Z800 Z820 Firmware revisions

[r1_rn22]

Hello community,

does someone of you know where to find previous BIOS versions now?

The FTP is no longer working, and the SP number is unknown to me

I am looking for 2 specific files: J63_0368.BIN and J63_0369.BIN

Thanks

[SDH]

Are you aware that the BIOS for a Z800 is different than the one for a Z820?

What are you trying to do? If it is to step up a very old BIOS in a Z820 we can help you. Addressing for the source SoftPaqs has changed and you need to know how to get those but they still are available. I posted on that a few weeks back here.

The prefix on the bin files tells what type of box they are for, and that does not change over time. It is the suffix that changes. Here's the latest for a Z800 and then the latest for a Z820:

7G5_0361

J63_0396

Since you state that two J63 bin files are what you need I'm assuming you truly have a need for the Z820... and then the question comes why did you mention Z800?

EDIT: I'd only upgrade BIOS from within BIOS... don't use the HP app included because that was written for older OS and you run a higher risk of bricking your motherboard by using the app from within a newer OS. 

[r1_rn22]

Thank you for coming back to me, SDH. I know the latest versions, but I want to figure out what went wrong on the upgrade. AFAIR, it was 3.65, 3.68 or 3.69 on the Z820 mainboard. I inspected the readouts, and new files with WinMerge, and could find lots of differences. Maybe I can merge the missing data (MAC adresses, UUID, Tags) from the corrupted file into a new BIOS file and upload that.

So, have you any idea where to find Z820 previous BIOS releases?

[SDH]

The ones I was referring to are very early... back from the 1.x era. Look up my recent post that gave the links to get those. As noted, the addressing is changed from the original document that HP released but I give in that post the links that work.

For more recent ones a few are listed in the HP access page if you look a bit there and go to the older versions links. Search for Z820 drivers. Filter for W7 64-bit. There are active links there if you look a bit. For some reason HP did not give active links for some of the OS links and does for others so you need to do a bit of work to find them. I don't know of any place that has all the older ones listed with their respective softpaqs but once you know how to search for them with the newer style of addressing that I gave a few weeks ago you likely could find them in the newer ftp servers if you also found their SoftPaq number.

Good luck on your project!

[SDH]

Found that link for you from a few weeks ago:

Solved: Re: Z620 Bios update for better support of RX 570 8GB - HP Support Community - 9026899

Note the difference in addressing from the old HP document and the new way that works. Plus, note the blocks of 500, and I figured out that this is how to recreate a link that will work if you only know the SoftPaq number. Same address ending with .exe gets you the actual download and if you change it to end with .html that will get you the HP ReadMe document instead.

Note that you may need to manually add the https at the beginning. At least here if I copy/paste over one that works the https is left off. Here's an example that is the right way:

https://ftp.hp.com/pub/softpaq/sp64501-65000/sp64701.html

That will get you the HP ReadMe. You can also change it to end with .cva and read that with notepad.

Edit: If you end up getting too frustrated a motherboard transplant is not too bad of an experience. Sounds like you're still missing a MAC and a UUID. I thought the MACs came from the two LOM chips.

[r1_rn22]

Thank you for your support. I want to do the following: The original Z820 BIOS was some version 3.6x - I could finally read out the old BIOS chip, the adapter socket had a weak solder, now I got the BIN image of the contents and compared it to the J63_0396.BIN, and found tons of differences, and a lot of empty spaces whilst they are filled with FF in the read out file.

Curious if its possible to merge both files and reflash a 3.6x. Its only 16MB of data, so a doable task.

The AMT MAC is now FF'ed, and the UUID has the last digits also FF'ed (this is where I filled the MAC of the DHCP on the 2nd LAN port). The first LAN port is deactive in Win10, and cant work (it has no valid MAC, and there is no way to fill it in).

I also figured out how to edit the system information: CTRL-A when you are in the BIOS.

So, to summarize it: I need the 3.6x BIOS binary, compare it to the readout file, merge, reflash. And then do the upgrade to 3.96 within the BIOS (I will never do this in Win10 again. Never, never, never ...).

Apart from the additional stress, its a good learning. As I have all the tools needed, and do hw engineering for a living, I have no problems unsoldering a TSOP16 and reading it out, in case the update fails again.

Btw, do you have any idea if its possible to dump the contents of the actual BIOS on the command line? Like readfl.exe > bios.bin ?

[r1_rn22]

FYI, I found the MAC.

it starts in the read out BIOS at location 00000FFF, and its different from the one I saw in the DHCP, so it should be the 2nd one that I miss. The value differs 1 at the last digit. Theres also space for that in alle other compared BIOS files. So at least, this should be fixed. I will try it tomorrow during daytime. I guess this

Interesting, that the 3.6x BIOS version did not update this part: J63 v01.02 03/09/2012 vs 3.96 plain file: J63 v03.96 10/28/2019. Looks like the update within Win10 did corrupt some of the data.

[SDH]

...way beyond my skill set.

Take a look at the post here:

HP Z620: updating the boot block to support V2 xeon CPUs? - HP Support Community - 6501237

MtothaJ there pulled off a flash of the chip and he's the only one I remember having done this. Maybe you could contact him via PM for advice. It was a remarkable accomplishment. Me... I just take an hour and transplant in something that (supposedly) works. You never know what you've got until the first boot my way, however. Good luck!

(You can look up his posts from your forum account.)

[r1_rn22]

Thanks for this link. I think it will work out now, with what is left from the corrupt BIOS update.

Two more things that I ask your support:

The BIOS has actually 2 MAC addresses ... one is 1 digit higher at the last position (for me like MAC: ..... AC and ..........AD). In the BIOS file the higher one is written - is this the AMT address? My corrupted BIOS file only contains this one at the regions mentioned.

Would you mind looking in your BIOS to confirm, if the AMT MAC is the higher number? It should be the first one.

The "Chassis Serial Number" that the BIOS needs to be filled in - this is not the barcode number of the board, its the number on the sticker on the under side of the unit, starting with CZC - right?

For the other topics, I think I need to unsolder the chip again and do a manual editing of the mentioned regions. But at least I can confirm, that my system booted with a full 3.96 BIOS image, put 1:1 in the flash memory (boot block date shown is 03/06/2013), ME version shown is 8.1.60.1561

To summarize: I will manually add the missing MAC address, and put the file onto the flash. The rest should work then. In parallel, I will also contact the person you mentioned.

Thanks for your support, highly appreciated!

[SDH]

Happy to help... keep us posted on the progress if you can.

All our HP workstations from this era are souped up Z420 and Z620 models; all are v2 with the 3/6/13 boot block date. I'm the souper upper and do all the firmware updates plus the BIOS tuning. We have no Z820 workstations... DGroves has worked mainly with the Z8xx series, instead.

The Z620 and Z820 both have two LOM chips on the motherboard. On the case rear the AMT labeled port is below that of the lesser port. The Z420 has only one LOM chip, always AMT type. I block off the non-AMT port on the rear with a blank network cable plug, so no one accidentally reattaches to that lesser port. In BIOS on the System Information page the upper port listed there is "NIC Controller (AMT)" and its MAC on the Z620 I just checked ends in C while the lower listed port on that page is just called "NIC Controller" with its MAC ending in D. Otherwise the MAC addresses are identical.

Under System IDs in the Security page both adjustable entry lines we leave blank, and the UUID on that page is not alterable, comes with the motherboard, and is always listed. That is supposedly unique in the universe and is what I've been told is harvested automatically by Microsoft in W10/W11 activation to form the basis for that workstation's net-based "digital license". This is how once a workstation has been activated you don't need to have the original long license code available to do clean installs of W10/W11. We use some expensive software that also checks the UUID locally before it will run as part of its licensing.

Regarding the Chassis Serial Number for the Z420 and Z620 these here all start with 2UA and an example is [Personal Information Removed]. Those seem to rise in alphanumeric values based on when we bought them. Each x above represents a numerical value, not a letter.

The bar code on the motherboard has left-side and right-side part numbers. The left represent all the possible "brandings" and the right-side ones represent the actual part numbers. You almost never can tell what a motherboard is branded for (Windows/Linux/sometimes TPM firmware version toggle). The right-side numbers are how you know if you have a v1 or a v2 in the case of the Zx20 workstations. The workstation's serial number is on the outside label(s) of the case but never on the motherboard. Other codes on some of the bottom case labels allow you to "brand" a virgin motherboard so it is licensed exactly as the original motherboard was. I've only gotten 2 of those virgin motherboards off eBay over the years. Doing that process is not for the faint of heart but you've proven you could do it if you needed to. I have my HowTo notes on that if you ever need them.

I think that covers it!