EME-156 mechanics

[Edstrom]

I have just joined the forum as I bargained a PCW8512 really cheap a few weeks back. It had no keyboard, disks or printer so I have poked around with it, bought a keyboard, changed drive belts and hooked up a HxC floppy emulator and it boots. The FD1 however had a jammed stepper so the head could not move even with quite some finger force. After reading through some threads here at the forum I ripped the stepper open and could get the head moving by gripping onto the axis directly. Then here is my problem, when I tilted the drive a brass cotter fell out and I have no idea where it came from, please help me if you can identify the part and where it belongs:

[||C|-|E||]

Ah, that is the write protect pin  In this link below yoy can see where it goes:

http://www.cpcwiki.eu/index.php/Changing_the_drive_belt

[Bryce]

You're going to need an original commercial disk (Game or utility) to re-align the drive as you have removed the stepper motor.

More info on the write pin here too: http://www.cpcwiki.eu/forum/amstrad-cpc-hardware/help-i-changed-the-drive-belt-and-now-my-disk-drive-is-suddenly-read-only/

Bryce.

[Edstrom]

Ahh, cool, I don't seem to the be the first who get this surprise! 

I have a CP/M PLUS - LOCOSCRIPT disk which I hope is readable still and that might suffice for re-alignment, I have no idea how to re-align but I am sure there is a description somewhere.


Thanks a lot for the prompt replies!

[Bryce]

Re-aligning isn't a simple process and I don't think there's a description anywhere here. What type of oscilloscope do you have? It's a little bit easier if you have a DSO with fancy triggering.

Bryce.

[Edstrom]

I have a Rigol DS1052E, quite cheap but work pretty well I think.


If you can give me an outline of the process I can give it  a try and write up the result for future reference

[Bryce]

That's more than good enough to re-align a drive.


Generally... The motor housing doesn't have fixing holes, it has a sort of cog construction, so the housing rotation is variable. You need to mount the housing with the screws not fully tightened (just enough so that the housing can be adjusted with some effort). Then get the drive to read a track (ie: type CAT and hit enter) while the scope is connected to the output of the head (at the connector end so as not to disturb the head movement). It probably won't read at all on the first attempt, but you should see pulses on the scope. Now turn the motor housing slightly (1°) left or right until the pulses have reached a peak height. The drive should also be successfully showing the directory at that stage, however you need the scope to know that the head is centred on the track. When you've centred the head on the track you can tighten up the screws fully.


Bryce.

[Edstrom]

Sounds straight forward, however I now got the head searching for track 0 without finding it, stuttering against the end position.


There is also a small collar with a screw on the outgoing axis from the stepper on the other side of the mount so to speak, I can move it easily maybe it needs tightening?
Do you need a picture or do you know what I am talking about?


However, I removed the track 0 and index sensors when I changed the belt so i think i need to calibrate those first right?
Should be as simple as put the scope at the cable signal and see when it is in position right, or is there more to it?


Sorry for the flood of questions, just feels great to see progress with this thought to be dead drive! :-)

[Bryce]

Oops, then the entire mechanism needs to be calibrated. If the head is "drilling" against the the case then either the TRK00 sensor needs to be adjusted or sometimes this can be calibrated out with the moter adjustment.

Bryce.

[mjj]

Hello, I'm new to the forum!

I'm working on fixing my workmate's 6128, which has an EME-156.

We replaced the old broken gooey belt with a new one and cleaned the head with IPA. The disk spun correctly but the head motor made horrible stepper noises when returning to track 0. In spite of that, the drive was able to return to track 0 eventually and even partially read some disks (some others gave retry/ignore/cancel errors).

I then lubricated the stepper's worm gear with silicone grease but it did not help (too viscous?), so I cleaned up the grease and applied shredder oil on BOTH the worm gear and the guide rod with a cotton swab.

After lubrication the stepper jamming went away, but then the drive stopped reading our 2 test disks completely. CAT returns retry/ign/canc on all 4 sides

Things I've tried so far:

I had to disassemble the index sensor but I think I put it back in the same position based on the lacquer remains.

I've measured the index period with the scope and it's 200 ms.

I've moved the track 0 sensor, spent about 1 hour trying many different positions. No help.

I also rotated the entire head stepper to maximize TP1's Vpp as suggested earlier by Bryce. Surprisingly, I had to move it a lot (about 10º) to find the maximum amplitude. But this didn't help either.

I've even moved the worm gear collar so that the "lever" (stopper?) hits the head carriage at the same time the carriage hits the end of the guide rod. I'm not sure this is the actual purpose of the lever...

At this point I've touched so many things that I'm clueless how to move on... To summarize: I think the index sensor is in place (how can I calibrate it with the scope?). I think the trk 0 is in place, too, and the new stepper orientation maximizes TP1's Vpp AFAIK but I'm not sure of anything anymore because the drive still can't read disks and it's making me crazy.

Any clue?

Did you make it work, Edstrom?

[CanonMan]

I've also got an EME-156 drive with a loose collar (if that's the correct name for it) on the worm drive shaft. What is the purpose of this collar? I haven't managed to figure it out yet!

[mjj]

I keep thinking it's an end stop back-up mechanism in case the optical track 0 sensor fails to prevent the carriage from smashing into the stepper's mounting plate. I adjusted it as follows: first I loosened the bolt that holds the collar and the shaft together (mine was quite tight actually), then I turned the shaft all the way towards track 0 (the outermost track). With the shaft (and the head carriage) in that position, I turned the collar so that the crank opposite to the bolt points down at 7-8 o'clock and tightened the bolt so that the incoming carriage gets on the way of the crank and prevents further rotation of the motor.

I've done some progress after a few more hours messing around: now the drive loads Indiana Jones completely!!! It never made it past the load screen before, so I'm really proud of the patience I didn't know I have!!!

I can't read the other 3 sides yet though, but with lots of time and patience it feels doable now.

I've realized that the index sensor position doesn't matter much. Ie. Indy loaded just fine while I moved it around carelessly.

The track 0 sensor, however, is very finicky. The way I'm adjusting it is as follows: With the scope probe on TP1, I do a CAT, which moves the head to track 0. If the sensor is not working correctly the head will actually land on either track 1 or track "-1", maybe even track 2 if it's way off. Note that the track 0 sensor can't do "sub-track" finetuning. Ie. if your head lands between track 0 and 1 the track 0 sensor won't help. In this case you need to rotate the whole stepper as Bryce explained above!! And an oscilloscope is absolutely required here because you need to rotate the motor in small angles looking to maximize TP1's Vpp, which in one of my disks peaks at ~300 mVpp. Anyway, back at track 0 sensor: when CAT errors out and the disk is still spinning, I turn the stepper manually one step at a time towards track 0 until the signal on TP1 disappears, which means I went past track 0 intro "track -1". This way I know which track number I landed onto by the number of steps I had to take manually, and which way to move the sensor. The signal at TP1 looks like a triangular, ~300 mVpp wave when there is data recorded on that track. There is a ~3VDC bias so use AC coupling. If you don't have a scope, the only way is trial and error moving the sensor around in submilimiter (0.5 mm?) steps.

[Bryce]

It sounds like the core of the stepper motor has broken. Inside the motor there's a permanent magnet (outer dark ring) held on the shaft with a nylon ring (that white bit in the middle). These are meant to be tightly fitted to the shaft so that they all turn together, however the nylon tends to get cracked and then the magnet spins but doesn't turn the shaft enough. This gives the exact randomness you are describing. The damage will be obvious when you open the motor because the magnet can turn/slide without turning/moving the shaft. It's also noticable by the fact that you calibrate the drive, it works a few times on some disks and then suddenly doesn't read any disks the next day.


The repair: You'll need to completely disassemble the motor (as shown in the picture). Then you need to superglue the nylon ring to the shaft. USE VERY LITTLE GLUE, OTHERWISE THE MOTOR WILL NEVER TURN AGAIN!!!!! And it needs to be in the correct position at the end of the shaft, not pushed against that steel plate. When it has all set you can reassemble the motor, but the drive need a complete re-alignment and calibration once you've finished.

Bryce.


Edit: Just found the thread where I described the whole thing in detail (with fancy diagrams too!): http://www.cpcwiki.eu/forum/amstrad-cpc-hardware/fixing-a-floppy-stepper-motor/

Edit 2: One last comment. I've done this repair on many many drives since then. The rubber ring mentioned in the other thread was often completely gone (drivebelt gooey condition) and the remaining rubber was inside the stepper casing. If this is the case, you'll have to carefully scrape it out. At the end of the shaft (inside the motor casing) there's a metal spring to pre-tension the shaft, most of the goo will be on that. Scrape it off with a plastic rod (a metal screwdriver won't work as there are lots of magnets in there that make life difficult). I didn't replace the rubber, I just reassembled it with no seal and all drives are still working, some of them since over 4 years now!

[mjj]

You nailed it.

I opened up the stepper and this is what I found: cracked nylon sleeve and magnet rotating freely around the shaft. Now I can't understand how I was able to read anything at all with the stepper like this.





I just IMed the owner with this new piece of information. He gave me written consent to superglue the motor under the influence of a few beers, so I'm going to do the right thing and proceed with the superglue but I have a couple of questions:

1. What's the risk of excess glue? As long as I let it dry completely before reassembling the stepper I can't see how it can permanently lock the motor.

2. What to do you think about the state of the black rubber on those photos? I don't know what the original shape was. It's not really rubbery... Not gooey either. Actually something in between. Should I remove it completely?


EDIT, question 3: what is the "correct" position of the sleeve/magnet at the end of the shaft and what's the tolerance? (ie. how far should the shaft protrude)

Thanks, your advice is really helpful!

[GeoffB17]

In case it's any encouragement, my drive was the first one that Bryce 'fixed'.   My drive is still working fine, the only problem - if it's a problem - is that the drive now works so much more silently/smoothly than it ever did before (that I can remember) that I frequently need to check that it is actually working!

Great to hear that he's been able to use the expertise to save a lot of other drives as well!

Geoff

[mjj]

Answering questions 2 and 3 myself: I ended up scrapping all the rubber remains with a cotton bud and IPA, I don't think that black crap was being useful anymore.


And I think the correct sleeve position is where the tiny leftover part of the nylon sleeve is now, as seen in the photos below:

[Bryce]

You nailed it.

I opened up the stepper and this is what I found: cracked nylon sleeve and magnet rotating freely around the shaft. Now I can't understand how I was able to read anything at all with the stepper like this.





I just IMed the owner with this new piece of information. He gave me written consent to superglue the motor under the influence of a few beers, so I'm going to do the right thing and proceed with the superglue but I have a couple of questions:

1. What's the risk of excess glue? As long as I let it dry completely before reassembling the stepper I can't see how it can permanently lock the motor.

2. What to do you think about the state of the black rubber on those photos? I don't know what the original shape was. It's not really rubbery... Not gooey either. Actually something in between. Should I remove it completely?


EDIT, question 3: what is the "correct" position of the sleeve/magnet at the end of the shaft and what's the tolerance? (ie. how far should the shaft protrude)

Thanks, your advice is really helpful!

Some answers:

1 - I had a drive that someone tried to glue, but they had pushed the magnet / nylon up against metal plate before it dried. Not only was the magnet in the wrong place, but the glue had got into the bushing on the plate and it took hours to get it to turn again.

2 - All that black goo in the second picture needs to go. The end should be a shiny nickel coloured surface (and it's a spring). The goo was originally a ring on the nylon.

3 - The correct position is pretty much where it is in your first picture. Plus/minus 1mm would still allow it to work, but may cause it to rub against one of the ends.


By the way, I don't actually remove the shaft from the drive to glue it, I do the repair with the shaft in place.

Bryce.

@GeoffB17: Great to hear that the oldest of these repairs is still working (since early 2015 I think). The extreme quietness suprised me too, I don't remember drives ever being that quiet back when they were new, but any drive that I've had to fix that way and fully calibrate are barely audible afterwards.

[mjj]

Thanks for the answers!


But I sill can't read disks. Not even Indy, which loaded fine with the broken sleeve.



I glued the magnet and the bigger part of the sleeve to the smaller part that remained at the inner end of the shaft. I let it dry long enough and it appears to spin freely without rubbing or sticking to the stator body.


Previous to that I had scraped all the rubber seal remains with IPA + bud.


I recalibrated the stepper angle to maximize Vpp at test point TP1. I took a few scope screenshots, which I'll post later today.


Then, I recalibrated the track 0 sensor so that CAT lands on the first track that has a visible signal on TP1. I'm assuming that this is track 0 and CAT goes to this track to search for the directory but I'm not even sure about that. Surprisingly enough, track 0 is being detected at a different stepper angle on different disks or even sides of the same disk and this is blowing my mind because it makes no sense whatsoever.


What I thought was the index sensor was just the LED that lights up through the index hole and that's why its position is somewhat irrelevant. I haven't really touched the "index sensor", if there is actually one.


I think I have everything the same way it was when Indy loaded up just fine.


However, I found a tiny itty bitty bearing ball (about ~1 mm in diameter). Is there a ball bearing on the carriage where the ball may have dropped off from?

[Bryce]

Hmmmm, I'm kind of running out of tips at this stage. Have you checked the entire PCB for dry joints, especially on the connector headers?
I've no idea where that ballbearing came from.

Bryce.

[mjj]

So... In the end my workmate just got a Gotek so I returned the 6128 to him today without being able to fix the drive

I checked the solder joints with a magnifier and couldn't see anything cracked.

I looked at the drive schematic and learned that TP5 is the output of track 0 sensor's collector so I was able to calibrate track 0 more accurately looking at TP5 with the scope, but it didn't help.


Thanks all for your help!! It this thread gets an update in the future I might ask for the drive back and give it a go... I don't like quitting things halfway

As promised and for the record, here are scope captures of TP1 while the disk spins with head on track 0. This is approximately the highest amplitude I was able to get after fiddling with the stepper angle. Some disks had higher Vpp than others. Even within the same track some stretches had notably lower amplitude than others. I guess these are areas that have been demagnetized or somehow worn out over time.

[Roverius]

Hi,


I know it's a old post, but i want to allign the stepper motor with the use of a scope.
It is a EME-156 model ME56PB31
I was searching for the TP to connect the scope to.
I don't find any TP on the board, so maybe someone can point me to where to look at.


Thanks.

[gerald]

It is a EME-156 model ME56PB31
I was searching for the TP to connect the scope to.
I don't find any TP on the board, so maybe someone can point me to where to look at.

the test points are grouped on a 5 pin connector at the upper right side on https://www.cpcwiki.eu/imgs/c/c4/EME-156_Bottom.jpg

[Roverius]

the test points are grouped on a 5 pin connector at the upper right side on https://www.cpcwiki.eu/imgs/c/c4/EME-156_Bottom.jpg

Thank you, how i've missed that :/ So, if i am right, i should test on TP1 that is on the left right of the 5 pins for allignment.  For track zero sensor, that should be TP5 then.
Sorry for this questions, but those are new to me

[gerald]

On TP conector :
pin 1 or pin 2 for head amplifier
pin 3 is GNDpin 4 is the index hole detection (high on index)
pin 5 for track0 sensor (low when on track 0)

You usually do not need to touch the track0 sensor.
You can use TP4 to adjust the rotation speed. You want a 200ms period between pulses.

[Roverius]

Thank you !


Drive is working perfectly again