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

Re: Oh dear, did I fry my KryoFlux?

Post by Ash »

Okay, I have never had a Sunday delivery happen before, weird times we live in... But I got the multimeter.

I tested the power adapter and that seems to be within acceptable ranges?
5v-line.jpg
12v-line.jpg
I went through and testing along the Molex adapter cable too, just because I could, the drive contacts; all looked good and within ranges.

Tested the pins 8, 26, 28, 30 and 34 -- everything was below 5v.

I decided to plugin the KryoFlux board again, reproduce the previous steps I have taken. But, nothing, nothing at all. The drive won't light up to seek now and the chip does not even get warm anymore. Software just responds with consistently:
Control command rejected by the device CM: maxtrack=0
I suspect one or the other certainly has something fried or damaged now, or I have somehow damaged the drive when disassembling the shielding earlier.
/* --- Text data --- */
User avatar
Ash
Posts: 10
Joined: Thu Jun 04, 2020 5:48 pm

Re: Oh dear, did I fry my Kyroflux?

Post by Ash »

SomeGuy wrote: Sun Jun 07, 2020 3:37 pm 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)?
Perhaps not a direct answer to your question, but on the page on which pictures reside on Google translate says:
The four wired 1N118 diodes, which connect the potentials of the stepper motor to the ground potential, are striking. They seem to have been integrated afterwards with the help of test points.
It is conceivable that the free-wheeling diodes integrated in the controller were of sufficient size, but the current flow via the internal ground led to potential increases and thus to faults in the more sensitive circuit parts of the signal evaluation. The additional diodes discharge at least part of the freewheeling current and thus dampen the interference potential.
/* --- Text data --- */
brightcaster
Posts: 316
Joined: Fri Nov 08, 2013 10:48 pm

Re: Oh dear, did I fry my KryoFlux?

Post by brightcaster »

I wouldn't say that all SFD-321B are risky, but the rev WT-04 certainly is! I once got a bulk from ebay (5 pieces because they where cheap and there was an easy modification to use them in the amiga) of which 3 went dead while copying or formatting a disk (intense stepper use)! On one of these boards the controller chip had a real hole in the black plastic after that. One of the two working drives had the diodes and the other one had a toshiba controller instead of the semco the others had.

So I guess that there was a batch of the controller chip which wasn't robust enough. So the people at samsung added the diodes for that batch.

In my case I never destroyed the amigas I connected the drives to. But as far as I remember the SFD-321B is is one of the rare drives which also uses the 12V! If they are used for the stepper driver it could well be possible that they find there way to the datalines, when the controller gets burst!

But there are so many different revisions to the SFD321B that some might work well for a long time.

BTW: The later SFD321B even don't have the DD/HD switch to recognize the notch and therefore don't read DD floppies well (as they are always in HD mode)!

So I would wait for the nec and give it a try. The samsung is excpected to be dead. As long as the LEDs on the kryoflux are flashing the cpu is working. If you're lucky everything will work again. If not the IC3 has to be replaced....


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 »

Good news! The NEC FD1231M arrived today. Probed the relevant pins and it seemed to be all within acceptable ranges.

I was able to succesfully get the KryoFlux to interface and read data off floppy disks! Hopefully writing will work too.

Thank you again David for all the help you've provided.
/* --- Text data --- */
brightcaster
Posts: 316
Joined: Fri Nov 08, 2013 10:48 pm

Re: Oh dear, did I fry my KryoFlux?

Post by brightcaster »

That's good news! Congrats!

David
User avatar
mr.vince
Posts: 2142
Joined: Tue Oct 05, 2010 5:48 pm

Re: Oh dear, did I fry my KryoFlux?

Post by mr.vince »

No, something must have gone wrong. We have used dozens of these drives. If the driver IC gets hot, it's likely it's burned. No part of the device should get hot, it was designed in a way it does not need cooling etc.

It's very clearly stated in the manual (I believe in bold red letters) that routing power through the KF board is a possibility but you need to know what you're doing and because of the nature of it this will void the warranty.

The repair, if you want to do it locally, is pretty much straightforward. Get a replacement IC (probably a Euro or less), find someone with a hot air gun (desoldering equipment) and have it swapped.
User avatar
Ash
Posts: 10
Joined: Thu Jun 04, 2020 5:48 pm

Re: Oh dear, did I fry my KryoFlux?

Post by Ash »

mr.vince wrote: Fri Jun 12, 2020 11:14 am No, something must have gone wrong. We have used dozens of these drives.
This is more curiousity than anything, does that include the models without the diodes?
mr.vince wrote: Fri Jun 12, 2020 11:14 amIf the driver IC gets hot, it's likely it's burned. No part of the device should get hot, it was designed in a way it does not need cooling etc.
Thank you for confirming, it actually works fine now, but I am using a different drive entirely. The first drive seems to be unusable now.
mr.vince wrote: Fri Jun 12, 2020 11:14 amIt's very clearly stated in the manual (I believe in bold red letters) that routing power through the KF board is a possibility but you need to know what you're doing and because of the nature of it this will void the warranty.
Please do not take this the wrong way, I only aim to help improve things, not to demand replacements which probably come at a cost to you guys. I do not dispute that the fault lies with me and have voided warranty, I suspected the drive itself was to blame. But if it is just from connecting the shipped power supply and Molex connector, then please read further.

If that is the case and you mentioned it is clearly stated in the manual, I feel that the manuals need updating regardless. It genuinely is not noticeably clear in the manual and I will break down the reasons why this is very murky to me:

In the quick start manual:
Warning: Do not connect or disconnect anything on your own unless you have read the main product manual as well. A false connection can damage the product and will void your warranty!
So, basically, go read the main manual and if you follow it, you should be golden… So, in the main manual…
KryoFlux is a fully bus powered device. Therefore no external power is needed. For special purposes the board allows to be powered externally. It is even possible to distribute power to an attached device (see 5.). Please note that the power rail for +5V is directly connected to the device’s CPU. A bad (cheap, unreliable, broken) power supply can damage your board as well as external devices. The PSU must deliver a minimum of 1A per power rail (+5V/+12V). Check the orientation before attaching the plug.
None of this suggests that having the power supply connected and the USB connected to the KryoFlux at the same time is actually going to lead to a problem, just that it is not necessary. It suggests that having a bad power supply would be an issue, I am using the official power supply which I also tested as can be seen in a previous post that showed that it was running correctly.

Perhaps it should have a mention that having a faulty drive connected to the same power connector can lead to something bad happening on that power supply when it is connected (but considering the drive made the chip heat up when the Molex wasn’t connected to the board, I don’t think this was the issue in my case)?

The section in red does not make it clear that merely connecting the cable voids the warranty or can cause a problem, but the placement of sentence suggests that these are connected sentences:
Incorrect orientation of the cable will DESTROY your KryoFlux board and/or your drive. You will also void your warranty (prebuilt boards).
If the second sentence was meant to be a general one, and I suspect it is not because it is repeated in other scenarios in the manual immediately after a particular issue, I would suggest that if voiding the warranty occurs from just connecting it to the power supply, then this should be stated separately and clearly in its’ own paragraph.

I have read the manual over and over and it was never really clear to me if the MOLEX connector can be used to power the board, the manual is not clear on this either. It suggests "For special purposes the board allows to be powered externally" in the very heading under "MOLEX", however I am sceptical it was referring to the MOLEX connector, even though it was in that section, but I only came to this conclusion after mulling over this for a while now.

It might be worth considering finding a way to clarify visually that KryoFlux to a powered MOLEX connector is a "no no" from a hardware experience point of view, people like to naturally just connect things when they see relevant connectors, and I noticed that AdamK did the same mistake in his thread where he posts his video, particularly at this timestamp (and mentions earlier that the connector is where power goes).

In summary, I still view myself still at fault because I did something the quickstart did not say I should do, but the other manual does not say that is actually even an issue, or if it does, it was not clear to me and it really needs updating.
/* --- Text data --- */
User avatar
mr.vince
Posts: 2142
Joined: Tue Oct 05, 2010 5:48 pm

Re: Oh dear, did I fry my KryoFlux?

Post by mr.vince »

Thanks for feedback.

I really think it's explained in detail. Quickstart shows how to set it up and clearly warns to read the main manual before attaching anything differently.

On page 4 there are several key facts:

(3) MOLEX power connector: KryoFlux is a fully bus powered device. Therefore no external power is needed.
(7) USB B connector: KryoFlux connects to the computer with a USB A to USB B plug. The board (NOT an attached drive!) is solely powered through USB.

Plus all the warnings in red.

We've had people attach Apple drives to the JTAG because they thought it looked like it would fit.

Sorry, I can't check if the drive has a diode or not. It's just that I've personally have used quite a few and we sold maybe a few hundred of these with no complaints in this regard so...
User avatar
mr.vince
Posts: 2142
Joined: Tue Oct 05, 2010 5:48 pm

Re: Oh dear, did I fry my KryoFlux?

Post by mr.vince »

Just in case anyone wonders: These connectors were placed for power users who might have very special needs and are willing to sacrifice the board (if anything goes wrong) because of that special need. E.g. 8" drives might need special care power rail - wise. Keep in mind you are using steampunk technology with a modern computer. 8" drives tend to use mains motors (120-240V) and you really don't want anything to go wrong here.
Post Reply