upgrading a dc7800 to a dc7900

[faithie999]

this is for peter, since i believe he is the pioneer with the dc7900.

i read your msg about the superiority of the 7900 over the 7800 with some envy!

not wanting to ditch the 7800 i have, i searched ebay for a 7900 mobo.

here is one that may be for the CMT, since it doesn't say SFF in the listing.

http://www.ebay.com/itm/HP-Compaq-dc7900-CMT-Mini-Tower-462431-001-Motherboard-460964-000-LGA775-/140815579700?pt=Motherboards&hash=item20c9433a34

from the pictures in the ebay listing, would you say it will fit in the 7800 case?  also, not there is no heat sink included.  since you have both the 7800 and 7900 to compare, would you say the heatsink from the 7800 mobo would fit on your 7900 mobo?

thanks!!

ken

[pete...@cruzio.com]

> this is for peter, since i believe he is the pioneer with the dc7900.
>
> i read your msg about the superiority of the 7900 over the 7800 with some
> envy!

A Q45/ICH10 will always be superior to a Q35/ICH9, in general terms, but
it is really the procs and the RAM which are determining the performance
not the chip sets.

Now, should the Q45/ICH10 BIOS be superior in some fundamental way to the
Q35/ICH9, as is, on the surface, true where ACHI is concerned, without
there being a "solution", then there is no case for the Q35/ICH9. However,
we DO have a solution for the Q35/ICH9, so there is indeed a case case for
the Q35/ICH9.

hp/Compaq's documentation for the DC7900 strongly implies that
motherboards in the SFF and the CMD are the same. They are not. The SFF
has a 4-slot mobo whereas the CMT has a 7-slot mobo with six slots being
usable and the seventh being reserved for an eSATA cable kit.

Just as with the DC7800 where the SFF has a 4-slot mobo and the CMT has a
7-slot mobo with six sots being usable and the seventh being reserved for
a LPT cable kit, or some such nonsense.

> not wanting to ditch the 7800 i have, i searched ebay for a 7900 mobo.
>
> here is one that may be for the CMT, since it doesn't say SFF in the
> listing.
>
> http://www.ebay.com/itm/HP-Compaq-dc7900-CMT-Mini-Tower-462431-001-Motherboard-460964-000-LGA775-/140815579700?pt=Motherboards&hash=item20c9433a34
>
> from the pictures in the ebay listing, would you say it will fit in the
> 7800 case? also, not there is no heat sink included. since you have both
> the 7800 and 7900 to compare, would you say the heatsink from the 7800
> mobo would fit on your 7900 mobo?

Won't fit the DC7800 SFF case, period. But, WILL fit the DC7800 CMT case.


Knowing what I know now, I would always go with a DC7900 over a DC7800.

However, the DC7800 experience was invaluable as learning and teaching
exercises, as we now have a possible solution for ACHI-less motherboards,
where OS X is concerned.

The various OSx86 Guides of years past ALWAYS taught that the SATA ports
should be set to ACHI-mode for OS X. Now we know why, don't we?

Sure, generic mobos, such as Gigabytes and others, often provide true IDE
device trees (standard with ICH7, optional using a "bridge" chip with the
later ICHes), and if push comes to shove we could always run our optical
and hard drives off of the IDE bus, as I have reluctantly done a few time.

But, the DC7800 and DC7900 provide NO IDE bus, so SATA is the only bus
available, unless one uses USB which, some may recall, I actually did for
a while on the DC7800, until the ins-and-outs of the "magic" IOATAFamily
kext was identified.

I believe the SFF and the CMT may be using the same cooler. In any case,
even with, say, a Gigabyte, the cooler is always a separate deal.

[faithie999]

thanks, peter.

what i now have is the 7800 CMT, and you have confirmed that the 7900 CMT mobo will fit in the 7800 CMT case.

once i put the ebay 7900 CMT mobo into my 7800 case, i plan on moving the proc and heatsink/fan from my current 7800 CMT mobo to the ebay 7900 CMT mobo.

i'll let you know how i make out.

ken

[pete...@cruzio.com]

> what i now have is the 7800 CMT, and you have confirmed that the 7900 CMT
> mobo will fit in the 7800 CMT case.

I misunderstood. I thought you had a 7800 SFF, and I wanted to clarify and
underscore what I believe is a significant error in the 7900
documentation.

> once i put the ebay 7900 CMT mobo into my 7800 case, i plan on moving the
> proc and heatsink/fan from my current 7800 CMT mobo to the ebay 7900 CMT
> mobo.

Should be OK.

The other differences include:

1) one PCI-e 1x slot is replaced by a PCI-e 4x slot in a PCI-e 16x form
factor, most probably intended for a RAID card.

> i'll let you know how i make out.

Please do.

[pete...@cruzio.com]

I found another DC7900 CMT mobo, one with "free" shipping.

I am, thereby, going to "upgrade" (if that's the correct word) my DC7800
CMT to a DC7900 CMT.

I have been busy eliminating some of the apparent anomalies in the DC7900
DSDT, such as variables which are named _Sxx, but which are NOT intended
to be "reserved words", which the initial "_" is ALWAYS supposed to
signify.

Basically, I have renamed all of these SxxZ [ * ] .

Also, the EUS1 and EUS2 have been renamed to the correct (for AAPL) EHC1
and EHC2, and all the USBxes have been renamed UCHx (also for AAPL).

There remains some small strangeness in the SATAs, but these are not
causing any problems at the present time.

Basically, the SATAs are named SLT0, SLT1, PRT2, PRT3 and PRT4, and, just
for consistency, these should be named PRT0, PRT1, PRT2, PRT3 and PRT4.

I MAY get around to those in the next revision.

USB 3.0 WAS working on the DC7800. It is not now working on the DC7900
although the driver is loading and all other indications are favorable. I
may have to check the installation of the four port card and possibly
re-seat it.


[ * ] For the uninitiated, all Class I railroad "reporting marks" may be
any combination of one to four letters, EXCEPTING the final letter MAY NOT
BE "U" (these are reserved for so-called "containers"), "X" (these are
reserved for non-Class I railroads, basically a business which owns a
railroad which DOES NOT directly interface with the Class Is ... a
"leased" locomotive might have a reporting mark such as "EMDX", meaning
Electro-Motive Corporation, Leased) [ ** ] and "Z" (I forget what this
signifies). So, I am using the very LAST letter in the Roman alphabet, Z,
for these purposes.

[ ** ] This created a big problem for The Chessie System, which had
selected CSX for its abbreviated name. Hence, why CSX equipment is ALL
"lettered" CSXT and not CSX, as CSX would be a disallowed reporting mark,
as it signifies a private owner, and NOT a Class I railroad.

[pete...@cruzio.com]

Rev. C has all the updates which I mentioned earlier.

However, Revs. A, B and C all work.

I have found a DC7900 CMT mobo for my DC7800.

The DC7800 experiment was interesting, and much was learned about mobos
which work fine on Windows (for which these machines were obviously
designed) but don't work on OS X due to SATA incompatibilities.

Besides which, the Q45/ICH10 is just a better design than the Q35/ICH9,
although there is nothing fundamentally wrong with a 35-series machine.
And, no SATA incompatibilities.

These DC7900 machines from Compuvest are really clean. They come with
brand new Chinese-imported, but HP-branded keyboards (Windows type), and a
compatible mouse (generic, Microsoft-type, not HP-branded). USB in both
cases, NOT PS/2 as are originally supplied (and are still supplied by some
sellers of these machines).

Shipping is fast, and the packaging is extra strong, too, with reinforced
corners and heat-sealed packing straps.

If you can get a $159 deal with free shipping, this is surely a very low
cost way to get a 45-series machine. Mine came with a C2D E8400 (3.0 GHz,
6 MB cache), but there are a few E8500 and E8600 machines out there (but
not necessarily from Compuvest, and surely not for $159).

Oh, the machine comes with Windows Vista Business OEM edition installed on
an 80 GB hard drive. The certificate is good, too, and it will allow you
to register your machine.

The RAM will be HP-branded 667 MHz, two sticks, but not necessarily
installed in the correct slots.

On these HP mobos, the black-colored RAM slot must be filled first, then
the third, then the second and, finally, the fourth. The second, third and
fourth slots are white-colored.

If you repopulate the mobo with fast RAM, you can get the machine to run
pretty fast. The E8xxx series procs has a 1333 MHz FSB, and 800 MHz RAM
seems to work well in these. There are NO hardware tuning options in the
BIOS, so not any old thing can be made to work.

I am sending this from ML on the DC7900.

ML for $19.99 ... gotta love it!

[faithie999]

i think the Z reporting mark designates a trailer used in trailer-on-flatcar service (with the flatcar normally privately owned, and would itself have an X reporting mark).

[pete...@cruzio.com]

> i think the Z reporting mark designates a trailer used in
> trailer-on-flatcar service (with the flatcar normally privately owned, and
> would itself have an X reporting mark).

Sounds right.

Trailer Train Corp uses ?TTX, as in ATTX, BTTX, CTTX, for their co-owned
fleet of various types of flats.

The individual railroads own the auto racks themselves.

So, most probably U=container, X=private owner or lessor, Z=TOFC.

[pete...@cruzio.com]

> i think the Z reporting mark designates a trailer used in
> trailer-on-flatcar service (with the flatcar normally privately owned, and
> would itself have an X reporting mark).

Y is apparently NOT a reserved letter.

Union Pacific ran out of locomotive numbers (only four numerals are
allowed on locomotives) what with the acquisition of Western Pacific,
Missouri Pacific, Missouri-Kansas-Texas, Chicago and Northwestern,
Southern Pacific and Rio Grand.

So, UP petitioned AAR (or whomever issues reporting marks) for Y as a
final letter so it could reletter its "yard" locomotives UPY. Apparently,
it took years to obtain a variance.

In the meantime, there were UP, SP, DRGW and some other locomotives which
were relettered WP until the ranges of numbers could be freed-up.

Then, UP gave EMD a $1 Billion contract (actually, more like $2 billion)
for 1,000 SD70Ms.

To be numbered UP 4000-4999.

UP liked them so much that hundreds more were ordered and the numbers of
UP SD70Ms now range from the high 3000s to the mid 5000s.

If that wasn't enough, UP sent 24 SD60Ms to EMD for a Tier 2 upgrade with
the 16-645F engines being replaced with 12-710Gs and the horsepower
reduced from 3,800 to 3,150. UP 9900-9923, so far.

These so-called 9900s have proven to be very popular here in California as
these so-called "ECO" locomotives produce very low emissions.

Sorry for the off-topic material.

[mosslack]

On Aug 5, 2012, at 3:28 PM, pete...@cruzio.com wrote:

Sorry for the off-topic material.

I don't know about everyone else, but I love train talk myself. 8^)

Just a message from mosslack...

______________________________

Alt-OS <+> GG <+> TBIE <+> Hack List

[faithie999]

on other forums the list nanny would ban us.  glad things are a bit more relaxed here.

[pete...@cruzio.com]

> on other forums the list nanny would ban us. glad things are a bit more
> relaxed here.

We're actually pretty "cool" here.


A testing version of the Rev. D DSDT is working now on the DC7900.

Mainly, changes in the AHCI port numbers to be more "reasonable", like
PRT0, PRT1, PRT2, ..., PRT5, rather than SLT0, SLT1, PRT2, ..., PRT5.

Hopefully, the bare DC7900 CMT mobo will soon arrive in decent condition
and I can finally convert the DC7800 CMT to a DC7900.

This entire experiment with the hp/Compaq systems (DC7800 and DC7900) has
been fun, and it has produced a number of VERY AFFORDABLE, yet quite
high-performance systems.

I looked at the details of a DELL system which was along the same lines as
the DC7800 SFF, and it was not a lot different.

Must have had the very same mechanical engineers doing the physical design!

[mosslack]

On Aug 5, 2012, at 5:56 PM, faithie999 wrote:

on other forums the list nanny would ban us.  glad things are a bit more relaxed here.

We aim to please!

[pete...@cruzio.com]

The four-port NEC/Renesas-based USB 3.0 card (from HK for about US$12 with
"free" shipping) which was previously reported not to be working is now
working.

All I had to do was re-seat the card.

Working under ML right now using the hacked PXHCD driver.

[PH]

On Sunday, August 5, 2012 12:04:17 PM UTC-7, PH wrote:

> i think the Z reporting mark designates a trailer used in
> trailer-on-flatcar service (with the flatcar normally privately owned, and
> would itself have an X reporting mark).

Trailer Train Corp uses ?TTX, as in ATTX, BTTX, CTTX, for their co-owned
fleet of various types of flats.

Trailer Train is cooperatively owned by all the very large railroads on a pro-rata basis. I believe UP, as the largest Class I, followed very closely by BNSF, is the largest single owner.

TTX Corp ran out of ?TTXes, so it established a new set of reporting marks: TT?Xes, as in TTAX, TTBX, TTCX, etcetera.

[phubai]

I _finally_ decided to do this myself. Just ordered a dc7900 mb from ebay, and will convert my superbly running dc7800. Since I'm running the latest version of Peter's DSDT (Rev F), I think the only change I'll need to make would be to remove the ioatafamily.kext; correct? I'll compare the guides to be sure.

[mosslack]

AFAIK, the IOATAFamily.kext should remain as is. The additional kext contained within should not cause any problems on the later system. 

From the main system of mosslack...

______________________________

[pete...@cruzio.com]

> I _finally_ decided to do this myself. Just ordered a dc7900 mb from ebay,
> and will convert my superbly running dc7800. Since I'm running the latest
> version of Peter's DSDT (Rev F), I think the only change I'll need to make
> would be to remove the ioatafamily.kext; correct? I'll compare the guides
> to be sure.

The DC7800 SFF/CMT DSDT works on a DC7900 SFF/CMT, but I recommend going
with the DC7900 DSDT on a DC7900 or on a DC7800 which has been converted
to a DC7900.

Most DC7900 CMT mobos are bare boards and DO NOT come with the I/O shield,
which is a bit different. No matter, you may run the board without a
shield.

Converting a DC7800 CMT to a DC7900 CMT is a straight across parts swap,
although the location of some cables is a tad different, but not all that
different.

So, simply remove the DC7800 CMT mobo and place it on a flat surface next
to the bare DC7900 CMT mobo and move the parts a group at a time.

Again, the DC7800 DSDT WILL run on a DC7900, but there are enough
differences that I recommend updating the DSDT.

Removing the ioatafamily plug-in is optional, but I would do that as well.

So:

1) install the completed DC7900 mobo and boot the system (updating the
BIOS as required, beforehand),

2) if that is successful, update the DSDT,

3) if that is successful, remove the PXII2 plug-in (it is not needed on a
BIOS which has AHCI.

At this point, you have a DC7900 CMT, but not its I/O shield, and all
should be well.

[pete...@cruzio.com]

> AFAIK, the IOATAFamily.kext should remain as is. The additional kext
> contained within should not cause any problems on the later system.

The IOATAFamily would remain the same ... it is REQUIRED.

What is surplus is the PXII2 plug-in.

The PXII plug in is required; the PXII2 plug-in is required only if the
BIOS does not support AHCI.

It is not an error to have the PXII2 plug-in present on a DC7900 as it
would never be used, anyway.

The DC7800 DSDT (Rev. F) will work on a DC7900, but it is best to update
to a DC7900 DSDT (Rev. F).

I have tried to maintain consistency between DC7800 and DC7900 DSDTs. Like
Revs give like function.

[pete...@cruzio.com]

>> I _finally_ decided to do this myself. Just ordered a dc7900 mb from
>> ebay, and will convert my superbly running dc7800. Since I'm running the
>> latest version of Peter's DSDT (Rev F), I think the only change I'll
>> need to make would be to remove the ioatafamily.kext; correct? I'll
>> compare the guides to be sure.

See previous e-mails regarding software changes.

If you got the DC7900 CMT mobo for $41, you did much better than I did.

As getting a DC7800 to work as a Hack is perhaps easiest with the iATKOS L
distro, which includes the PXII2 plugin, I have gone with Lion for my
DC7800s.

As for the DC7900, I have gone with Mountain Lion, and I have not been
disappointed.

So, also, for my DC7800 CMT which was converted to a DC7900 CMT, except
that I kept that one on Lion, and it is presently stored ready for use,
but unused.

I am a great believer in the E8400 C2D proc and its HUGE 6 MB cache.

Also, I removed the 2 GB of 667 MHz RAM from the true DC7900 CMT and
installed it into the converted DC7900, thereby making that machine 4 GB
of 667 MHz RAM, and then I borrowed 4 GB of fast RAM from another machine
to repopulate the true DC7900.

Here is what my true DC7900 now appears to be:

Hardware Overview:

Model Name: Mac Pro
Model Identifier: MacPro3,1
Processor Name: Intel Core 2 Duo
Processor Speed: 2.99 GHz <--- Really, 3.0 GHz
Number of Processors: 1
Total Number of Cores: 2
L2 Cache: 6 MB
Memory: 4 GB
Bus Speed: 1.33 GHz
Boot ROM Version: MP31.006C.B05
SMC Version (system): 1.30f3
Serial Number (system): G8135543XYL
Hardware UUID: 1D8EC6D6-0B60-556A-8461-03A6C13EC67B

However, the system thinks the RAM is running at 800 MHz. Perhaps it is.

Anyway, I had specified AUTO, so the BIOS automagically made any
adjustments which it believed to be appropriate.

Incidentally, EVERY 8400GS video card which I have installed in this
machine has been recognized and has been usable.

The same cannot be said for some earlier machines.

[phubai]

Thanks Doug!

[phubai]

On Tuesday, November 27, 2012 1:27:35 PM UTC-5, PH wrote:

The DC7800 SFF/CMT DSDT works on a DC7900 SFF/CMT, but I recommend going
with the DC7900 DSDT on a DC7900 or on a DC7800 which has been converted
to a DC7900.

Most DC7900 CMT mobos are bare boards and DO NOT come with the I/O shield,
which is a bit different. No matter, you may run the board without a
shield.

Converting a DC7800 CMT to a DC7900 CMT is a straight across parts swap,
although the location of some cables is a tad different, but not all that
different.

So, simply remove the DC7800 CMT mobo and place it on a flat surface next
to the bare DC7900 CMT mobo and move the parts a group at a time.

Again, the DC7800 DSDT WILL run on a DC7900, but there are enough
differences that I recommend updating the DSDT.

Removing the ioatafamily plug-in is optional, but I would do that as well.

So:

1) install the completed DC7900 mobo and boot the system (updating the
BIOS as required, beforehand),

2) if that is successful, update the DSDT,

3) if that is successful, remove the PXII2 plug-in (it is not needed on a
BIOS which has AHCI.

At this point, you have a DC7900 CMT, but not its I/O shield, and all
should be well.

Thanks Peter! Once I get the MB installed and working, I'll let you know, and will request the DC7900 DSDT from you. Unless REV C is the latest DSDT for the DC7900? That's the version that is linked to your guide for the 7900.

I appreciate all of the help and info!

Bob

[phubai]

I've installed the new MB as it came in this afternoon. I'm having a problem I've not had before, but have heard of, and that is it's not recognizing my Macintosh USB keyboard when I enter into setup. Of course, that's where I was trying to go with the F10 key, but even if I don't touch the keyboard, it ends up in the same place, and nothing I hit at this point gets me any further:

The striped affect is simply the result of the blinds and sun! I really don't want to buy another keyboard, but if I have to, I will. I had access to all of my work computers previously, and would borrow from one of them if need be, but alas I'm retired and they won't let me back in the door.

Oh, my USB ports are powering up, as I've plugged other devices into them, changed ports for the keyboard, etc. I haven't cussed yet...am saving some favorites for later!

Bob

[Kris Tilford]

On Nov 30, 2012, at 1:31 PM, phubai wrote:

> I really don't want to buy another keyboard, but if I have to, I will.

I don't think this is a keyboard issue. It's likely a BIOS issue. I've
got one computer that requires an old PS2 keyboard to enter BIOS
setup, so if you have an old PS2 keyboard try that. PS2 keyboards
should be nearly free at places like Goodwill or such. On Mac
keyboards some keys have different functions. I've found that the
decimal "." key in a numerical keypad is often the same as the "Del"
key of a PC keyboard, and will enter BIOS setup on PCs that use the
"Del" key. If your keyboard has one of the "lock" keys (caps lock, num
lock, etc) inadvertently pressed this can effect things too.

In the photo there's something about PXE which is for network boot
over built-in ethernet. You'll want to disable this in BIOS once you
get into the BIOS setup and can change things. Right now it may be
expecting to boot from ethernet?

[phubai]

Interesting...I can't remember whether I was using this keyboard or one from work when I set up the DC7800, but either way, it was USB.

I lived on a boat for almost 10 years up until 2 years ago, and unloaded a lot of extras in order to get everything I actually needed on the boat. Keyboards didn't make the cut! Until recently, I only had notebooks.

Thanks for the tip...I'll see if I can find a PS2 keyboard. As you said, they should almost be giveaways.

Bob

[mosslack]

On Nov 30, 2012, at 3:19 PM, phubai wrote:

Interesting...I can't remember whether I was using this keyboard or one from work when I set up the DC7800, but either way, it was USB.

I lived on a boat for almost 10 years up until 2 years ago, and unloaded a lot of extras in order to get everything I actually needed on the boat. Keyboards didn't make the cut! Until recently, I only had notebooks.

Thanks for the tip...I'll see if I can find a PS2 keyboard. As you said, they should almost be giveaways.

Bob

Up until the time I moved, I literally had a truckload of them. Now I only have a few left, but I keep them around just for this reason. Not so often anymore, but there was a time that those certain systems only worked with PS2 when trying to use a Mac USB keyboard.

[faithie999]

i agree, this is a vexing problem.

i have a couple of hacks that require a PS2 keyboard if i want to access BIOS, but i have others that work fine in BIOS setup with USB kb's.

but, i think you're saying your usb keyboard won't work once your installer stick has booted and you are trying to begin the OS X install.

i have a gigabyte H55M-S2V that misbehaves in that way.  what i have found that different installer stick preparation methods use a different selection of E/E kexts and boot.plists.  when i first tried to install ML on this box, i made a unibeast USB stick which booted fine but alas, at the first setup screen, the KB wouldn't respond (also tried a PS2 kb, and it didn't work either).

i have a few other installer sticks that i have prepared for other hacks.  i tried them all in the gigabyte, just to see if  one of them would enable the KB to work at the installer screen.  lo and behold, an installer stick i prepared for an install of lion on my HP probook booted and the usb keyboard responded.

so, i went back to the unibeast ML installer stick and replaced its Extra folder with the Extra folder from the HP Lion installer stick, and it worked!

so, i would encourage you to find other installer stick prep methods (myhack, kakewalk, probably others) and see if any of them work for you.

another idea would be to copy the Extra folder from a working OS X install and put it on your installer stick in place of the one that's there.

good luck

ken

[pete...@cruzio.com]

> i agree, this is a vexing problem.
>
> i have a couple of hacks that require a PS2 keyboard if i want to access
> BIOS, but i have others that work fine in BIOS setup with USB kb's.

The BIOS examines the environment and if anything changed from the
immediately previous boot, you will get that menu.

The usual case is the number of attached drives has changed (deleting the
Windows drive, for example). Hit RETURN to accept the new configuration.

Thereafter, you may still go into the setup screen or to boot OS X.

The BIOS accepts Mac keyboards just fine.

The black "Pro" keyboard is favored over the white "Mini" keyboard.

There is something about the "Mini" keyboard which is just not right.

A Windows keyboard may also be used.

EXPECT to get, during -v, a message which says the machine could not
enumerate something within the USB system, possibly the "hub" which is
within the Apple Inc keyboard.

As with ANY Hack, you can get occasional KPs.

On my DC7900, I have gotten them because of Voodoo, because of the USB
system, and because of the video card.

Usually, this happens BEFORE Mountain Lion is at the Desktop, but it can
also happen immediately after ML is at the Desktop.

KPs which occur after minutes or even hours of operation are much less
frequent.

I have been running my DC7900 CMT as my primary machine since I first
installed ML on it, and it has proved to be exceptionally reliable.

A DC7800 CMT converted to a DC7900 CMT would be just as reliable.

[phubai]

Thanks for the responses...I have two hard drives and didn't change anything from the way they were in the DC7800. Everything is just as it was on the 7800; the only change I've made is the mobo itself. The screenshot I provided happens at boot regardless of what I try. Initially, I was hitting F10 during the boot process to try to get to setup. Once I got to the screen, I thought, success! Unfortunately, I can't get past that screen...it's like my keyboard isn't even there. I've tried booting a live linux distro from a DVD, and a Unibeast USB with the same result. Again, it's completely unresponsive once that screen appears.

I do indeed have the silver Mac USB keyboard rather that the black one unfortunately. I have/had somewhere, a USB to PS2 adapter that I'd like to find and try. If I can't find it, I may order one, and try to find a PS2 keyboard to have around. 

Could I have purchased a bad motherboard?

Bob

[Christian Wacker]

I've had weird successes and failures with Apple keyboards on non-Apple hardware. Do you have even a cheap $5 usb keyboard laying around to test with? (not that i'd suggest wasting money on any cheap keyboard from places like Walmart. Avoid ONN brand. Horrid.)
I'm also not sure how well an Apple keyboard would work with a USB to PS/2 dongle, as no Apple product used ps/2, and the keyboards have to be compatible with a converter to work with it.

---

From: phubai <rjb...@gmail.com>
Sent: Saturday, December 01, 2012 08:51
To: hq...@googlegroups.com
Subject: [HQ-A] Re: upgrading a dc7800 to a dc7900

--
To unsubscribe: hq-a+uns...@googlegroups.com
List options: http://groups.google.com/group/hq-a/subscribe
 
HQ-A homepage: http://hq-a.weebly.com/
Group Files Page: http://bit.ly/ia8E2l
 
 

[phubai]

Good point on the adapter...I just ordered an HP PS2 keyboard from Amazon for $4.97 and will just go over everything to make sure I have it right. I tried to boot with a DOS bootable USB as well, figuring I could at least get to a dos prompt, but I get the same screen. Until I can get into and setup the bios, I'm afraid I'm done. Hopefully the PS2 keyboard will let me get there.

Thanks!

[mosslack]

On Dec 1, 2012, at 9:51 AM, phubai wrote:

Could I have purchased a bad motherboard?

Bob

Absolutely possible, but unusual that it would only go that far. It obviously has been through the post with no problems so it's difficult to know whether the problem is as long as you have an unknown in the Apple keyboard. A fewf questions, do any of the other F keys work at the initial screen? Does one of the other hard drives have Windows installed on it? If so, have you tried booting it with the other drive disconnected? Is this the same keyboard you used with the DC7800 and did it work okay with it?

[phubai]

My thoughts as well Doug...I don't have Windows installed on either hard drive in this machine. I do however have another hard drive with a windows installation on it that I might try to see if it makes a difference.

None of the function keys, or function key combinations work at all. I've tried everything including arrow keys, enter, but it's not recognizing any input from the keyboard at all.

This is the same keyboard as was working on the DC7800, yes. The only difference is the mobo; I simply removed the 7800 mobo and installed the 7900 mobo. Didn't change anything at all hardware related.

I'll try the windows drive...it's on a 2.5 in drive and I'll just set it in there temporarily to see if it makes a difference at all, and will post when finished.

Thanks!

[mosslack]

On Dec 1, 2012, at 12:14 PM, phubai wrote:

This is the same keyboard as was working on the DC7800, yes. The only difference is the mobo; I simply removed the 7800 mobo and installed the 7900 mobo. Didn't change anything at all hardware related.

I'll try the windows drive...it's on a 2.5 in drive and I'll just set it in there temporarily to see if it makes a difference at all, and will post when finished.

Thanks!

You might also want to confirm the keyboards operation with other systems as well. Keyboards do go bad from time to time.

[phubai]

RE: Windows Drive

Same result...the system recognizes the drive just fine, and that appears on the same screen vs the other two drives I have in the case...but...I can't get any further than that. 

Thanks!

[pete...@cruzio.com]

> Could I have purchased a bad motherboard?

Yes, it is possible.

The USB system in the DC7800/DC7900 is particularly weak and may not be
providing the attached USB device with enough power.

On my true DC7900, the FIRST boot attempt always fails, and I have to
power-cycle the machine. The second and all subsequent boots always work.

-v gives me a cryptic message which says something about being unable to
"enumerate" a USB device which is attached.

I am using a USB KVA (keyboard/mouse-video-audio) switch between my
machines and my single keyboard/mouse, plus, sometimes, a bluetooth
wireless mouse.

Connecting a keyboard directly to another USB port, say, the front panel
USB port, will give me two keyboards, one of which always works.

This problem does not exist on any of my Shuttles, ASRocks or Gigabytes.

[Kris Tilford]

On Dec 1, 2012, at 11:22 AM, phubai wrote:

> Same result...the system recognizes the drive just fine, and that
> appears on the same screen vs the other two drives I have in the
> case...but...I can't get any further than that.

Did you mean the initial BIOS screen, or the initial bootloader
screen. The photo was BIOS screen.

If you're talking about BIOS, then this is a BIOS issue. In a chicken-
or-egg scenario, some BIOS have setting to enable USB at boot, so if
that's disabled in the BIOS, you'd need a PS2 keyboard to enable the
USB to work. Bad USB on the motherboard would probably show some
popped capacitors you could see on inspection.

***It appears LIKELY that you need a BIOS update. In the revision
history of the BIOS for the DC7900 it says: "Fixes an issue where some
USB SmartCard keyboards do not function prior to booting Windows." I'm
not sure what "SmartCard" is, but a mention of USB keyboards not
functioning is probably the same problem you have. I think the newest
revision is 1.26 Rev.A ? There's evidently some addition firmware
called "ME" (Management Engine) that's part of "AMT" (Intel Active
Management Technology), and can be upgraded via a flash AFTER
upgrading the BIOS. Perhaps this "ME" disables the USB so that in a
business environment the computer is fully secure? You'd probably need
to remove the CMOS battery and perhaps move a jumper on the
motherboard to reset such a security setting. See this for BIOS updates:
<http://h20000.www2.hp.com/bizsupport/TechSupport/ProductList.jsp?lang=en&cc=us&taskId=135&prodTypeId=12454&prodSeriesId=3785403
>

If you're talking about bootloader, this is a known issue with
Chameleon, and by extension Chimera, on some motherboards, and again,
it's related to the motherboard BIOS. In the Chameleon issue, you
should be able to enter BIOS setup using a USB keyboard, and change
BIOS settings and save such changes, but after this, when the
Chameleon screen comes up the keyboard is completely non-responsive,
and you'll be unable to select anything, and the boot will proceed
according to default setting for the bootloader.

[pete...@cruzio.com]

>> Same result...the system recognizes the drive just fine, and that
>> appears on the same screen vs the other two drives I have in the
>> case...but...I can't get any further than that.
>
> Did you mean the initial BIOS screen, or the initial bootloader
> screen. The photo was BIOS screen.

At least, get the BIOS setup properly.

Connect ONE keyboard to the front panel port and get into the setup screen.

Then, select the option to set all values to defaults.

Finally, save.

I suppose the same could be accomplished by clearing the CMOS and then
letting the first power-on thereafter force going into the setup screen.

There are options for turning off many things. I opt for turning off
legacy devices which I will never use, such as serial and/or parallel
ports, where possible.

For a DC7900 the default is IDE for the hard drives, and AHCI must be
selected in order to get that option, which OS X expects and requires.

On my DC7900 CMT, OS X lists the SATA devices as follows:

Unknown AHCI Standard Controller: <--- Obviously ICH10-D0 *

Vendor: Unknown <--- Obviously Intel
Product: AHCI Standard Controller <--- Good guess!
Link Speed: 3 Gigabit <--- Will negotiate to 1.5 for the optical drive
Description: AHCI Version 1.20 Supported

* I suppose it is possible to force Intel ICH10-D0 using a modification to
the DSDT, but I elected not to go that way, although I DID do so for the
gigabit Ethernet port.

[pete...@cruzio.com]

> Connect ONE keyboard to the front panel port and get into the setup
> screen.
>
> Then, select the option to set all values to defaults.
>
> Finally, save.

I checked the DC7900 BIOS with one of my other non-hp/Compaq machines and
the USB power definition is the same:

"AAPL,current-available",
0x05DC,
"AAPL,current-extra",
0x044C,
"AAPL,current-in-sleep",
0x09C4,
"device-id",
Buffer (0x04)
{
0x34, 0x3A, 0x00, 0x00
}

[phubai]

It is indeed a BIOS issue...I never get to the bootloader at all. I do have power on the USB ports, however, I'm not sure how much at this point. Once my ps2 keyboard arrives, I'll know a little more. I'll do an inspection of the board as you suggest. Nothing was obvious when I first unwrapped it, but I'll look a little closer to be sure.

Thanks!

[phubai]

I will try to do that before giving up on the mobo for sure. As soon as my ps2 keyboard comes in, I'll have a chance to go through it. I've tried all of the USB ports, and while they all seem to have power, I don't know how much. Regardless, switching to other USB ports with the keyboard had no affect whatsoever. Thanks for the settings!

Bob

[phubai]

PS2 kb came today which enabled me to update and setup the BIOS for the new mobo. Initial boot failed to black screen, so I removed my Graphics card, and booted to the default resolution without a problem. I ran the new multibeast to see how it did using that to install REV F of the DC7900 DSDT.aml (Thanks Peter!), booted into the system (still without a graphics card) and my USB wasn't working at all. Booted into the backup partition, installed the DSDT.aml manually, re-booted and I had USB. I did not use multibeast at all on the backup partition, only manually installing the DSDT. I shutdown, installed the graphics card, and booted into the system fine, graphics working. I've compared my main HD with the backup partition, and removed a couple of kexts, and verified that the correct DSDT is installed. I'm now rebuilding permissions and will see if I can boot into my main drive.

One thing that has happened, and it may be a setting I'm missing in BIOS, but, at boot, I go to Chameleon and there choose one of my drives and or partitions. At that point, Chameleon is not recognizing my USB keyboard yet; I must use the PS2 to select and boot the partition I want. After I boot into OSX, my PS2 keyboard no longer works, but my USB (at least on the backup partition) is working fine. I would like to use either/or but it seems when I added the PS2 enabling kexts from within multibeast, I somehow disabled my USB (thus explaining why I've chosen to delete the PS2 enabling kexts located on my main drive and install). I will probably try it again, once I have my main partition booting, but my main concern is the necessity for the PS2 keyboard at boot. That seems to be a BIOS setting that is not enabling USB at boot like the DC7800 did. 

UPDATE: I'm now in to my main partition and with graphics as well, but on boot, I again had to use the PS2 kb to work with Chameleon. After that, it no longer works, and the USB keyboard does.

Suggestions as to what I'm missing in BIOS? The only thing I did once I updated it was set date/time and set to defaults and then add AHCI. I didn't see anything else that appeared related. I found all of the USB ports enabled but that was under security in the BIOS settings.

Thanks!

[Kris Tilford]

On Dec 6, 2012, at 12:46 PM, phubai wrote:

> UPDATE: I'm now in to my main partition and with graphics as well,
> but on boot, I again had to use the PS2 kb to work with Chameleon.
> After that, it no longer works, and the USB keyboard does.

I've got a computer with nearly identical issues. I think the problem
is with Chameleon? On my computer I can enter BIOS setup or one-time
boot menu directly from either keyboard, but once Chameleon appears
the USB keyboard is non-responsive and the boot partition must be
selected with the PS2 keyboard. Since BIOS recognizes both keyboards
initially, I think this means it's a Chameleon issue?

[phubai]

That's interesting...BIOS isn't responding to my USB keyboard but, I do have a white Mac keyboard which may be a problem in itself. Perhaps a standard PC USB keyboard is more easily recognized. But I can't get into BIOS with it regardless and have to rely on the PS2 keyboard for that and my Chameleon selection. I haven't tried any of the other updated fixes for BIOS available on the HP support site...maybe it would do something. 

At least you can enter BIOS with your USB which I still can't. Of course that doesn't help you get past Chameleon. Chameleon worked fine for me and my USB keyboard with the DC7800 mobo. That suggests to me a change in the BIOS at least.

[pete...@cruzio.com]

> Suggestions as to what I'm missing in BIOS? The only thing I did once I
> updated it was set date/time and set to defaults and then add AHCI. I
> didn't see anything else that appeared related. I found all of the USB
> ports enabled but that was under security in the BIOS settings.

I've got no clue.

I only have USB keyboards and mice.

There are a few things which can cause USB to hang, but I have not seen
those in a very long time.

The DC7900 (and I would also assume the DC7800) can be intolerant of the
amount of current which a KVM + a KB and mouse wants, which is why,
sometimes, I use a second KB and insert that into the front panel, if only
to enter a CR or an F-key.

[phubai]

On Thursday, December 6, 2012 1:46:11 PM UTC-5, phubai wrote:

 

Suggestions as to what I'm missing in BIOS? The only thing I did once I updated it was set date/time and set to defaults and then add AHCI. I didn't see anything else that appeared related. I found all of the USB ports enabled but that was under security in the BIOS settings.

Thanks!

***Update***Everything is working fine with the exception of USB keyboard at BIOS. Once I select my drive/partition in Chameleon, the USB keyboard works fine. I still can't enter BIOS with the USB keyboard, SO...I installed PS2 enabling kexts with multibeast and will just use the PS2 keyboard for the time being. At least the PS2 keyboard works, regardless of whether I need to get into the BIOS, select an OS, or as I am now, post to a forum, and for $5.xx, it didn't hurt too much. Not a bad keyboard either; a new HP desktop standard issue (PS2).