Oh dear, did I fry my KryoFlux?

All questions about how to use KryoFlux go here.
User avatar
Ash
Posts: 10
Joined: Thu Jun 04, 2020 5:48 pm

Oh dear, did I fry my KryoFlux?

Post by Ash »

I was overly excited to have obtained a KryoFlux yesterday and set forward to setting it up. I am can accept blame and understand if I need to buy a new KryoFlux if this is my fault (even though it makes me sad that the delivery of a new board will probably take a while again). However, my biggest concern is understanding where things went wrong, because I do not want to fry another board.

I acquired the premium package, so I wouldn’t be plugging in anything that would unexpectedly cause an issue and followed much of the instructions covered in the quick-start guide with a supported drive. I have acquired a Samsung SFD-321B (I have never used this drive before, but the seller who provided it to me told me it works fine) which was described as supported on these forums.

I was careful with the instructions and ensure that data line, met with 1. I plugged the cable in with the notch orientation displayed on the drive.
drive-connector.jpg
I believe that these pins were correct, followed through the instructions; very carefully verified the cable was aligned to the connector multiple times through the course of testing this.
Where I think I may have screwed up, I did not read the instructions properly and I did plug in the power external power supply’s Molex connector also into the KryoFlux board; following the instructions, I did not actually turn on the external power supply until much later once I had gone through running the “dtc -c2” commands. I would see the drive spin up, but all I would get in the command prompt is:
Control command rejected by the device CM: maxtrack=0
I went carefully through the instructions, realized the Molex connector was connected to the board and it should not be and removed it.
I then went through much of the tutorial again, trying to do things the right way. By this point, the controller when connected fresh to USB always returned:
Control command rejected by the device CM: maxtrack=0
This was regardless if the drive data cable was connected or not to the drive, regardless if power was being provided to the drive or not.

I had noticed a burning smell and immediately disconnected everything, rechecked all the cables and everything were in the right place, they were. I went through everything again, very carefully checking the temperature of each component at each step. Eventually, I found when I turn on the power supply, that was supplying power to the floppy drive, one of the chips on the KryoFlux becomes too hot to touch within seconds (highlighted in the picture). Turning off the external power supply to the drive causes it to stop getting so hot.
kyroflux-suspicious-chip.jpg
At this point, I am not sure:
  • If that chip is expected to get that warm? I looked up the chip on Google it seemed to be a SN74LVC244A, an octal buffer/driver with 3 state outputs, nothing about that description seemed like it should be generating that much heat; which is why I suspect I’ve fried this board of dreams.
  • If the floppy drive was faulty to begin with, causing this.
  • If the floppy drive is faulty because of what I did earlier.
  • If I fried the KryoFlux earlier and this is the result of that.
  • If I damaged the provided power supply through my mishap earlier.
I have ordered a new floppy drive, this time a NEC FD1231M which is apparently supported according to this forum; but I am not too enthused about plugging it in on the off chance of frying that one too.

I am not sure where to approach this at this point; judging from the other comments on other issues on this forum, I would really appreciate advice that does not require me obtaining a multi-meter or frying more things.

Any guidance on where to go on from here would be appreciated.

Edit: Fixed some typos that were bothering me rereading this.
Last edited by Ash on Sat Jun 06, 2020 2:45 pm, edited 2 times in total.
/* --- Text data --- */
brightcaster
Posts: 302
Joined: Fri Nov 08, 2013 10:48 pm

Re: Oh dear, did I fry my Kyroflux?

Post by brightcaster »

That looks as if you are feeding voltage from the drive into the kryoflux board! I would disconnect the drive from the board, power up the drive alone and look for the pins 8, 26, 28, 30 and 34 on the drive's connector. There shouldn't be more than 5V!

Later Samsungs SFD-321B contain four diodes like seen in this pictures:

https://www.richis-lab.de/images/floppy/13.jpg

https://www.richis-lab.de/images/floppy/11.jpg

They are 1N118 from the stepper to GND. Without them the driver/controller of the stepper gets burst with intensive stepping sometimes (like calibrating the drive)....

So if you don't have them on the drive's board and the kryoflux board keeps cool while disconected to the drive you may be lucky frying the drive only! The Kryoflux data lines are able to sink 50mA per pin for a while. That's quiet a lot, and chances are good that they survived the torture...

David
User avatar
Ash
Posts: 10
Joined: Thu Jun 04, 2020 5:48 pm

Re: Oh dear, did I fry my KryoFlux?

Post by Ash »

brightcaster wrote: Sat Jun 06, 2020 10:44 am That looks as if you are feeding voltage from the drive into the kryoflux board!
So, the earlier connection I made to the KryoFlux board with the Molex connection should not have done anything?
brightcaster wrote: Sat Jun 06, 2020 10:44 amI would disconnect the drive from the board, power up the drive alone and look for the pins 8, 26, 28, 30 and 34 on the drive's connector. There shouldn't be more than 5V!
I have ordered a new multimeter. I understand I would have to switch the multimeter to measure voltage when testing these pins, when it comes to placing the probes, is that for example...

Place negative probe on 8, positive probe on 26, then move on to negative probe on 28 and positive on 30 etc?

I expect considering what I write below, this will not be necessary for this drive. However, I would like to exercise better caution in future if I can, so this is still invaluable knowledge for me.
brightcaster wrote: Sat Jun 06, 2020 10:44 am Later Samsungs SFD-321B contain four diodes like seen in this pictures:

https://www.richis-lab.de/images/floppy/13.jpg

https://www.richis-lab.de/images/floppy/11.jpg
I disassembled the drive and it seems there are no diodes like you have indicated.
diodeless-connector.jpg
brightcaster wrote: Sat Jun 06, 2020 10:44 amThey are 1N118 from the stepper to GND. Without them the driver/controller of the stepper gets burst with intensive stepping sometimes (like calibrating the drive)....
I am really grateful for your help, even if the board didn't survive; I at least now have a good impression what what the cause was.
brightcaster wrote: Sat Jun 06, 2020 10:44 amSo if you don't have them on the drives board and the kryoflux board keeps cool while disconected to the drive you may be lucky frying the drive only! The Kryoflux data lines are able to sink 50mA per pin for a while. That's quiet a lot, and chances are good that they survived the torture...
Do you think I would be safe plugging in the NEC FD1231M I ordered into the same power supply and KryoFlux board or should I be doing other tests before doing so?
Last edited by Ash on Sat Jun 06, 2020 2:45 pm, edited 2 times in total.
/* --- Text data --- */
brightcaster
Posts: 302
Joined: Fri Nov 08, 2013 10:48 pm

Re: Oh dear, did I fry my Kyroflux?

Post by brightcaster »

You should measure the pins with the negative probe connected to GND. That's the black wires on the molex or the metal shielding of the drive.

To be save, you can measure also the output voltage of your supply. It should be 12V on the yellow wire and 5V on the red one (negative probe connected to one of the black wires again).

If that's okay I would give the NEC a chance...

David
User avatar
Ash
Posts: 10
Joined: Thu Jun 04, 2020 5:48 pm

Re: Oh dear, did I fry my KryoFlux?

Post by Ash »

brightcaster wrote: Sat Jun 06, 2020 1:12 pmYou should measure the pins with the negative probe connected to GND. That's the black wires on the molex or the metal shielding of the drive.
Ah, I understand!
brightcaster wrote: Sat Jun 06, 2020 1:12 pmTo be save, you can measure also the output voltage of your supply. It should be 12V on the yellow wire and 5V on the red one (negative probe connected to one of the black wires again).
Will do~

I will post an update on how this progresses.
/* --- Text data --- */
ZrX
Posts: 618
Joined: Tue Dec 06, 2011 9:09 pm

Re: Oh dear, did I fry my KryoFlux?

Post by ZrX »

Where do you live, if the board has been damaged and it would be viable to try getting it repaired instead of buying a new board?
User avatar
Ash
Posts: 10
Joined: Thu Jun 04, 2020 5:48 pm

Re: Oh dear, did I fry my KryoFlux?

Post by Ash »

ZrX wrote: Sat Jun 06, 2020 6:12 pm Where do you live
I am located in Northern Ireland. It would not be a big deal to mail the board to somewhere in Europe (although the mail system is a bit slow at the moment).
ZrX wrote: Sat Jun 06, 2020 6:12 pmif the board has been damaged and it would be viable to try getting it repaired instead of buying a new board?
Open to trying anything. Considering a new KryoFlux board costs roughly £100; it is kind of in that sweet spot, where it might not really to be worth most people's time to test, replace components for cheaper with no board schematic readily available.

I could alternatively try doing it myself, although, lacking the appropriate tools for this sort of repair, I would suspect I would end up spending more than £100 in acquiring said tooling and practice material -- would need to research more to find out!
/* --- Text data --- */
brightcaster
Posts: 302
Joined: Fri Nov 08, 2013 10:48 pm

Re: Oh dear, did I fry my KryoFlux?

Post by brightcaster »

The schematics are already on your computer! There is a directory called schematics in your kryoflux installation. The circuit in question is IC3...

David
User avatar
Ash
Posts: 10
Joined: Thu Jun 04, 2020 5:48 pm

Re: Oh dear, did I fry my KryoFlux?

Post by Ash »

brightcaster wrote: Sun Jun 07, 2020 12:36 pm The schematics are already on your computer! There is a directory called schematics in your kryoflux installation.
Oh wow! I stand corrected. This is really helpful. ♥️
brightcaster wrote: Sun Jun 07, 2020 12:36 pm The circuit in question is IC3...
Thank you again~
/* --- Text data --- */
SomeGuy
Posts: 283
Joined: Wed Feb 18, 2015 8:18 pm

Re: Oh dear, did I fry my Kyroflux?

Post by SomeGuy »

brightcaster wrote: Sat Jun 06, 2020 10:44 am Later Samsungs SFD-321B contain four diodes like seen in this pictures:

https://www.richis-lab.de/images/floppy/13.jpg

https://www.richis-lab.de/images/floppy/11.jpg

They are 1N118 from the stepper to GND. Without them the driver/controller of the stepper gets burst with intensive stepping sometimes (like calibrating the drive)....

So if you don't have them on the drive's board and the kryoflux board keeps cool while disconected to the drive you may be lucky frying the drive only! The Kryoflux data lines are able to sink 50mA per pin for a while. That's quiet a lot, and chances are good that they survived the torture...

David
Just to be clear, are you saying this model of drive without the diodes installed can actually potentially damage a kryoflux (and perhaps other controllers)?
Post Reply