Please or Register to create posts and topics.

EVC2 Beta Software

Quote from elmor on 2020-01-01, 01:03

So the chip is detected correctly, but the data is all 00 or FF? Are you using the SPI_BOOSTER for improved GPIO drive strength? https://www.elmorlabs.com/index.php/product/evc2n2-spi-booster-adapter/ (it can be ordered for free for existing customers, you can contact me if you need it). Do you have more information about this "SPI readout lockdown"?

Oh i didnt order it, because usually chinese airmail is checked here and taxed 
Didn't have the resources to cover the price plus the higher shipping  
I'll contact you here or on OCN - need to check how to PM on here 🙂

No SPI_Booster so far, but voltage looks to be consistent jumping from 3.08v to 3.10v - and it looks also to hold it when i unplug the EVC2 via usb 
Detection wasn't working, and still isn't working - but i just set the parameters by myself 
Unsure if it matters but blank checking and verification work - yet it didn't wipe the bios chip
Read and export "works" but it has no data inside
Could backup it once but seems like i locked it - checking with MeInfo, read and write access is locked after my first update

I have no 2nd pc right now to verify, if i'll get access to the flashchip when the machine is not running (normaly should)
But so far it's locked, and every readout is empty + there is no detection - tho voltage readouts match

how does the SPI Booster work, i mean when would i actually need it ?
Is there so far no way to use UNIX with the EVC2 - so i technically could flash it via android command shell ~ through OTG cable

EDIT:
Blank checking / even tho "completed" doesn't wipe the flashchip, or does it ?
Send a mail to store@elmorlabs.com about the shipment request ~ hopefully the correct one 🙂

Spoiler

As mentioned it shouldn't cost you anything to order it as long as you enter the coupon code I sent you. Can't speak for import tolls and taxes.

The SPI Booster has nothing to do with the supply voltage, it's helping the CSn/WPn/HOLDn pin ability to pull the signal high. I found that on some systems where the chip is in-circuit (soldered to a board) with external power supplied, those pins are sometimes pulled down with very low resistance. This means they need to be forced high to be able to talk to the chip.

Your first step is to make sure the chip is detected to ensure the connection is good. It will read out the chip id even if it's not supported and report it's an unknown chip. Are you supplying power from the EVC2 or is the board self-powered? How did you connect the EVC2 to the chip?

Thank you, and for the explenation 🙂
Ya, they normaly check bigger taxed parcels with insurence and trow their 20% import tax on it by will :')
Soo if it's gods which are not declared without any visible recipe to orient themself about the price ~ they just don't put tax on it ^^
Well before 2019 we had not such kind of problem ^^' now they just search for money grabbing opportunities 😀

About the connection - it seems to be user error or EVC2 error 
SOIC clip was double checked, red cable faces outwards towards VMOD1 header, near Vin2
Cable goes over the OLED screen - CLIP red cable is oriented at the bottom left corner of the W25Q32BV
Like the tutorials normaly mention
Red from SPI_CSn -> CS

Spoiler


It had constant voltage, was enough to even power by itself
Didn't suply extra 3.0v as Vcc frpm the board was suplied
Disabled Automatic Vcc, as it already got one 
Checked also slower SPI frequency and once the OC mode 
As it didn't find it, i picked myself the needed specs, made a device ID Clone of the FV version (tried to make it detect that way) in the XML
Saddly no results, it looks like it reads something , but all the exports are empty 

Got SPI flashing unlocked by a HDA pinmod and can use ftp(w)indows to remotly SPI flash it
Later on wiped IntelME and reconfigured the VSCC table of the bios to know the flashchip location/size/device ID
(was missmatching on my buggy bios)
But this didn't help detection, nothing rly changed, same chip same device ID - all locks wiped / but yet no SOIC clip readout 🤔
Saddly can't confirm if it would work with own provided power, away from it recieving it's own
Had the same struggle with the X370 Taichi at start, readout only worked after i could provide own Vcc voltage
No detection worked, if there was any active power suplied to it
Maybe rly SPI booster issue ? 
* OLED display rotation is a great addition 😀
** only interesting thing i saw was that ftpw readout disabled SPI header readout / and EVC readout disabled (temporary) ftpw readout, if run at the same time

If it doesn't find anything by itself (it should at least report found unknown flash part), it means something with the I/O is wrong. From your description you've connected it correctly, assuming "CLIP red cable is oriented at the bottom left corner of the W25Q32BV" means on your board and not the pinout picture from the datasheet since there it should be oriented at the top left corner (pin#1 /CS). 

Reading from the chip (which also blank checking is doing) is a "blind" process. It can be done even if there's no chip present, which will result in all 00 or FF just like you see. Similarly the erase function blindly erases and waits for a finished indication, meaning erase can be started but will never finish if the connection is bad.

Do you have a DMM (multimeter) available? If so you could connect the EVC2 and check what voltages are on the CSn/WPn/HOLDn pins. They should be "high", meaning the same or close to Vcc. If they're much lower, say half or even 0 V, there's the pull-down issue I mentioned.

You say you can't try with the board powered off, and supply Vcc from the EVC2. Why is that? In some cases the on-board master can block I/O to the chip, which is probably why the Taichi worked only when powered through the EVC2.

 

Do you have a DMM (multimeter) available? If so you could connect the EVC2 and check what voltages are on the CSn/WPn/HOLDn pins. They should be "high", meaning the same or close to Vcc. If they're much lower, say half or even 0 V, there's the pull-down issue I mentioned.

I don't have one right now by hand , would need to borrow it - but i will check it in the next 2-3 days to know what's up

Reading from the chip (which also blank checking is doing) is a "blind" process. It can be done even if there's no chip present, which will result in all 00 or FF just like you see. Similarly the erase function blindly erases and waits for a finished indication, meaning erase can be started but will never finish if the connection is bad.

Do i understand it right ~  both blank check and test erase (chip search tab) are not actually erasing anything on the chip ?
I was always worried that the written test bytes it does write - will wipe important data of the flashchip , soo never rly run that option unless i was going to wipe the whole flashchip anyways 

You say you can't try with the board powered off, and supply Vcc from the EVC2. Why is that? In some cases the on-board master can block I/O to the chip, which is probably why the Taichi worked only when powered through the EVC2.

The issue is, if i was going to semi-brick this machine, i won't have any other windows device to restore it
And as we miss any other way to use the internal SPI flasher, except via Windows
(portable installations don't load the drivers correctly, zadig doesn't help) 🙁
Soo i had to make sure, i get a clean dump of it ~ which finally worked by puting IntelME R&W lockdown into menufacturing mode, to dump it and modify

If i mess something up right now, it would mean going to visit colleagues ~ just to get any kind of windows machine running / and disassemenbling there the notebook
* getting access to the chip was alright, half disassembled but still fully running ^^''
I could technically do SPI flashes through debian, android via OTG (runs 2 systems with MHL support), liveboot UNIX/debian, and hackintosh (UNIX) 
But in reality , it only works so far with a real windows installation ~ there is no way to set up the EVC from a portable windows and any kind of wine, doesn't load the drivers correctly
Tho my actuall issue is, that my main taichi machine is in sleep mode, since the last experiment where the PSU died at the end :') 

On the ryzen, live erasing the flashchip worked but writing hanged up 😀
But on this intel machine, the chip can be erased and written ~ while in working condition / which requires a cold boot to load the bios of it

Do i understand it right ~  both blank check and test erase (chip search tab) are not actually erasing anything on the chip ?
I was always worried that the written test bytes it does write - will wipe important data of the flashchip , soo never rly run that option unless i was going to wipe the whole flashchip anyways 

Blank check doesn't erase, it reads back data from the chip and checks if it's all = FF (erased state). Anything saying erase does erase the chip. The test chip option erases and writes a test pattern to the chip, then tries to read it back to verify everything is working as expected. What I meant with the above is that those procedures are not communicating with the chip but is blindly sending or receiving data. When you read all FF it means it keeps reading high on the MISO pin. When you read all 00 it means you're reading low on the MISO pin. There's no way to know if that data is actually from the chip or if it just happens to be the pin state. You can try it without the EVC2 connected to anything and see it still completes.

Ok, I think I understand your situation now. That complicates things a bit 🙂 There are a few options with various degree of risk, but it sounds like the best option might be to wait for the SPI Booster and a working second PC to proceed.

Did I understand it correctly that you have a second Debian PC? How about setting up Vmware Player with a Windows image on that and then do USB device passthrough of the EVC2? I would like to add Linux support for the EVC2, but it would be way too much work compared to how many would use it.

Quote from elmor on 2020-01-02, 02:25
Did I understand it correctly that you have a second Debian PC? How about setting up Vmware Player with a Windows image on that and then do USB device passthrough of the EVC2? I would like to add Linux support for the EVC2, but it would be way too much work compared to how many would use it.

It's a multi OS running on a Nexus 5, one part is debian via MHL cable, the other is native android - running at the same time
I could also boot up a whole ubuntu touch envoirement or run any kind of linux as live usb
Have 2 phones and could access the EVC2 (via OTG cable), if i know how - through linux
But i'm missing "installed windows" machine for the EVC suite & liveboot windows portable (win10&7) can't get the HID drivers to run

I mean i'm down for experiments 🙂
After all this thing is fine now, i can ftpw flash, or also flash through EFI/MS-DOS 
And at worst have my full backup here 
IntelME is wiped, and there are no flashing restrictions whatsoever
Just wonder why it can't read it out, where ftpw clearly stops on EVC readout try = it does try, but no data passes through ~ while voltage is consistent 🙂

New version 1.0.1.2, lots of updates 🙂

Hello,

I am one of the newbies that got the EVC2 after der8auer used and mentioned it in his last video. From what I can read in the description and what der8auer mentioned in his video, it can also be used on analog voltage controllers. I mainly got it to gain control over my GTX1080 which uses an uP9511 for some DICE/LN2. Is this chip supported and if yes how would I set it up ?

Apart from that thank you for building such a nice tool.

Quote from Aaron on 2020-04-23, 06:48

Hello,

I am one of the newbies that got the EVC2 after der8auer used and mentioned it in his last video. From what I can read in the description and what der8auer mentioned in his video, it can also be used on analog voltage controllers. I mainly got it to gain control over my GTX1080 which uses an uP9511 for some DICE/LN2. Is this chip supported and if yes how would I set it up ?

Apart from that thank you for building such a nice tool.

Hello, yes it's possible. Check out this topic: https://www.elmorlabs.com/index.php/forum/topic/evc2-vmod1-header-tutorial/

Let me know if anything isn't clear 🙂