PCjr Drive Calibration?

[jason]

I ran the floppy drive test with the JJ test plug installed and got a code “0202” which I believe means “Side 0 - Read - Missing Address Mark”?

This is what I mean by the codes’ not being terribly illuminating. Basically you can interpret this as “couldn’t read the disk”. It doesn’t really pinpoint the problem, though it probably rules out various controller issues. In fact, it leaves open the question of whether the test correctly wrote something on the disk for it to read back. But given your inability to read other disks, I think we can safely assume a reading problem.

Address marks delimit the metadata area for each sector on the disk. See: https://nerdlypleasures.blogspot.com/2015/11/ibm-pc-floppy-disks-deeper-look-at-disk.html

Thanks for the link, that will be an interesting read when I have the time!

I haven’t personally tried ImageDisk on a PCjr but am not surprised that it didn’t work. It accesses floppy controller hardware directly—not through the BIOS or DOS—and makes assumptions that don’t apply to the PCjr, including the IO port to use. The PCjr is mostly PC-compatible from the BIOS up, but at the hardware level it’s a rather different animal. You need something that specifically supports the PCjr.

Copy II PC 6.0 can test drive speed on a PCjr:

Code: Select all

copyiijr a: /t

You shouldn’t really need a program, however, when you have the drive out of the chassis and can see the bars on the pulley. If the bearings are reasonably lubricated and you don’t have a diskette in the drive, the drive should be able to reach 300 RPM and well beyond even with a slightly loose belt.

When viewed from the bottom, the drive shaft turns counterclockwise. If the speed is too low, the bars on the pulley will appear to move clockwise. If it’s too fast, the bars will move counterclockwise. If the speed is way off, the bars may not be of much help, but you can get a sense of which way it’s wrong more intuitively: A quiet, leisurely spin? It’s too slow. Spinning in high gear with noticeable motor noise? Too fast.

Usually speed problems are due to gummed-up bearings and old belts. Once you get the bearings lubed, the factory speed setting should be fairly close to right. Then it’s a matter of getting enough tension on the belt to avoid a sharp drop in speed when you put a diskette in the drive.

Copyiijr seems to confirm what I'm seeing when I test the drive on the bench standalone. I can get a good RPM on it (I believe reasonably stable) without a disk inserted, but when I put a disk in is when it starts to vary some (both via CopyIIjr output and I can hear a slight variation in sound). So I'm guessing the belt might be slipping a little bit? I tried flipping the belt over without any luck getting it to behave better. I might have to try to find some new belts. I know console5 is out of the fabric reinforced ones right now.

Just for the heck of it, I did try a rubber band to replace the drive belt and get the same behavior...so maybe it is a greasing issue or something is up with the drive speed circuit?

Which bearings are you talking about lubing? The spindle pully or the spindle drive motor (or both)? I suspect the pulley has a bearing somewhere underneath it although I'm not sure how to get it off safely without risking damage to it. The motor should just unscrew although I'd have to get inside it to grease it I imagine? Any recommended grease, or any old automotive type will do?

Thanks for the all the advice!

[Brutman]

Some variation in the drive RPM is normal when you have media inserted. I think that if you are using what is reported by CopyIIJr then the target RPM range is between 197 and 203, with 200 being spot on. There will be some variation, but it should be in that range.

(Keep in mind that CopyIIJr targets 200.0, while the real RPM is 300. Or at least that is what my older version of CopyIIJr reports.)

I think this was mentioned already but if the additional friction of a diskette is causing a large variation in speed then there is usually a tension adjustment that can be made. You want just enough tension to avoid slipping; going to far will damage the drive in the long run.

Anything in the path that might move should be lubricated. So this means both the motor spindle and the center hub spindle. I use a light sewing machine oil, and I use it sparingly. Giving it time to soak in and migrate is advisable. You need to ensure that you don't get it on the belt or anything the belt touches, as that causes slipping.

[jason]

I've had a little more time to look at the drive, and it gives me a good opportunity to mess with my new scope! I'm very rusty on using these but I managed to get a few images. I took the belt off the bad drive and lubed up the motor a little. It seemed to help it run a little more consistent speed but it's still fairly irregular.

Here's an image of the good working drive tachometer signal (yellow/ch1) and motor (blue/ch2)


Here's an image of the bad drive tachometer (yellow/ch1) and motor (blue/ch2). Note that I had to really zoom in to see the tachometer voltage.


By the manual (linked previously: http://www.primrosebank.net/computers/m ... floppy.pdf) page 91, Channel 1 was connected to J3-A1, and channel 2 was connected to J3-A2.

I'll keep poking around as I have time and shake off all my rusty circuit work which I haven't touched in ...many years, lol. Hopefully learning something in the process .

[DoctorOctal]

I can get a good RPM on it (I believe reasonably stable) without a disk inserted, but when I put a disk in is when it starts to vary some (both via CopyIIjr output and I can hear a slight variation in sound).

If you get the right speed without a disk inserted, your problem is almost certainly mechanical, not electrical.

So I’m guessing the belt might be slipping a little bit?

Yes, due to friction in the bearings and/or a loose/worn belt.

Which bearings are you talking about lubing? The spindle pully or the spindle drive motor (or both)? I suspect the pulley has a bearing somewhere underneath it although I’m not sure how to get it off safely without risking damage to it.

I went into detail about this in my earlier post, so go back over that. You don’t have to take the big pulley off—just put oil around the shaft and let it seep in as I described. You don’t have to take the motor out, either. Just put a few drops of oil on the shaft under the small pulley and work them in.

Any recommended grease, or any old automotive type will do?

No grease, just a very light machine oil, as Mike said. It should be very thin and runny.

“Zoom Spout” oil is cheap, widely available, and should work well. It’s sold under several different brands and with somewhat different labeling, but in general you should look for something that’s described as “sewing machine oil”, “turbine oil”, “good for high-speed shafts”, and the like.

You could clean out and replace the grease in the collet bearing, but at this point you’re almost certainly better off just reviving the grease that’s there with a few drops of oil.

Because the oil is very runny, it’s easy to make a mess. You need to take care to not get oil on the belt. Also, as I warned in my previous post, you should only put a few drops of oil on the collet bearing—too much, and it will run down to the bottom of the collet and make a mess on your diskettes.

I found that I couldn’t control the flow precisely enough with the spout on the bottle, so I got some 1 oz (30 mL) squeeze bottles with 1.5" 20-gauge needles. Here’s something similar to what I got: <https://www.amazon.com/Hicet-Bottles-Stainless-Needle-Novinex/dp/B076YZY5ZK/> This has been great. I can place oil a drop at a time exactly where I want it.

To recap the process:

1. Take off the belt, then lubricate the collet bearing, main shaft bearing, and motor bearing. These should all turn easily, smoothly, and fairly quietly. This is very important. Don’t skip this step.
2. Once you’ve eliminated excess friction and thoroughly cleaned up any wayward oil, put the belt back on and verify the speed of the drive without a diskette present. The factory setting should give you a speed that’s close to correct, even if the belt is old. If the speed is a bit off, you can adjust it with the potentiometer.
3. Close a diskette in the drive and check the speed. If the speed drops significantly or becomes erratic, add tension on the belt just until the belt stops slipping and the speed is nearly what it was without the diskette in place. Once the belt has proper tension, you can tweak the speed with the potentiometer if necessary. This should only be a fine adjustment. If you’ve tightened the belt as much as you can and the speed is still pretty far off, you’ll need to replace the belt. Don’t try to use the potentiometer to make up for a big speed deficit.

[jason]

I can get a good RPM on it (I believe reasonably stable) without a disk inserted, but when I put a disk in is when it starts to vary some (both via CopyIIjr output and I can hear a slight variation in sound).

If you get the right speed without a disk inserted, your problem is almost certainly mechanical, not electrical.

I should have mentioned the above plots from the scope were with the bad drive without a belt installed.

So I’m guessing the belt might be slipping a little bit?

Yes, due to friction in the bearings and/or a loose/worn belt.

Which bearings are you talking about lubing? The spindle pully or the spindle drive motor (or both)? I suspect the pulley has a bearing somewhere underneath it although I’m not sure how to get it off safely without risking damage to it.

I went into detail about this in my earlier post, so go back over that. You don’t have to take the big pulley off—just put oil around the shaft and let it seep in as I described. You don’t have to take the motor out, either. Just put a few drops of oil on the shaft under the small pulley and work them in.

Any recommended grease, or any old automotive type will do?

No grease, just a very light machine oil, as Mike said. It should be very thin and runny.

“Zoom Spout” oil is cheap, widely available, and should work well. It’s sold under several different brands and with somewhat different labeling, but in general you should look for something that’s described as “sewing machine oil”, “turbine oil”, “good for high-speed shafts”, and the like.

You could clean out and replace the grease in the collet bearing, but at this point you’re almost certainly better off just reviving the grease that’s there with a few drops of oil.

Because the oil is very runny, it’s easy to make a mess. You need to take care to not get oil on the belt. Also, as I warned in my previous post, you should only put a few drops of oil on the collet bearing—too much, and it will run down to the bottom of the collet and make a mess on your diskettes.

I found that I couldn’t control the flow precisely enough with the spout on the bottle, so I got some 1 oz (30 mL) squeeze bottles with 1.5" 20-gauge needles. Here’s something similar to what I got: <https://www.amazon.com/Hicet-Bottles-Stainless-Needle-Novinex/dp/B076YZY5ZK/> This has been great. I can place oil a drop at a time exactly where I want it.

To recap the process:

1. Take off the belt, then lubricate the collet bearing, main shaft bearing, and motor bearing. These should all turn easily, smoothly, and fairly quietly. This is very important. Don’t skip this step.
2. Once you’ve eliminated excess friction and thoroughly cleaned up any wayward oil, put the belt back on and verify the speed of the drive without a diskette present. The factory setting should give you a speed that’s close to correct, even if the belt is old. If the speed is a bit off, you can adjust it with the potentiometer.
3. Close a diskette in the drive and check the speed. If the speed drops significantly or becomes erratic, add tension on the belt just until the belt stops slipping and the speed is nearly what it was without the diskette in place. Once the belt has proper tension, you can tweak the speed with the potentiometer if necessary. This should only be a fine adjustment. If you’ve tightened the belt as much as you can and the speed is still pretty far off, you’ll need to replace the belt. Don’t try to use the potentiometer to make up for a big speed deficit.

Thanks for the recap and the links. I'm sorry I missed the details originally, with our still quite young son around the house I lose track of things often and somehow missed those steps. Sometimes it can be a week or more before I can get back to a project; So, I apologize for missing the details you so carefully laid out. Thank you for pointing me back at them. I'll go back over these steps when I have time and see how things look.

[jason]

Alright, I've made some progress!

I investigated the circuit more this weekend and started zeroing in on the tachometer signal. It struck my as odd that the tach signal was SO far off in voltage, it's supposed to be around 8 volt peak to peak, I was seeing voltages in the 10s or 100s of millivolts. And looking at the circuit the tach is fairly independent of the rest of the circuit.

I ultimately decided to open up the top of the motor housing to see where the tach wires ran and took the top off as well as the iron core and a plastic spacer.

First off, I discovered that me oiling the top of the "motor" under the sticker...wasn't really the motor. All I had done is oil the tachometer (and possibly cause issues with the magnetic sensing circuitry, ). I cleaned that all up easily enough.

Then I started comparing the two drives and noticed that when I turned the good drive, I could clearly feel the stepper motor. But on the bad drive, it felt a little too smooth. So, to make a long story short, the iron core for the tachometer was not on very snuggly to the plastic spacer that attaches it to the motor shaft (and the afore mentioned oiling certainly didn't help , although it was having issues before I did that). In looking at the iron core, it looks like it may have had glue on it at some point holding it to the plastic, I'm not really sure. But fortunately I discovered that if I flipped it upside down, then it fit nice and snug on the plastic spacer and the motor shaft. Put it back together, and now it holds speed great, even under load!

The other issue with the drive turning on with no connection to ground on the motor enable signal was the Q13 transistor (PN2222A). I replaced it with a new one and now the drive does not turn until you ground the motor enable signal.

Hopefully the rest of the drive works fine, I'll be testing that next and will post what the next series of issues are

[jason]

Okay, so did some testing tonight. Good news and bad news.

I ran the internal diagnostics and got error code "B". With the "JJ" diagnostic plug, I get error code "0204". Unfortunately, neither of these is in the Hardware and Maintenance Service manual.

I booted into DOS 3.3 on my jrIDE and the drive can read existing disk directories (i.e. "dir" works fine) but has issues actually running things.

I formatted the disk...and curiously..it only formatted head 0. So, I got a 180k disk out of "format a: /s". But...the disk will boot just fine. Why did DOS decide that I only had a either a single sided drive or a single sided disk and only formatted one head....

I peeked inside the drive head and can see heads on both upper and lower halves. I believe head 0 is the bottom one?

Off to read the Qumetrak 142 manual and see what I can learn. But this is better than I started, so.....huzzah!

[jason]

Okay, got a few hours while the kiddo was napping to look at the drive again.

I followed the diagnostics in the SAM's Computer Facts for the Disk Drive/Disk Drive Adapter Board. Specifically page 12 for "Disk Drive will not switch sides". The instructions started leading to an issue with Q7 (which I was already suspect of Q7/Q8). However, after some additional poking around and examining chip U4 more closely, it appears it is the culprit. Pin 5 is always high regardless of switching the logic on the input pints 6 and 7, thus causing the Q7 transistor to always open up to ground on the collector side. I did test grounding U4 pin 5, and sure enough Q7 then shuts off and I get 12 volts in the collector.

So off to find a 75451 or equivalent. Will follow up with more when I have replaced the chip.

I've also attached a PDF of an update to the circuit diagram I made. Some of the text was hard to read, so I poked around and filled in the areas where the labels were unreadable. Hopefully it's useful to someone. The diagram is still not completely accurate as there are still several resistors missing that were added to the as-built units (typically from +5 or gnd to a connection point).

[jason]

And success!

I replaced U4 with a 'new' 75451 and now it appears to be working normally again. I still need to do a little more testing, but everything seems to work just fine.

Does anyone know how I would go about calibrating the track and heads without the original Qumetrak disk they say you need? Or might anyone have a flux image of the disk? I'm curious how I would eventually go about fixing track/head placement issues in the future?

[Brutman]

I've been following along silently, but I have to say, great job! I have not achieved your level of debugging skill yet - not even close ...