BlueSCSI on Macintosh Portable

[Paolo B]

So, seems there’s light at the end of the tunnel…

First off, if I use the same SCSI device on the external bus, everything works. So, it must be about the way the Portable is accessing the internal bus, which is likely managed by the software… and, bingo! When I use System 6.0.8, everything works fine, even if the device is mounted on the internal bus.

Will now try a fresh new installation of System 7.1.

 

[SuperSVGA]

The Conner drive was specifically designed to be a SCSI 0 ID device, somehow hard locked.

Technically you can change it, there is a block of four jumpers on the drive's PCB and the first three are the SCSI ID.

 

[Paolo B]

So, today I spent some time testing step by step different incremental configurations. System 6, System 7, with various HD drivers. Everything works fine, until the moment I add in to the picture the Wi-Fi thing. At that point the crash happens again (I guess the same situation already reported by @JDW).

 

[JDW]

At that point the crash happens again (I guess the same situation already reported by @JDW).

I confirmed that the "always powered" solution proposed by @Androda works. No more freezes upon wake from sleep if WIFI was active when the machine went to Sleep, but I am not using it at the moment because it requires a male jumper pin to be inserted into one of the holes in the PDS connector, and should that male pin ever slip out, you will have voltage on that wire which could short on something. I suppose I could add hot glue on it to lessen the likelihood of that happening. I don't know. @Paolo B , you might want to give it a try (if you have a diode), see how it works, then let us know your thoughts on it. Since I don't have a battery and use the Portable Battery Eliminator as a result, I don't need to worry about the BSv2 being powered even during sleep. There is no battery to drain on my Portable.

 

[Paolo B]

I confirmed that the "always powered" solution proposed by @Androda works. No more freezes upon wake from sleep if WIFI was active when the machine went to Sleep, but I am not using it at the moment because it requires a male jumper pin to be inserted into one of the holes in the PDS connector, and should that male pin ever slip out, you will have voltage on that wire which could short on something. I suppose I could add hot glue on it to lessen the likelihood of that happening. I don't know. @Paolo B , you might want to give it a try (if you have a diode), see how it works, then let us know your thoughts on it. Since I don't have a battery and use the Portable Battery Eliminator as a result, I don't need to worry about the BSv2 being powered even during sleep. There is no battery to drain on my Portable.

I see we're eventually converging, I also have collected evidence that an "always on" solution works (I used a bench PSU).
Bar for the busted pico which is still a complete mystery to me, in the end we expereince the same issues.
So, I will consider the proposed solution, however I only have PSU grade diodes at hand, I need to source something suitable. Any recommendation would be very appreciated, I am not familiar with the selection of the right specs for this kind of component. Thanks.
Surely, though, the idea of having the led always lit is not so appealing, but this could be easily addressed.

 

[JDW]

Not sure what a "PSU grade" diode is, so if you know the part number (often printed on the diode body), that would help identify what you have.

I believe that any diode other than a Zener and other than a 1N4148 (which don't come in 1A varieties) should suffice for a basic test. Silicon diodes rated for 1A like the 1N4004 variety will have a larger voltage drop than a Schottky, but it still powered the BSv2 in my tests. I posted the voltages here.

All diodes have a voltage drop across them that varies by the amount of current that flows through them. According to Androda's BSv2 page, current draw can reach 750mA. At that level, a Schottky diode rated for 1A makes more sense than a silicon diode, because the reduced voltage drop of the Schottky would be more apparent.

 

[Paolo B]

Not sure what a "PSU grade" diode is

Yeah, not a very technical description, I must admit… I just go by heart, I must still have somewhere some diodes I bought for refurbishing a busted PSU rectifier. So, heavy duty, chunky ones.

By the indications you shared, I should be able to source an appropriate component, hopefully on Yahoo auctions for avoiding all the complications with Mouser et al. when you buy small quantities.
It will take some time, but I will surely share the outcome.

 

[JDW]

If by "chunky" you means something like this...



...well, those will work too!

They are normal silicon diodes like a 1N4004, but they are chunky boys because they handle 6 Amps! For long term use, their size may be a bother, but such diodes will work for testing purposes just fine, if you have something like that.

 

[Paolo B]

So, I have tried the proposed hack (tapping +5V from PDS position A2 for feeding the BlueSCSI board and preventing the Wi-Fi crash), but strange (and scary!) things happened.
At first, I just wanted to monitor the status of the A2 pin in different conditions (off, sleep, battery, PSU etc). So, I connected the two probes of the multimeter. Red probe on A2 (+5V) and black on B1 (Ground). Even with the 9V battery alone, I could read 5 V. So far so good.

However, as I powered up the Portable, it started crashing showing a garbled screen and throwing one error (sad Mac) after the other.
Luckily, after removing the probes, everything went back to normal. Whatever happened, I prefer not to push my luck.

All in all, I think it’s a software thing that causes the crash, as the original DynaPort was an externally powered device and probably the working logic of the driver had been written based on this assumption.
Besides, I am afraid that an always-on solution would likely quickly drain the 9V battery, even with the machine off.

 

[JDW]

@Paolo B
So it would seem you have your meter set properly to measure Volts (not Amps), and the fact you could read 5V indicates you have the RED probe inserted into the proper jack in your meter to measure Volts (not Amps). So I am rather bewildered by why your meter, measuring only volts, would cause enough power draw on your PDS slot to result in crashes. You seem to indicate there are no crashes whatsoever with the meter totally removed. What is the make and model of your meter?

Speaking of the BSv2 always-powered situation draining your 9V battery, I am quite curious how fast it drains even without BSv2 connected at all. Do you have past experience in knowing that?

The Macintosh Portable Battery Eliminator by @Androda is the only solution to prevent battery drain, but it has the following two caveats:

1. Incompatible with software that only works when you have the AC Adapter/Charger plugged in because the Eliminator acts like a never draining battery that isn't being charged.

2. There is no 9V battery replacement and therefore PRAM is reset every time you remove power from the wall socket.

However, the benefits are that you never need another lead acid or 9V battery again. You won't be able to take your Portable to Starbucks and impress people like mad, but most of us don't do that anyway, so it's not a major loss. So far, I've been living with the two aforementioned caveats. The only one of the two which pains me is the lack of PRAM retention. There are a number of settings I need to setup, aside from the Date/Time, every time I power on the machine. It irks me if I don't. But there's nothing I can do about that, so I live with it.

 

[Paolo B]

@JDW
The meter is a Fluke, don’t remember by heart the model name, but it’s a recent one. By default, it works in auto range mode, so maybe this caused some fluctuation on the drained current which disturbed the CPU operations. However, I must admit it was very strange as the voltage rail should be decoupled from the processor.

As for the 9V battery, I am not using the machine on daily basis, I normally remove the big battery when I store it away, but I leave the 9V connected. And it lasts many months, for sure, I don‘t remember the last time I replaced it.

Anyhow, I think I will also accept that the Wi-Fi may lead to a system crash, it’s not that I need to be constantly connected. I just now have to find a way to easily enable / disable the Wi-Fi operations by enabling / disabling the corresponding software bits without having to re-enter the configuration parameters anew every time. Any suggestion is welcome.

 

[JDW]

Well, because you have a Fluke, I doubt Auto-Range had anything to do with it. Perhaps @Androda might have some good speculation, but I myself really have no idea why the mere presence of that meter (measuring only voltage, no less) would trigger one Sad Mac after another. Very odd indeed.

As to answering your other question about how to more easily enable/disable WIFI, I need to wait until after my move this week. Everything is in boxes, and except for a for replies here and there (like this one), my mind is bit preoccupied with other matters. But it is a good question because avoiding that crash is important, I think.

 

[Paolo B]

Everything is in boxes, and except for a for replies here and there (like this one), my mind is bit preoccupied with other matters.

I wish you a smooth relocation, all the rest is now unimportant.

 

[Androda]

Looking at the schematic, PDS pins 2 through 4 on columns A, B, and C are all the main 5v source. There aren't any control signals mixed in over there to cause problems. This gives me one idea for why it might cause problems: accidental inversion of the intended pin.

If the measurement was done backward accidentally and thus on A30 instead of A2, that could cause issues. A30 is the "bus error" indicator pin, so if the connection was shaky at all even in voltage mode it might have been enough to make the system think there were loads of errors happening.

As for the DaynaPort software being "turned off" or "disconnected" somehow to prevent the system wake crash, I don't have any information on how to do that. The original Dayna SCSI to Ethernet hardware was always powered on via external adapters, and we may have to replicate that with an always-on power source for BlueSCSI to function properly through sleep mode.

 

[Paolo B]

I measured on the second line of pins starting from the rear of the machine. The rows are marked on the board, from 1 to 32, so I should have picked the right one.
I was very surprised to see such a messy situation triggered by a simple voltage measurement, and I got very scared as at first the machine refused to boot, displaying only the cursor without going any further.
I disconnect everything (9 V, main battery, power supply) and let it rest for some time, after that it was again OK.
Honestly can’t imagine any explanation for that, maybe some short of some kind (where, how, caused by what?), the board was recapped many years ago with tantalums, it still looks perfect.

Anyhow, I guess we should come to pacts with the idea that the way the whole thing works is perfectly consistent with the design of the device it replicates. And that without having the possibility to tweak the legacy software, this “bug” is there to stay. Honestly, it’s not a show stopper, just a small nuisance. Once you know it, you can live with it.