BlueSCSI v2 - DaynaPORT WiFi!

[Paolo B]

So, today I tried to figure out why the modified BlueSCSI board is no longer working.
At first, I suspected the hand-made SCSI cable was the culprit, as the modification to the board was looking neat. In fact, I could identify one line that was maybe a little loose at the connector. But eventually the cable seems to be fine, I can connect and use a SCSI2SD (only other SCSI device I have at hand).
So, I’ve tried absolutely anything I could figure out with the BlueSCSI board / SD card, but to no avail.
Before modifying the board, I made a copy of all the files on the SD card. The HD images are deemed as OK by Disk Jokey but are not longer recognized.
I updated the firmware of the Pico, formatted many times the SD, tried with virgin images. Nothing, the SCSI bus just hangs.
So, I reverted the mods, resoldered the resistor in and removed the 4.7k Ohm ones.
The BuleSCSI seems to be bricked, the Portable doesn’t even get to the disk with the pulsating question mark.
If I connect an external SCSI device, it’s not recognized, so it seems as though something is totally messed up with the SCSI chain.
Is there any way for understanding a little more about what is going on?

 

[JDW]

If I connect an external SCSI device, it’s not recognized...

Meaning, you disconnected your bricked BlueSCSI from your Portable and connected only a known-good and fully bootable SCSI drive to your Portable externally (no other SCSI device attached other than that external), and your Portable still would not boot from it?

---

Also, can you confirm your Portable is still working by inserting a known-good boot floppy into the real floppy drive? (Or FloppyEMU externally?)

If you can get your Portable to boot from a floppy, I would recommend running MacTest to confirm all else is well with your Portable. Use version 2.0 (DL#3) for a 5126, or use 1.0.2 for the 5120:

https://macintoshgarden.org/apps/mactest-portable

I used DL#3 on my 5126 and can confirm it boots into 6.0.4 just fine and all the tests in MacTest work. (You will need a serial cable to do the SCC test, and an 800K & 1.44MB floppy to do all the floppy tests.)

 

[Paolo B]

Meaning, you disconnected your bricked BlueSCSI from your Portable and connected only a known-good and fully bootable SCSI drive to your Portable externally (no other SCSI device attached other than that external), and your Portable still would not boot from it?

No, it would. With SCSI2SD unit I have 4 devices ID=0,1,2,3 on the internal SCSI connector, plus a real MO drive (ID=5) on the external connector. Everything works. If I attempt the same with the BlueSCSI, neither the internal devices nor the external one are recognised. If I unplug the BlueSCSI, then I can boot from floppy or from the external MO drive.
Can't get if it is an hardware issue (besides soldering and desoldering the three resistors, I obviously haven't touched anything) or what else... Very annoying indeed...

 

[JDW]

@Paolo B
Thank you for clarifying. I am relieved that your Portable has not been affected.

I see from your earlier screenshot that you have a version of BSv2 older than mine.

YOURS:

MINE:

It is a fact that my 2023.03a version worked fine with the dual resistor mod. I actually did that mod before I even tested my BlueSCSI in my Portable. It just worked. Not sure if I was just lucky, or blessed, or both.

What is truly puzzling is the fact that @Androda said he did the same dual resistor mod on his 2022.12a BSv2 and had no issues. He did not mention if he has a PICO or PICO W installed, but it's clear that you and I both have the PICO W installed. If Androda also had a PICO W installed, then the only thing possibly different between your unit and his would be firmware version.

I realize you said you updated firmware on your PICO W. But I had a BSv2 in the past which I wanted to upgrade to WIFI. I purchased a PICO W, desoldered the regular PICO, soldered in the PICO W, then flashed the firmware, and I found it didn't work. After much hair pulling, I desoldered the PICO W (not too difficult because I have a Hakko FR-301), flashed the firmware to to the PICO W, then re-soldered the PICO W back to the BSv2, and I found it worked fine. So I am wondering if the same oddness that hit me with that particular BSv2 in the past is also now afflicting you.

 

[Paolo B]

So I am wondering if the same oddness that hit me with that particular BSv2 in the past is also now afflicting you.

In fact, I did the same as you did, I swapped the Pico with a Pico W, it just worked fine.
After making the mod for the sleep freeze and before testing the unit I upgraded the firmware to the latest one.
Today I even tried to downgrade it again. Fundamentally, I now have the same hardware, the same firmware, the same SD with the same ini and the same images, but they no longer work. The log file reads everything OK.
One afternoon down the drain...
Tomorrow I will try your suggestion: desolder the pico, re-patch the firmware, solder it back.

 

[JDW]

Tomorrow I will try your suggestion: desolder the pico, re-patch the firmware, solder it back.

Sorry to hear nothing else worked. Desoldering that PICO is always a last resort, even if you have a good tool like the FR-301 desoldering gun.

Honestly, until now, I've always wondered why I seemed to be the lone man on the planet with these oddball BlueSCSIv2 issues, mainly because I don't always read reports from other people who have the same issues I have. I went through my old PMs just now so I could remember my BSv2 issues more clearly.

The BSv2 I mentioned having had trouble with in the past is a revision 2022.12a, just like yours. It originally was acquired from Kay Koba and was a legitimate BSv2, not a fork. (My 34-pin BSv2 mentioned in this thread is a hardware fork of BSv2, made by Joakim Larsson. Not sold by him, but rather given to me freely only because he had extras.) My 2022.12a board had the normal PICO in it originally, but when the WIFI news hit, I wanted to swap it out for the PICO W. I spoke to Eric H. about that prior to my swap. He cautioned me against it because others had failed to successfully desolder the PICO.

As you may expect, (1) skill, (2) desoldering experience, and (3) not having the right desoldering tool play a huge role in one's success or failure of desoldering a PICO from a BSv2. I have the skill, general desoldering experience and the right tool (FR-301), so I was able to desolder the PICO cleanly, and then I soldered in the PICO W, and then flashed it. The flashing seemed to have worked fine (no errors); but as I said, it would not function when I later tested it. That ultimately led me to desolder the PICO W, flash it outside the BSv2 (which Eric told me was possible and no problem to do), and then I soldered it back into the BSv2, and then it started working! But why? Nobody knows!

So it is based on my own experience in having to desolder-and-flash the PICO W that I wrote my past experience for you. That's honestly all the only trick I can think of to assist you.

But even if desoldering your PICO W and flashing outside the BSv2 works (after you get it soldered back to the BSv2), the question still remains: what happened to cause all this mess in the first place?

Yes, we know it was the resistor mod that bricked your BSv2, but why? And that is the question I cannot answer.

Lastly, I would like to mention one more thing for the record. Pixel peepers who examine my forked 34-pin BSv2 resistor mod photo may be led to think I used the wrong resistor value, but that is in fact my stupid iPhone's fault. Modern iPhones most often produce brilliant photos; but in this case, it altered colors and instead of showing a red stripe, that stripe looks brown in the photo. I opened my Portable just now and used a magnifying glass to confirm that my resistor is indeed YEL, VIO, RED, GLD = 4.7kΩ. (VIO = violet = purple)

 

[Paolo B]

Thanks for your support @JDW, I appreciate it.
Today I squandered another hour and half for desoldering the PicoW, not the easiest of the jobs, as the fit is very tight and removing all the solder is very tricky.
Now I am twice as lost as I was before, though. The Pico W seems to be dead. There’s no visible damage, but definitely I cannot mount it on my Mac using the usb cable+pressing the “BOOTSEL” button. Nothing happens, not even the power led glows.
So, I took out of the drawer the original Pico: in this case as I connect it to the power feeding usb cable, the led on the board turns green. Similarly, though, I can’t mount it for flashing the rom either. Curious to note: after a few seconds feeding power via the usb cable, the board of the plain Pico becomes super hot. With the Pico W it gets mildly warm.
Still, when I plug (no solder, “a la socketing style” the plain Pico and try to use the BlueSCSI, it does not function. The two leds on the board turn on, the one on the Pico turns green, but the unit still hangs the scsi chain.

Now, beside the disappointment of going from a working unit (bar for the fastidious freeze after sleep) to a bricked one for unknown reasons, if this can help the developer community I can still make some tests if anyone has any suggestion. Else, I will throw it in the garbage and will source another unit. Sh*t sometimes happens, hopefully I will have better luck with the next purchase.

 

[JDW]

@Paolo B
I'm terribly sorry to hear that.

What baffles me is that this resistor mod is merely the addition of two 4.7kΩ "pull-up resistors." You are pulling the voltage level of those two pins on the PICO up to 5V through resistors. So even if the PICO side was completely ground, it wouldn't be a dead short to 5V due to the 4.7kΩ resistors being there. Indeed, that is their purpose — to prevent 5V and GND from shorting.

Based on your detailed explanation, it would appear that the Raspberry Pi chip was fried. There are only 3 things that I can think of which would have caused that:

1. You were rubbing a balloon on your head while scuffing stocking feet on a push rug while tapping your fingers on the pi chip. (My humorous way of suggesting an ESD spike from your body perhaps zapped the PI chip.)

2. Your soldering iron isn't ESD safe and an ESD spike hit the Pi when you soldered the Pi-side pins.

3. You measured each of those resistors after having soldered them in place. Believe me, I fried a chip on my LC575 board simply by measuring resistance, and that episode quite nearly fried my brain!

Now if none of the above 3 things happened, then I will need to defer to @eric and @Androda on this matter, as they have significantly more experience with designing the BlueSCSI devices than I do as a feeble end user who often needs lot of help myself.

Yet more mysterious are the two other things you mentioned:

1. Your old PICO, which worked fine before (which I assume means it was flushable in the past), now will not mount on your computer. You cannot flash it at all. That is a real mystery!!

2. When you connect the PICO to your BSv2 (I assume your PICO still has some older version of the BSv2 firmware on it), the BSv2 doesn't function, which indicates that something on the BSv2 board must have been fried as well. That too is a mystery because all this mod consists of is two pull-up resistors.

Again, I eagerly await hearing what @Androda and @eric think about this.

 

[Paolo B]

Of course I would be very curious to know what caused the issue.
Above all, I would like to know whether sourcing a new Pico W would suffice or if chances are that the board is also fried.
As for the root causes, I can only make speculations:
- ESD: a possibility, unlikely, but it is a possibility
- testing after soldering the resistors: yes, I did it, but after the first run failed, when the issue was already there, so to speak
- loose connections in the SCSI cable: unlikely, and in any case I wouldn’t know how such an event could damage the electronics

 

[JDW]

@Paolo B
Because @Androda and @eric area not replying, I strongly recommend sending them both a PM.

 

[Paolo B]

So, I concluded both Picos are dead.
Maybe the temperature for the desoldering tip was too high (I use 350 for most of the jobs, idea is to have a quick hot spot rather than persisting too long and heat up the entire nearby area), or maybe I inadvertently shorted with my own fingers some pins along the sides while it was connected to the power.
They both visually look very fine, but surely they are gone.
Anyhow, I will get a new Pico W and hope for the board to be still healthy (I will have to re-do the resistors thing, though...).
It would still not explain how any damage could have been caused when I made the simple resistors mod, but I don't want to waste too much time behind this small thing.

 

[Androda]

So, I concluded both Picos are dead.

I've been totally unavailable the last few days, but the previous post saying the Picos "get hot" when plugged in very strongly pointed to hardware issues on the Pico units.

The added 4.7k ohm resistors are unlikely to cause these hardware failures, as later BlueSCSI versions use exactly this configuration - just with SMD resistors on the base PCB.

It was more likely caused by swapping the picos, as that's a much larger soldering job with potential to damage the primary 3.3v regulator inductors given how close they are to the 3.3v output pin.

On the standard Pico which gets really hot, I would suggest checking the 3.3v regulator output. It's probably closer to 5v now - regulators tend to "fail short" and pass through the full input voltage. If you're in the USA and would like, I could perform a "post mortem" and try to determine exactly what failed. But these sticks are only $4 or $6 each (plus shipping) so the replacement cost is fairly minimal to begin with.

 

[Androda]

Because @Androda and @eric area not replying

Your repeated at-mentioning me and Eric on this issue does not increase its visibility - both of us are already watching this thread.

Paolo's comment asking for help with the non-functional BlueSCSI was two days ago. Not a week, not a month, two days. Given that people have responsibilities outside the retro hardware world, I don't think a two day delay is unreasonable.

 

[Paolo B]

So, first of all, my sincere gratitude and appreciation to @JDW for his support, we’re on the same time zone, we’re both enjoying some days off (golden week), so our exchanges can be quick.
And of course thanks to @Androda for giving his feedback and sharing his expertise on this unfortunate case.

That said, yes, good guess, the voltage regulator seems to be gone in the baseline Pico (the one that gets super hot): I read 1.1V at 3V3_EN and 4.3V at 3V3. So, busted.

On the Pico W (the one that doesn’t lit up), I read 3.3 V sharp on the 3V3 pin, and 2.9V on the second pin (don’t know if it’s correct or not, though, I would need to read through the documentati on).

My best guess for the busted one: it’s a possibility that I accidentally connected the power connector rotated by 180 deg, de facto swapping 5V with 12 V. The connector has no poka-yoke (suggestion for improvement for the next spec update). Unlikely, but I don’t have any other explanation.

As for the Pico W, indeed I have no clue. The soldering job was done neatly and with minimal impact, must have been something else.

 

[Paolo B]

So, today the Pico W arrived. After updating the firmware I slotted it in the BS board. Now, when I try to boot the Portable off the BS board, the Pico flashes 5 times indicating that there‘s no (valid?) SD card. The images on the card are the same as before this little nightmare begun and they are validated by Disk Jokey. Could it be that the BS board also got damaged as consequence of the busted voltage regulator on Pico?

 

[JDW]

Could it be that the BS board also got damaged as consequence of the busted voltage regulator on Pico?

That is a question I unfortunately cannot answer, but I am still wracking my brains over how the addition of 2 resistors could cause such a catastrophe.

I found your earlier post where you mentioned the "possibility" of ESD, but there's always the possibility for that. I tend to be rather impervious against that. In all my years of dealing with electronics, I've not zapped something with ESD. I've zapped many a board with other things, but not ESD.

You said earlier that you might have "accidentally connected the power connector rotated by 180 deg, de facto swapping 5V with 12 V." That connector sockets into the BSv2 host board and flows through the board before touching the PICO. If you did inject 12V into a 5V line, that certainly would have fried something. But the question is, did you? Even so, you have two BSv2 boards, both which don't work, right?

 

[Paolo B]

I also don’t think it’s an ESD issue.
As for the 5-12V accidental swap, I also don’t think I did it, but - objectively - I noticed that it is *possible* to fit the power connector upside down, as there’s no connector casing, so to speak, just the four prongs, which of course are not offering any kind of “poka-yoke” feature.
By the way, to my humble opinion a more traditional “HD molex” full size type would have been more practical, considering the fact that the board is surely meant to be used on vintage hardware, as a replacement for the aging spinning hard drives.
Anyhow, I will need to check the power lines routes for understanding if any damage could have occurred as a consequence of hypothetical accidental voltage swap and/or damage to the voltage regulator on the pico.
But hopefully the designers will have a prompt feedback to this doubt and will be able to suggest some quick test on the hardware.

 

[JDW]

I really wish I had the quick answer for you, @Paolo B. I know you're just like me, hoping for quick answers because it's Golden Week here in Japan and you'd prefer to quickly get this issue resolved so you can move on to more productive tasks such as actually using your drive!

The problem with feeding 12V into a 5V line (not saying you did that, mind you), is that it doesn't always result in a burned component that you could see with a magnifying glass. Such over-voltages can zap ICs but not physically show you the chip is zapped.

But let's give you the benefit of the doubt and rule out ESD or reversing the connector on that power harness. We know it cannot be the physical presence of those two resistors because they are there to prevent shorts. Is your soldering iron "ESD safe" such that it would be impossible for the iron to inject any sort of spike into the circuit? That's the only thing I can think of because when you soldered those two resistors, one side of the resistor touches the PICO, but the other side touches the BSv2 board's 5V line. So if there was an over-voltage on that 5V line, any chip on the main BSv2 board which is powered by that line would be hit by the spike.

I doubt it's that because in my years of using soldering irons, I've never experienced anything like that. I am just pulling hair, on your behalf, trying to make sense of what you have described for us. I know you must be frustrated to death!

 

[Androda]

the Pico flashes 5 times indicating that there‘s no (valid?) SD card

5 flashes means the Pico is unable to connect to the SD card, as mentioned in the troubleshooting document.

One important note is that the SD card is powered from the Pico's 3.3v regulator. So when your standard Pico's regulator started generating over 4 volts, the SD card may have been damaged. I would suggest trying to set up another card to see if that helps resolve anything.

it is *possible* to fit the power connector upside down

Similar to the original full size hard drive connector, the floppy power connector is polarized. I just tried to plug in a standard floppy power cable upside-down and it only inserts about 1/4 the normal distance before it becomes very hard to keep going (fingers are sore).

BlueSCSI also includes a handy silkscreen showing where the red (5v) and yellow (12v) wires of a normal connector normally go. The risk of incorrect installation when using the standard connector and cable appears to be minimal.

As an aside, it might make sense to create a different thread for your debugging - we're pretty far afield of the DaynaPort WiFi subject.