• Please review our updated Terms and Rules here
  • From now on we will require that a prefix is set for any items in the sales area. We have created regions and locations for this. We also require that you select a delivery option before posting your listing. This will hopefully help us streamline the things that get listed for sales here and help local people better advertise their items, especially for local only sales. New sales rules are also coming, so stay tuned.

Looking for an 8" floppy drive

commodorejohn

Veteran Member
Joined
Jul 6, 2010
Messages
3,302
Location
California, USA
I'm currently in the process of stockpiling bits and pieces for a potential project, and one thing I've been interested in is adding an 8" floppy disk drive to the mix. However, I've never used 8" disks before, so I'm not sure how much variation there was in the hardware. I have a stockpile of about 70 DSDD disks in IBM's 1024-byte/sector format; is this soft-sectored? Will pretty much any soft-sectored double-sided 8" drive work with this, as long as I've got a controller that can read it? Does anyone have suitable equipment to spare?
 
8" Tandons - Not For Sale:

8" Tandons - Not For Sale:

I can't sell you one of my drives but I can recommend that you buy a later model 8" drive without the AC Motors.

I bought two new Tandon TM848s half-height DSDD 8inch drives around 1986 for my MSC-LAT1, and they we a much better design, taking advantage of the motor technology used in 5 1/4inch drives.

I had a pair of old AC motor 8" full height drive and never liked them; I would be disappointed to ever find them in my garage.

Before you buy 8" drives, it would be a good idea to download some 8" drive manuals (PDF) and look over them to see what you'd be willing to buy. My favorite online source for documentation is:
http://maben.homeip.net/static/S100/
 
Last edited:
My opinion is different, having used and owning many 8" drives, including the TM-848s. The 848 has one big flaw--the 24VDC is used only for the drive motor--it's dropped to 12V by a 7812 regulator for everything else. Two failures are very typical--the capacitor on the input to the regulator fails short, with the result that a picofuse on the PCB gets blown--or, the 7812, even with the shunt resistor, is sorely overloaded and sometimes fails.

My preference is for AC-motor drives--absolutely no trouble with them. My day-in, day-out drives are Qumetrack DT/8s and Siemens FDD-200s. Shugart SA-850s are also very nice drives. All are built like battleships. If you're looking for a drive for those IBMs (yes, they're soft-sectored), be careful not to grab a single-sided drive, the most common of which seems to be the Shugart SA-801. A good drive, but with an obvious limitation. Whatever, if you get an AC motor drive, be sure it matches your line power. A 220V 50Hz drive isn't going to work well without substantial changes on a 120V 60Hz circuit.

Japanese 8" drives can be a little strange, particularly ones from NEC, so if you find one, be sure to get a manual for it. Early on, there really wasn't a standard for 8" drive connectors, so a manual is valuable in any case. One notable zinger is the old Calcomp 8" drives with power to the drive being fed through the the same 50-conductor ribbon cable used for signals. Plug it in upside-down for a real surprise. Other older drives, such as PerSci, Innovex, Orbis, etc. have their own peculiarities.

Note that for all of these drives, you'll need a 24V/5V (and possibly -5) power supply. 50-to-34 pin adapters, if you don't want to roll one yourself can be had from John Wilson at DBit.
 
Thanks for the input, both of you. The main concern here is that I don't want one badly enough to spend crazy money on it (some eBay sellers list them for upwards of $200,) and there are actually a couple TM-848s on eBay in the $50-60 price range, albeit in "untested" and less than pristine condition. How reparable are these things, typically?
 
Oh, generally speaking, if the heads are in good condition, they're quite repairable. I believe that Bitsavers has a maintenance manual for them, so they're not mysterious. They were a standard fixture on TRS-80 Model 16s (and their 16-bit cousins).

Are your IBM disks in the System/34-style magazines?
 
Any idea whether this is a one-sided or two-sided drive? The manual I found lists 848-1 and 848-2 variants, but the sticker on this drive says 848-E, which isn't mentioned.
 
Double-sided--look at the bottom left area on the PCB. See that connector with the "409" label on it? Note that there are two black cables coming out from it. A single sided drive would have only one.

Note that yellow electrolytic just below the power connector? That the one to suspect if you get no response. It tends to fail silently.

I have no idea what that green patch wire connected to pin 48 is supposed to do. I suspect that it can be safely removed. Note the mashed jumper pins near the power connector.
 
Last edited:
While I agree with Chuck that the 24/12v thing isn't a great design, I haven't had any related bad experience with the TM848s; I've used (and still use) a number of TM848s, some of which have seen a fair bit of use, and no problems with the 12V supply.

Note that there are a number of different versions of the TM848, some of which are easier to repair than others; the -E is a relatively modern design (double-sided) using a microprocessor and direct drive, unlike earlier versions and most other drives of the day that used belt drive.
 
Any idea whether this is a one-sided or two-sided drive? The manual I found lists 848-1 and 848-2 variants, but the sticker on this drive says 848-E, which isn't mentioned.
It's quite different from the TM848-2... its printed circuit boards are smaller and sparce on the chassis in contrast to the TM848-2. It may be that -E was an economy model made for a big production line with all the excess cost stripped out. Tandy certainly had the clout to get cost reduction deal for their expected volume.
 
It's quite different from the TM848-2... its printed circuit boards are smaller and sparce on the chassis in contrast to the TM848-2. It may be that -E was an economy model made for a big production line with all the excess cost stripped out. Tandy certainly had the clout to get cost reduction deal for their expected volume.

Not to mention Jugi Tandon's penchant to produce the cheapest product possible.
 
One drawback of the AC motor drives was that the motor ran continually.

Did anyone ever come up with a timer off circuit that worked? The snag always seemed to be that the controller assumed the motor was running, and didn't wait for it to come up to speed. Mind you, all my experience has been with the Motorola Exorciser so that may be an oddity.
 
Not to mention Jugi Tandon's penchant to produce the cheapest product possible.
Nothing inherently wrong with that, as long as the quality doesn't suffer. Some of his other products certainly violated that rule, but I wouldn't say that the TM848-E was an 'economy' model with all that that implies; just following the trend of replacing a lot of TTL with a microprocessor (8748 in this case) and a few support chips, as with most electronics lowering the cost while actually improving reliability.
 
Last edited:
One drawback of the AC motor drives was that the motor ran continually.
Not as bad as it sounds; many drives also had a head load solenoid, so although the disk would be spinning in its jacket at least the head only contacted the disk when the drive was selected.

Avoiding the spinning of unselected disks (and saving power) was the main reason for IBM's clever oft-maligned twisted cable hack.
 
If you've got an AC motor, you have exceptionally stable speed, due mostly to the lock-in to line frequency and the sheer mass of the assembly. AFAIK, all AC-motor drives have head-load solenoids. One benefit of the AC motor is that the spindle is moving when you insert a floppy, so it always seats correctly (almost all 5.25" floppy drive have to spin the motor to do this and have a circuit to test when the door is being closed).

Finally, note that the Tandon 848's motor control line is the opposite sense of a 5.25" drive. That is, it's "motor off", not "motor on" when active-low. I can see why this was done--if you're replacing an AC-motor drive, there's no motor control line, so the default should be "run all the time". You can't simply run the motor control off the head-load signal, as a typical head-load time is 50 msec., while the Tandon specs motor-up-to-speed at 700 msec.

There are 5.25" floppies with head-load solenoids, so it isn't a bad idea. I think economics drove the industry in its elimination, rather than good practice.
 
Just as a data point, one of the Tandon TM848-series drives in my TRS-80 Model 12 had the shorted filter capacitor failure. There was no fuse to blow; instead, it just dragged down the 24V supply. It was a black plastic-cased axial leaded capacitor on the input side of the regulator, and replacing it was an easy fix.

A while back, I got a pair of TM-848-series drives mounted in an external drive enclosure. The enclosure is of the same style as the external drive enclosures typical of TRS-80 computers, but it's comically large. The whole thing looks like a clown's giant novelty floppy drive. I've successfully used it with a KryoFlux via the FDADAP adapter. I cleaned up the innards of the unit, replaced the fan filter, unplugged one of the drives (the KryoFlux didn't seem to be happy with two drives on the cable), and brought out a short cable to power the FDADAP from the enclosure's power supply. It all seems to work nicely. The main drawback of the setup is that it's a bit of a pain to open up the enclosure and dismount the drive's PCB to get at the heads for cleaning, which may be a frequent problem when I start trying to image lots of questionable old disks. It may not be the optimal DS 8" drive type for me to use, but it's what I have.

I agree that 8" drive prices on eBay have generally been ridiculously high. What you end up with may be heavily influenced by what eventually gets posted at a reasonable price; if you end up having some options to choose from, I suggest that you consider ease of access to the heads for cleaning.
 
I mounted a white LED inside one of my full-height 8" drives so I could see what I was doing when cleaning it. It also serves as a visible "power on" indicator.
 
Floppy motor speed tolerance is more critical than speed stability

Floppy motor speed tolerance is more critical than speed stability

If you've got an AC motor, you have exceptionally stable speed, due mostly to the lock-in to line frequency and the sheer mass of the assembly.
Stable doesn't really matter as long as the speed is within +/-3% of target. That's the figure that the 8" IBM formats use to assure compatibility between systems under worst case combinations.

Speed variance within this 3% tolerance range is not enough to trip the PLL.

However, I'd bet that if you surveyed people, 90% would believe the PLL tolerance is more critical. The drive speed being within 3% is more critical - beyond that you risk clobbering your tracks and losing data.

Tandon's 848-2 and -1 use the DC motor controls spec at +/- 1.5% as is common among 5 1/4" drives. I prefer the extra precision.
 
Last edited:
How many controllers made since 1985 even use a PLL? Most use a digital counter-type data separator, such as the WD9216, No VFO needed. (I've heard the 9216 referred to as a "jog-locked-loop") The original 5150 FDC board used a PLL, with IBM hybrid circuits, but by the time of the 5160, that was history.

As far as speed variations, it isn't long-term speed changes that'll getcha, it's ISV. DC motors with their lower mass tend to fall victim to this more than the AC motor drives.

If you're doing a data separator in an MCU, it's usually quite satisfactory for most applications to simply establish a bandpass-type discriminator as in "anything between this and this is such and such a length pulse", perhaps adding a long-term correction factor.
 
...digital counter-type...
Exactly the right way to do it.

If you're doing a data separator in an MCU...bandpass-type discriminator...
No. You'll limit yourself to a specific data encoding format... once you're in the firmware, there is no need to limit yourself like circuitry. The answer is in the first quote above.

...How many controllers made since 1985 even use a PLL...
If I didn't address it first someone would have posted a PLL argument. Just saving time.
 
Back
Top