Another Strange 1050 Fault

[cwc]

I’ve got four 1050s (2 x Atarimax Happy, 1 x US Doubler clone and 1 x original Happy. This last drive is now giving me problems. It reads and writes perfectly but often has problems formatting (known good) disks in SD, ED and DD. At the moment ED formats seems to work more often than SD or DD ones but formatting in all densities does occasionally work.

 

It’s not the drive mechanism as I’ve swapped it into one of the other drives and it works perfectly. It’s not the Happy as I’ve tried that in another drive and it works perfectly too. Swapping a known good mechanism into the problem drive results in the same formatting issues. The culprit appears to be the main 1050 board which is a Revision C3.

 

Symptoms (testing using Prosystem for Windows via USB): When the drive receives the format command it spins up and begins formatting immediately. On SD and ED formats the head gets to the last track (39?). More often than not instead of stepping back on the verify phase the head grinds back to track 0 then back to track 39 and the format stops with an error. Occasionally on DD formats, the head reaches track 39, pauses, goes one track backwards in the verify phase then stops with an error.

 

Any ideas what might be causing these problems? Any help would be greatly appreciated.

[_The Doctor__]

1. First place to look is speed RPM needs to be correct. Use an rpm tester. make sure its dead on or one rev slower.

2. Check voltages and power are clean. (voltmeter then oscope)

3. Gently make sure are chips are pushed in.

4. Gently wiggle connectors (just enough to move them around but not bend / break anything and only gripping plastic not wire)

[Kyle22]

Possibly, if not any of the above, it could be the analog amplifier circuitry between the head and the FDC.

[cwc]

Many thanks. RPM is fine. Chips are pushed in. Connectors are fine. I will look at voltages next.

Edited June 9, 2015 by cwc

[cwc]

Using the relevant Test Points I get 12.47v and 4.88v. One of the other 1050s I have gives 12.18v and 5.09v. Looks like the faulty drive's 12v supply is too high and its 5v supply is too low. Any ideas? Replace Voltage regulators / large caps?

[1050]

2% plus or minus regulation shows you just barely outside normal, change the regulators if you want. Get 7805A and 7812A for 2% regulation, other numbers may have even looser regulation. The real measure is the supply side to be regulated, you should have 9 VDC going in to the 5 volt regulator and close to 20 going into the 12 volt one. If these are lacking then look to caps/diodes. Be sure to check these and your first choice of test points while attempting format.

[_The Doctor__]

@1050 you are spot on again! Glad you are here!

 

@kyle22 If the power supply ends up clean, the amp circuit is indeed the next step

 

one day all the disk repair forum threads need to be consolidated and organized... a quick list of things to do in general order generated.

[cwc]

Many thanks. Great to have you all on board! Voltage measurements: I've got 11.5v going in to Q7 (5v regulator) and 4.9v coming out. The 4.9v remains steady during a (failed) SD format. I've got 23v going into Q8 (12v regulator) and 12.5v coming out. The 12.5v drops down to 12.3v lowest during a (failed) SD format then goes back up to 12.5v. On Test Point 13 (5v) I'm still getting 4.88v which remains steady during a (failed SD format). On Test Point 14 (12v) I'm still getting 12.47v which drops to 12.23v lowest during a (failed) SD format. Only issue appears to be the slightly high input on Q7 but it appears to be being regulated down to the right level. Nothing noticeably wrong with caps. Solder joints look OK. Checked some other voltages across the board as per Sam's Computerfacts and all were fine. Swapped U13 (FDC) with known working one from one of my other 1050s and the symptoms were still there.

 

Any more ideas? Assume we now need to look at the circuitry between the FDC and the head as Kyle22 suggested?

Edited June 10, 2015 by cwc

[1050]

Other than throwing chips at it, I think it's time for a frequency counter and scope. Two 3086 chips can be changed for new and you might try a different 6532 to see if the problem follows it, but unlikely is my guess. C36 needs adjusting is my hunch as I have several just like yours with difference, they do work fine once warm but when cold will not work well at all. Most of these got this way by being heavily used Happy drives that used to work just fine. Not blaming Happy here, just seems to happen a lot that they tend to fail in this manner - cold only format fails at first and gradually getting worse as time goes by. I'm thinking data separator circuits needs a bit of fine tuning after 3 decades.

 

Nothing wrong with those supply voltages but both regulators just over/under where they should be. Another long shot in my book, but you are sure to correct that problem by replacement easy enough. Have yet to get one fixed with this problem, so I'm out of helpful ideas at this point.

[cwc]

Many thanks. An update: I tried swapping the 6532 with a known working one from one of my other 1050s and the symptoms remain. Adjusting C36 gradually clockwise improved formatting up to a point where disks would get almost to the end of the verify stage on an SD format, but I could get no better than that. I'm ordering new 7805, 7812 and 3086s. I will replace them and provide an update next week. Just before I place the order - is there anything else I could try and replace at this stage?

[1050]

Many thanks for the update, a vital and all too often missing aspect of these types of posts !! Very much appreciated here.

 

Your adjustment technique for C36 is not standard but lacking a scope, it's all we can do. This adjustment centers the frequency of the data separator to 250 Khz, one then also needs to adjust a resistor, VR3, for one eighth time frame duty cycle at that frequency to have the circuit doing it's best job. I can't find my reference for this information today, but I do remember seeing a scope reading of adjustment for the 2793 FDC, too bad it's not on the Sam's photofact or 1050 field service manual, when I find it, I'll post a link to it. Tweak VR3 and then C36, one then the other until fixed?

 

Nothing to replace until you can find something broken in my opinion.

Edited June 20, 2015 by 1050

[cwc]

Update: I tried adjusting VR3 / C36 without success, losing the ability to read/write in the process. So I put VR3 / C36 back to their original positions so that the drive now reads and writes in all densities perfectly once more. I've just noticed that aborted formats now do not wipe the disk - the data remains intact - so, there is no write during format. Current symptoms: When the drive receives the format command it spins up and begins formatting immediately. The head gets to the last track (39?) and instead of stepping back on the verify phase the head steps back very slowly track by track and the format eventually times out with an error. As, after a period of unreliability, the drive has finally lost the ability to format (but, as far as I can tell, works perfectly in all other respects) does this suggest a failed component i.e. something that would interfere with the signals to commence / undertake / complete the verify stage of a format? Any ideas? I'm certainly no expert but the U13 FDC datasheet suggests that formatting simply involves writing and as writing is fine in all other respects, there can't be anything major wrong with the drive? Seems a crying shame that on a 30 year old drive I can read disks of a similar age backwards and forwards to and from a pc via usb(!), but something relatively minor could be stopping the drive from functioning just as it did when it left the factory all those years ago...

[MrMartian]

So, it can write to a disk, but not format it? That doesn't make any sense... Formatting is no different than writing a sector, just a track at a time...

[1050]

You may have found something broken. Any chance you have the head jack J6 plugged in backwards at the front of the board? This jack uses zero based numbering with pin zero as ground at the rear position on the board - it will be the braided/bare shield wire. Please use an ohm meter on it and verify 1.5 ohms across the erase head, pins 2 and 4. Should have 7.3 ohms between pins 2 and 3 and again between pins 2 and 1. With 15 ohms between pins 3 and 1. All this is just the head unplugged. Head should be OK if all numbers given are good. After all the diatribe below, I make a sketch of a five pinned jack numbered 0 thru 4, draw head windings where they are supposed to go and reverse it. It don't look good for the erase head. The read/write head is just swapped around, should work fine either way, but the 1.5 ohm winding in the erase head will NOT tolerate 12 volts across it for very long without letting out the magic smoke. Best of luck in this regard.

 

Erase power comes from U17, a 74LS06 inverter using two gates of it paralleled, perhaps it's been killed and needs replaced? You should NOT have any kind of data retained after a format command and operation - big defect here. R2 is a really big 2 watt, 150 ohm resistor right next to the J6 head jack and is also in this erase power circuit, measure it to see if it has opened up to kill the erase function. One other device that can kill the erase function is CR10, a diode in line with pin 3 of jack J6, checking it's traces you can verify that it is across pins 2 and 4 of jack J6. It it's shorted, no erase happens.

 

I believe the circuit normally works at 5 volts during writes, but is boosted to 12 volt during format by Q10, a transistor very close to R2, as is R3 another part of the 12 volt boosting circuit. Other than the mentioned final drivers above, everything else in the erase circuit is logic level signals coming from elsewhere, the front 3086 being one of these devices I mentioned in a previous post that might be changed out just for the heck of it.

 

If your erase head survived, try again with VR3 / C36?

[cwc]

Many thanks. A further Update: J6 was plugged in correctly and all the resistance readings are fine. R2 resitsance is fine as well. Replaced U17, CR10 and U21 (3086) (I forgot what a delicate operation desoldering 1050 chips was - but I managed it - new chips nicely socketed). The result is: format still fails on verify stage (drive formats to track 39 then grinds back to track 0 then back up to track 39) BUT the disk is now being erased once again. I checked normal read and write operations and they are fine as ever.

 

This is very unscientfic, but when I was idly comparing resistance readings with two of my other drives (while powered off), When I took a reading across U17 pins 9 and 6 in the problem drive the multimeter (20k setting) stayed at "1." whereas the other two drives read around 6.5. Replacing U17 has proved that it is not the cause of the issues in the problem drive but perhaps this information might be useful to someone who is better at electronics than me in pointing the waytowards which component(s) may have failed......

 

Any further help or ideas would be greatly appreciated.

 

**Edit** - Drive is not writing properly - Writes (to disks formatted in working drives) appear to work fine, but cannot be read in the problem drive or other drives (Prosystem Error - "Error Identifying Disk Format"]

Edited July 5, 2015 by cwc

[thax]

hello there, have you checked the small pressure pad above the head? if thats gone or overworn/very dirty/ that can cause problems. and also its worth checking the head is clean.

ive always found these to be most common problem on the 1050.

 

cheers

Edited July 6, 2015 by thax

[1050]

Ahhh, but I just love the unscientific method grasshopper. You are hired for sheer diligence. I'm casting a fishy eye at CR8 with that news of differing ohms between pin 6 and 9. Trouble is that they go to so many other parts I can't conjure up the exact reason for it. But can you, under power, verify that you have 4.7 volts at the cathode (striped end) of the 4.7 volt zener diode, CR8? It's located closest to pin one of U21, the front 3086. But another trouble with ohms readings of this type is that one has to double check the polarity of the meter probes and duplicate that aspect in all such attempts. You may get a different reading because the red probe was where the black probe was used on the other machine in other words. Red on pin 6 for both readings and THEN if they are different you actually may have something is another way to put it. If it's a genuine, valid different reading here, then very good work done here in finding it.

 

Edited information wrap up seems to show that erase functions fine now but writes are problematic. Good news on the head, I feared the worst. So it appears that only 5 volt write function is not up to par, CR8 is supplying balanced bias to the write amp and could be the only wrong now if it's shorted or drifted away from 4.7 volts. Resistors rarely fail so I am casting about some here blaming anything nearby with a PN junction. Everything else close by has been replaced so not the problem. Next in line is U18 and U15 both flip flops and pure logic level here. And here we start running out of rope to hang ourselves with...

 

thax brings up a point I was starting to wonder about some too. The basics. Head needs to be crystal ball clean and looking like glass. Please apply some WD-40 to a Q-tip and spend a full minute massaging it (works well but is very, very slow), dry, then clean the remains of that solvent away with alcohol Q-tip. And double check the felt pad for clean and proper fuzzyness. These two items are the highest disorders out there. thax has that part dead right.

[Guest]

i've found that alcohol on a q-tip (cotton bud) can also work

although if you head down this route - it goes without saying "don't waste the good stuff"

[1050]

The type of alcohol used plays a big part in getting good results though. It should be 93% or better isopropyl alcohol, but it turns out that strong flavor is very hard to find in some places. For me, in the states, I just have to look at more bottles and NOT settle for the usual 50% version as that flavor of rubbing alcohol gets next to nothing done where the strong stuff will peel the gummed oxide off so quickly I was stunned such a difference in results was even possible. WD-40 is common enough and it does work given enough time so I've changed my tune somewhat. 50% alcohol should be good enough to take the wiped away WD-40 off after it's done a proper job of it. Use both and get it done right the first time.

 

I purchased a not working 1050 that the seller swore up and down he had cleaned with alcohol till his fingers hurt. It had a massive patch of oxide still on it. He obviously had used the weak alcohol. I used WD-40 on it after years of just sitting around here and it slowly dissolved every part of that patch. I did this as a specific test for WD-40 suitability at this task and it works great, it just takes extra time is all. You know it's working when the white Q-tip slowly takes on a red mud color. And I've had oxide patches so small as to be next to invisible, even when I was looking right at it with bright lights I couldn't really see it, I did see the dirt on the Q-tip though.

[BillC]

I don't know about the US, but in Canada many pharmacies carry 99% ISA in addition to Rubbing Alcohol(which is usually 70% IPA here). I wouldn't recommend cleaning with rubbing alcohol as the residue left behind(from the other 30%) might damage a floppy.

[Kyle22]

I don't know about the US, but in Canada many pharmacies carry 99% ISA in addition to Rubbing Alcohol(which is usually 70% IPA here). I wouldn't recommend cleaning with rubbing alcohol as the residue left behind(from the other 30%) might damage a floppy.

The higher the percent, the better. But, do not fear the "residue," it's only water, H2O.

[cwc]

Update: Great news! Drive is formatting again in all densities and is fully functional. Tonight I established that the write issues after replacing U17, U21 and CR10 were due to a bad replacement CR10 diode. Doh! So I soldered the original back in and found that formatting worked. The culprit was therefore U21 (3086). To test the drive I have: copied unprotected SD, ED and DD disks with format before write using the Happy Sector Copier; ran the Happy diagnostics; successfully booted from a range of SD/DD/protected/unprotected disks in my ageing collection; formatted disks in all densities with and without verify using Prosystem via USB; formatted disks in SD and DD in Happy Warp Speed DOS and copied protected disks with Happy backup. Disks created in the *ex*-problem drive all read fine on all my other drives and I couldn't be happier! Note: The resistance measurement on pins 6 and 9 of U17 remains at "1." whereas I swear that a reading with probes in the same pins on my other drives is around 6.5 (Is this a red herring?) I can also confirm, for completeness, that the voltage at the striped end of CR8 in the ex-problem drive while powered on was bang on 4.7v. Nothing more to say except a BIG THANK YOU to everyone who has helped me here and in particular 1050 who was there from the beginning and whose suggestions have been invaluable in pointing the way to solving the problem.

[1050]

Fantastic, couldn't be happier if it were mine. You are most welcome and deserve full credit here - you've got a knack for this and quite a stubborn streak too. Both were essential in your success. Good job all around.

 

Only way to really know which of U17 or U21 it was would be to put one of them back in I'm thinking, but they are both essential to the erase/writing functions. When the drive goes quiet as if it's not there, it's time to replace the rear 3086 - they are known to go bad there and do that. Checking one I have here, I get '1' on 20K, 2000, and even 200 ohms settings between pins 6 and 9 of U17 either polarity, just FYI. Don't know what to make of your other two. But as usual I started this exercise testing pins 6 and 9 of the wrong chip U21 and there I got conflicting results until it was determined that one of the probes was sliding down to make contact with a resistor lead in that locality. As soon as I took pains to insure that didn't happen, my readings became consistent - it was still the wrong chip, but they were consistent. So I understand all about 'I swear I saw it'.

 

We threw some chips at it and won the gamble, sweet. Your head is clean too !!