Amiga 1200 repair analysis

[tester23]

Hi,

I hope it's ok to post repair help requests. I am stuck...
Amiga 1200. Working with original Amiga DD drive. Kickstart 3.1.

• Day 1: using Amiga for several hours straight. Formatting new diskettes and running demos for verification.
• Day 2: Amiga ignores floppy after checking diskette and boots straight from CF Card. Realize that I can't format diskettes or read anything. Except: first 20 seconds after Amiga boots.

What I tried so far:

Amiga Test Kit - alignment test and read test:

• genuine floppy shows all sectors of top head ok, all on bottom side broken.
• pc converted drive: all bottom one broken, random output on top side.
• tested that with several new diskettes. Both DD and HD type.

Amiga Test Kit - floppy pinout test

• Amiga to Drive signals look ok: motors spins, head is stepping, "side" pin is acting as expected, computer reacts to changed disk, drive is doing the "tick" when empty.

Other:

• Measured the resistances from floppy to Paula - all correct.
• Measured the decoupling 100pF caps and 27 or 68 Ohm resistors. All ok.
• Changed floppy cable 3 times already.
• Audio is correct
• Mouse is correct
• Memory tests are passing

What I found wrong or suspected:

• VCC (5V) was a bit low, around 4.7V using a meanwell psu. VCC to GND resistance was 7 Ohm - an almost short.
  Found the shorted capacitor C16A (0.22uF). Replaced with new.
  Resistance is now back to 60 Ohm and it behaves as expected (capacitor charging via multimeter).
  VCC is now 4.9 when running with CF and floppy drive attached.
• After fixing the short both drives are acting erratically - still impossible to boot, impossible to format or read any disk.
  Drive read test shows fully broken bottom part and randomly broken top side of each floppy and each drive.
• Paula runs at around 40 Deg C. when doing nothing and nothing connected. Not sure if it's normal for this chip.
• Alice is running also around 40 C, but I read it's 'normal' for this one.
• DKRD signal from floppy has a filtering capacitor 100pF. This capacitor to ground resistance is 1 kOhm. All others are 1.6 kOhm. I changed the cap to the new one, but it was the same value measured. Reinstalled the original capacitor back.

Next steps:

• Ordered new Paula chip just in case. My version is: 8364R7PL
• Ordered new 74HC86D just in case - it is used for calculation of 2 signals sent to floppy - but it's for writing as I understand.
• Ordered a "new to me" tested Sony MPF disk drive, to convert and test in amiga.
  I am not sure if "too low" voltage could damage the floppy or I somehow managed to get really unlucky to damage both drives on Day 1 and Day 2 with diskettes.
• I am now testing all traces from Paula to everywhere searching for a break, but original symptom with only 1 side of drive broken is ruining my analysis.
• Check with oscilloscope all signals on the floppy. Especially DKRD.

Can't understand:

• How can it be broken in half - only bottom part of disk broken always on every drive, if SIDE signal is ok. If Paula would be damaged then it would consistently fail all the time as the data come always on the same trace.

So far finding the short on the C16A was the only fault I confirmed without doubt.

Ideas anyone??

 

[arnljot]

Mechanical problems in the drive is to me the most likely candidate. Have a look on YT on cleaning drives, Jan Beta, Adrians Digital Basement and a few other of the reputable ones. Any talk of touching the track zero sensor and you click out.

Clean it, dust it out and gently clean the heads with isopropanol and put some new white litiumgrease on the rod that transfer motion from the motor to the head

 

[tester23]

Mechanical problems in the drive is to me the most likely candidate. Have a look on YT on cleaning drives, Jan Beta, Adrians Digital Basement and a few other of the reputable ones. Any talk of touching the track zero sensor and you click out.

Clean it, dust it out and gently clean the heads with isopropanol and put some new white litiumgrease on the rod that transfer motion from the motor to the head

Hi, the PC drive I had was realigned by me in 2021. Completely stripped, cleaned, recapped, adjusted. It would be weird to have both drives failing on the same day. On the amiga drive I only cleaned everything, but learned my lesson and didn't touch the sensors.
I will get the new pc drive that's easy to convert in some days (sony mpf920) and try this one. Now I don't have access to anything more and if then system works, I will know for sure that both drives failed and I can focus on them.

My other idea is setting up a gotek. Gotek should not have any mechanical issues, heads, sensor etc etc.


@arnljot
if you have the possibility and time, would you be so kind to confirm that my temperature on Paula and Alice are correct or not (37-40 Deg C) and empty board resistance on 5V line to GND should be around 60 Ohm?
My board is 1d4.

 

[arnljot]

Gotek is a good idea, I missed the part where you had tested two drives and found a shorted cap.

Still, one side "dead" seem so drive side related to me, only you see it on two drives which I agree with you points to the computger side of the floppy cable. Only it would be wild if data from either side of the floppy was sent down different datalines, wildly inefficient.

fwiw, Preplexity.ai says:
This pattern strongly suggests a failing Paula floppy read channel or related filtering, not the drives or SIDE signal logic.6502+1

What the symptoms imply​

• Only one disk side consistently failing on multiple drives and disks, while SIDE toggles correctly, points to the analog read path for one head, not mechanics or alignment.amibay
• The fact that it works for ~20 seconds after a cold boot and then degrades fits a marginal component that drifts when it warms (Paula internal analog circuitry, a nearby transistor, or RC network).amiga+1
• “Broken bottom side, random top side” on different drives, with the same DKRD line and same Paula, rules out both drives and cable and strongly implicates the controller side of the read path.amiga+1

Why “half the disk” can fail​

• Both heads share control lines (MTR, STEP, DIR, SIDE), but each head has its own analog path (head coil, pre‑amp in the drive, then into Paula through a small number of passive parts). A fault affecting only one channel can corrupt all sectors of that side while the other remains good.6502+1
• Amiga Test Kit showing a clean pattern on one side and consistently bad on the other is exactly what you would expect if one head’s read signal is attenuated, noisy, or badly shaped before it reaches Paula’s digital PLL/decoder.YouTubeamibay

Things that are probably fine​

• Drives: Multiple drives (original and PC‑converted) show the same “one side bad” behavior, so both being bad in the same way is very unlikely.amibay+1
• SIDE, motor, step, index, disk‑change lines: You already see correct stepping, motor control, SIDE toggling, and OS reacting to disk insertion/eject – that matches a working digital side of the controller and CIAs.eab.abime+2

High‑value checks​

In order of usefulness:

• Compare DKRD waveform cold vs after 30–60 seconds:
  
  • Use a good, known‑working DD disk with a repeating pattern (e.g. a game or demo that works in another Amiga).
  • Probe DKRD at the drive and again at the Paula‑side of the series resistor or filter.
  • Look at both sides (SIDE low vs high) and see if the amplitude or shape changes significantly once the failure appears.amibayYouTube
• Carefully re‑inspect and reflow the small passives in the floppy read path:
  
  • Any 100 pF caps, 27–68 Ω resistors, and the DKRD filter parts around Paula. A hairline crack or heat‑sensitive resistor can slightly bias the signal so it only works cold.amibay
  • That one DKRD filter node with 1 kΩ to ground versus 1.6 kΩ on the others is suspicious; it may indicate a leaky Paula input or a borderline resistor on that net, even if the capacitor itself is good.amibay
• Check for contamination or flux residue:
  
  • After the C16A short repair, make sure there is no conductive flux or debris between the DKRD pin, its series resistor, and ground or 5 V traces. Even megaohm‑level leakage can distort a weak analog pulse.YouTubeamibay

About voltage and Paula temperature​

• Brief operation at 4.7 V is unlikely by itself to kill Paula, but running with a hard short (C16A) may have stressed some internal section, especially an analog input stage.6502+1
• ~40 °C on Paula and Alice idle is within the typical range for an A1200 and not a clear fault indicator on its own.6502+1

Planned replacements​

• New Paula (8364R7) is the most promising next step:
  
  • Given: both sides and multiple drives affected in a highly asymmetric, thermal‑sensitive way, and all passives/cables check out.
  • If the new Paula restores full two‑sided read reliability, that confirms an internal failure in the original chip’s floppy read circuitry.6502+1
• 74HC86: mainly involved in write‑related logic; swapping it is low risk but unlikely to cure a pure read‑side, side‑select‑specific problem.amibay
• “New to me” Sony MPF drive: good to have as a reference, but if it shows the same “bottom side always broken” pattern, it further nails the problem to the motherboard signal path or Paula.amibay+1

Practical recommendation​

Given everything already tested, the most probable fault is a partially damaged Paula input or the very small set of passives in the DKRD/head read path, with behavior drifting as the board warms up. Verifying DKRD quality with a scope before and after the failure window, then swapping Paula, is the most direct way to confirm and resolve the issue.

 

[tester23]

Yes, similar ideas. As I will be waiting probably around 1 week to get Paula here, I can try to get the scope on the pins with DTRD. As other AI chatbot told me, often lower head is weaker signal. Then maybe if I have a "start to die" Paula and gotek could help. I need to set it up. Never used one before and got one randomly with other stuff.


EDIT:

Latest update: I found Gotek; Updated to latest firmware 3.44. Put the ADF file there..... IT WORKS. So if the output signal from floppy is perfect then it's working. I now await for the real floppy.
I am now thinking if maybe it was possible that I somehow contaminated or damaged the heads because I got a pack of around 50 DD diskettes, which turned out to be around 20 totally unusable and 30 formattable using XCopy. If these diskettes were somehow a problem maybe they managed to somehow damage my drives.

Other scenario is that my shorted capacitor maybe did something to drive(s). I.e. lower Voltage leading to increased current on some pins?? I don't know.

Still it's possible that Paula needs to be replaced if brand "new" floppy signal won't be good enough.

Opinions anyone on this one??

EDIT:
Probably last update.
I have no clue but it's fixed. After fitting the random Gotek and checking if perfect signal is recognized; I reconnected the Amiga drive and now I knew that Paula IS ABLE to see the tracks on both sides.
And then I had "a moment"; What if my >>good<< disk is not good. I found one of disks that arrived written externally by not any of my drives.
Disassembled the stepper and the head attachment to be able to move head freely with fingers only - and here it was... low side perfect read along with top.
Then reassembled the whole head transport and adjusted position of the motor to have the same result. It's fixed now.

BUT:
I HAVE NO IDEA, how heads got moved. It is not possible. All screws were in good location and firm.
Only 2 things which were done:
- I found the capacitor which was causing short.
- I was formatting lots of floppies the day before fail and those floppies were very old.
But still... no method to damage the floppy head.

 

[awmosoft]

It is possible hat VCC (5V) damage the floppy head. I had once blew up a HD diskdrive in my Amiga 4000, because the caps were leaking.