Here's a funny one. I have an original Apple II drive I am repairing. The first thing I did was run a drive speed test both on Apple-Cillin and Locksmith, it said the drive speed was going from one extreme to the other but the pitch of the motor wasn't changing, although it did seem fast. I traced this down to a problem on the motor board with a faulty 2SC2785 transistor and the Sony CX065B chip, I replaced these and the I could hear the motor was moving a lot slower. Apple Cillin and Locksmith still said the drive was erratic in speed though which I thought was odd. I built an optical tachometer and put it on the drive spindle, the drive was very slow 220rpm and I tuned it up so it was running at 300rpm and I monitored it for a few minutes it was perfect! Apple-Cillin and Locksmith still reported it being erratic. I absolutely know this if FALSE. I can only assume it's writing some data to the drive and timing it to read it to get the drive speed, sometimes it may be missing hence the erratic drive speed. Does anyone have any ideas. Is this a stepper motor alignment issue perhaps?
Nice job troubleshooting the speed issue.
ALPS drives have been know to fail on the speed control circuit, both on Apple IIs and on Commodore 1541 drives (which share the same basic ALPS mechanism)
If both usitlities are reporting that the drive is errratic then perhaps it is. A small amount of speed jitter is normal but if it is extreme it could be because of a slipping belt, stretched belt, poor motor lubricati0n or other mechanical issue.
Your eyes can deceive you, and those strobes are not 100% accurate anyway.
The most inportant thing is that the drive read and write diskettes properly.
So the next test to do is to try and read a known-good diskette written from ANOTHER of your drives and see if this drive reads it properly.
And then write a diskette with this drive and then attempt a read test with a known-good disk drive.
If you're successful then ejnoy using your newly repaired drive. If not, then more trubleshooting needs to be done.
Thanks Baldrick, you're right about the ALPS drives and the controller being the same as the 1541, that's where I got the schematic :-) The spindle is 100% accurate, I wasn't using a strobe on the spindle, I was actually using a little trigger on the spindle which cuts an opto-sensor and measures down to uS accuracy, the drive did run at 300rpm +/- a few rpm for a few minutes, it's certainly not a drive speed issue, I did think about the slipping belt. I will do more research though. Thanks Youngie....
It's interesting that I've seen something similar with the software reading, one thing to make sure is that all the connections are clean because there may be something that's not getting back to the motherboard which exhibits the odd speed readings.On the belt, if it wasn't used for a while it could be warped, while this doesn't mean it will slip it does add to the slipability. One thing I do and it's up to you if you want to try something similar is I use a CRC belt conditioner. It's made for drive belts, but I'm expecting the rubber is close enough. Maybe someone can set me straight on that. In addition to that after conditioning I set the belt on a special Solo cup that's ben cut to allow it to "flair out", I put the belt on the cup so the open end is down and then place that on something round and wider to allow the cup to reshape the belt. After that I mount the belt inside out and rotate it around the drive path a few times over a few days before flipping it rightside out then remounting. May be overkill but I've had drive stop exhibiting the odd speeds after doing this along with the other cleaning stuff.
Interesting method described above by jeff d for using a plastic cup as a mandrel. What about boiling the drive belt in water to redistribute any plasticizers?
It's important to realize how these software utilities showing drive speed work. The disk II interface receives no speed signal from the drive. Nor does the drive even have a speed sensor. What the software does is to simply continuously read the data of a fixed track. It's looking for a specific byte pattern, probably something like the start of sector 0 marker. And it measures the time how long it takes for this pattern to repeat, in order to derive the drive's speed.
So, as Baldrick already said, you need a properly formatted, known good disk, otherwise these utilities won't work. And, of course, the read head, the analog circuit, the connection cable and the disk II interface must be working fine. If those were bad, and the read bit stream was therefore unreliable, then the software will incorrectly derive speed values which are all over the place.
If you made yourself a technical device to monitor the spindle's speed with µs accuracy, and you see the rpm is correct and stable, then that is good enough anyway. You would need to look into other reasons, like making sure the disk is good, checking the analog circuitry etc.
Using the software utilities to check rpm is only a final step for fine tuning, once everything else is working.
For anyone interested in this, and it may be of use to someone, even though my drive was spinning at spot-on 300rpm the disk speed utilities reported it as erratic, this was traced down to a broken wrre (green wire) actually going to the drive head. Once I fixed this the utilities reported the drive speed correctly and it all worked fine ;-)
We are always interested in the end result solution. And a broken wire to the read write head would most certainly affect the results of the read write action of the drive, especially as the head slides back and forth on its sled.
Quite often the original poster solves his problem, then we never hear from him again and we never really know the result of the troubleshooting. So thank you for that.
Actually that's not true. Apex II writes its own pattern on one of the tracks and reads it back. This is why all the data on the diskette used during the calibration gets destroyed. Adrian has a pretty good video on that:
Also smartphone strobe tachometers are pretty accurate when it comes to speeds of about 300 RPM. I've used RPM Meter on my Apple II floppy after calibrating it with Apex II and it's exactly 300 RPM. Of course it's a bit harder to calibrate in a dark room with just the phone strobing and giving you seizures, so it's more useful for quickly measuring the RPMs afterwards.
How accurate is RPM Meter? Well, I set it to 1000 RPM and used a photo-diode right agaist my phone's LED and the frequency counter on my oscilloscope:
16.577 x 60 = 994.62 RPM, so let's just say ± half a percent.
Yeah, every speed utility softwre I have ever used warns you to use a "scratch disk" which will be overwritten during the speed measurement process.
The softwares write a pulse to the disk and then read it back.
The way that speed tests are normally coded they essentially zero out a whole track, usually track zero, then they write a few FFs in one spot. Then after that they read, and time how long it takes before they read the non-zero. They do that a few times and then average the time, then compute RPMs from that. The zeroing of the track is why they recommend using a scratch disk. A really sophisticated speed test might run the same test on a few tracks, like say, 16 and 34, and see if they come out the same as zero (they should of course). But from observing drives running a number of different speed test software over the years, I never saw one that did that. Ones that were on diagnostics disks often used some other track than zero, because they knew that the disk image as shipped didn't use that track and it was safe for them to overwrite.
I know you're sharp, but did you miss that LED bulbs will strobe at the correct frequency (anywhere in the world) wihtout having to rely on a programmer generating an accurate strobe? Don't need to fight any strobe lighting, the LED lighting just works!
No LED bulb in the world would strobe at 5 Hz (which is what is needed to measure 300 RPM), unless it was specifically designed for some insane rave parties.
I think he's referring to using the strobe indicator disc on the drive spindle which requires a 60 (or 50) Hz strobe light.
But general lighting LEDs don't strobe smoothly - there's too much decay time (and many are rectified to DC) so you'll have a hard time resolving the strobe disc. Not to mention that household current isn't always dead-on 50 or 60 Hz.
Fluorescent tubes work better for strobe disc resolution, but better still is a smartphone app that strobes the phone's internal LED flash lamp like "RPM Meter & Torch Light" for iOS. Annoying ads, but it does the trick.
True. I have LED lighting everywhere in my house and none of the bulbs strobe at the mains frequency, otherwise they would have driven me insane. I do have a cheap chinese dimmer for LEDs bulbs in one of the bedrooms that uses PWM with varying duty cycle. It strobes at 120 Hz when dimmed all the way down and I absolutely hate it. I guess I could use it along with the markers on the wheel, but then I would not know if it's hitting the exact RPM or one of the harmonics. Using the free iPhone/Android apps that can strobe at the full revolution frequency is the easiest.