotton tipped swab (not the cleaning disk). The felt pads wear and stop pushing the disk evenly down onto the R/W gap. Some pads used an adhesive like rubber cement and could easily be damaged by bad disks or improper insertion / removal. Alcohol can soften the adhesive and the pad becomes misaligned or gets rolled up. Use a mirror to inspect the pad. Nearly all the pads on single sided drives of this vintage are going to be defective. One other thing comes to mind. The connectors and their pins often have fractured solder joints. Check the joints under a microscope or touch them all up to correct the fractured solder joints. On that old software for the 400/800, you'll need to load the "translator disk" to get the old software to run properly on the XL / XE machines. The MPI and 1050 drives had no adjustments for head azimuth, but the TM100 mechs did have an azimuth adjustment that needed to be performed using a scope with the alignment disk.

I gotta say.... those are some good looking drives!

12:21 , as someone that worked in an injection moulding factory, yeah moulds are incredibly expensive, upwards of $100,000 so that's a major cost saver

Awesome that you're using the FujiNet, I'm one of the devs on the project. If you want to boot your disk from FujiNet with Basic disabled, hold the option key down when booting until you see the Blue Screen boot, otherwise Basic is left enabled. Same way you do when booting Gorf, as it's an Atari thing, not a FujiNet thing :D If you just "tap" the Option key, basic is still enabled, which can be a problem for some disks/applications. Happy FujiNet'ing!

Atari was selling so many VCS game consoles, that they had an unbelievable quantity of 6507 and 6532s in their parts store, and were getting them for vastly discounted prices. It was heavily incentivized for engineers to build peripherals using those two chips.

The top and bottom of the 810 are indeed the same, and was one of the nice touches that Kevin McKinsey made to make this thing easier to produce.

In early 1982, Atari made a deal with Tandon. Atari would buy TM100-1's in large quantity to fulfill the 810 orders, and Tandon would design the replacement disk drive for the 1000 series (which became the XL systems), the 1050 (which used a Tandon TM50-1 mech).

To answer your formatting question: The FORMAT command itself in the drive firmware is expected to format the disk, then verify it, while returning the number of bad sectors it finds. If this number is 0 (and the command acknowledges as COMPLETE), then the disk utility package will write a filesystem to it. Otherwise, it will return the # of bad sectors, the disk utility package will re-issue the FORMAT command again. If this fails again, then an ERROR 173 (bad disk format) will be issued.

Adrian needs the well known Translator disk for 800 to 800XL compatibility.

Makes me miss my 1200XL. I spent so many hours punching in code from old Computer Games magazines and saving it on cassettes. At least until I finally got a disk drive for it. Seeing these things just lights up all kinds of old memories.

The early 810s like the one in the middle, use the mech from Magnetic Peripherals Inc. (MPI aka Control Data). The later 810s use a single sided mech from Tandon (basically a TM100-1 without its analog board) Original 810s had no data seperator (completely disregarding WD's technical note on the 1771 stating that data seperators were strongly encouraged), which lead to companies like Percom selling Data seperator boards. Later 810s had the Grass Valley modification, which included an improved analog board, an improved motor control board, and an improved ROM which changed the sector interleave to improve performance. The Atari disk drives also have a particular quirk that they are tach'd at 288 RPM, rather than 300 RPM. Also, thanks for the shout-out on the FujiNet :) -Thom from FujiNet

The addon board was a fix released fairly early in the 810's life - the drive will work if you plug the 1771 straight into the socket but it'll be more prone to read errors. Also they used CPUs rejected from the 2600; the console ran the 6507 slightly overclocked, and any that failed to cope eventually ended up being repurposed. Also as a heads up, GORF and EP-CYG-4 are incompatible with XL/XE machines.

I'm a Commodore guy but it's nice to see some different stuff. By the way these long videos are great and i love them !

I'm glad to see the Atari get some love. It just occurred to me that I didn't get any ads. Did you forget to monetize? Also, we were supposed to have a test today, teacher.

That "mod board" is the external data separator in the FDC socket. That was an "optional" thing with FDC chips that turned out to not really be so optional after all. And the CPU ("maybe that's the ROM") is the same 6507 as used in the Atari 2600. The 6532 RIOT is another of the three main chips from the Atari 2600. And yes, both chips are relatively common as long as there are junk 2600s to steal them from. The interesting thing about the 1771 chip is that apparently it's an inverted-logic bus version. So all the sector data on the floppy is inverted! Way back in '08 or so I was ripping all sorts of disks with a Catweasel board (PCI version in a Linux system) and found that you have to invert all the sector data when reading Atari disks. There's also the issue of the slightly slower motor speed which gave them one extra sector per track.

Fun fact, the creator of the SIO interface used it as a basis for creating USB.

you are magician :-) ... my dad entire life repaired tape/cassete recorders, lots of mechanics, belts, heads, etc .. this was very close :-)

The 810 can be upgraded to true double density with a Happy board, which replaced the 6507 with a 6502. Faster SIO too.

Ah! (paused at 23:08 so if you cover this, sorry)... the CPU in the first drive is a Rockwell C010745, a 6507 made by Synertek. It looks to be a MOS 6502 type CPU with only 13 address lines and some other pins dropped to fit a 28 pin DIP. So it only works to 8Kb (the Hitachi chip is RAM), and the GI 9316B chip is mask ROM.

As someone who grew up with a 130XE in the house, happy to see the Atari 8-bit line getting more love. The 1050 was my daily drive(r) (huhuhuh) and the one we got for our machine kept having read issues and went in for warranty service multiple times at the local dealer. That wasn't common though I don't think. I just acquired a 130XE that's in mint condition, save a non-working Start key but i picked up a FujiNet and can't wait to mess around with it once I fix the keyboard.