You should either power it via its usb port or via the SWD, not both. Especially if you are powering them from different sources. Keep just the ground connection, two data lines and the reset if you want to power it via the ledger's usb. Later edit: although a lot of serial programmers (microchip, atmel, arduino, etc.) provide VDD to the target, it seems that this is not the case here (ST link nucleo). See the reply from Repligon below.
ST-link on nucleo doesn't provide power to the target. VDD_TARGET pin is there to read target's operating voltage. It won't even connet if voltage on this pin is too low.
in that case i guess it was resetting cause vcc and ground shorted(since it was swapped the vcc on the ledger was connected to the usb chassi on the st link) Oh well, that's some quite resistant hardware XD
@@Repligon I was pretty sure that was correct, but when I looked at a couple of ST devkit schematics, and checked a devkit I have in front of me, the resistor connecting VDD_TARGET to the programmer isn't even fitted! I've been connecting it religiously, but apparently on my devkit it's a do-nothing pin. A 10k pulldown and nothing else..
So you basically made a PCB so that the pins are at a right angle? Instead of just poking them through the PCB of the ledger? Why is that? Why not just solder the pins directly to the PCB and, looking at the pinout, directly put the ledger with it's pins into the debugger?
Tip: you can flash the ST-Link with the firmware of the Segger J-Link, enabling more features and compatibility. IDK if it works with non ST devices, but they say it would be "illegal"
Gosh, I love your videos especially the ones that talk about niche subjects in computer security such as Debugging Hardware. we need more. Love & respect
@@RoiEXLab Yes I also like to use ENTLÖTLITZE. For any soldering job ENTLÖTLITZE is a must have. The greatness of ENTLÖTLITZE must be capitalized here.
@@RoiEXLab Huh, TIL. I'm not even going to try to pronounce that but thanks for including it - I wonder if it translates to "copper mesh" by any chance. Google translates it to "braid" which i guess is correct, its also called solder braid too.
I once had to implement something on a FPGA for university. I hated it because the hardware world is so different than the software world I'm so used to :(
Shouldn't even be too hard as both of them are german (afaik). Would be interesting to see what a hardware guy and a software guy could to together. The soldering would still be shit though xD
When you want a soldering alloy to leak somewhere deep and narrow, use much flux (you really can't overshoot with it) and slightly (just slightly) overheated soldering iron. You touching your place with the tip and holding a little and then you can literally see how the alloy suddenly seatles down. Then you stop heating but let it cool down before you wil move it any kind. And in general use more flux and use a good one like FluxPlus, that can be leaved without washing and don't worry that it can corrode the copper.
I'm not sure trying to power the Ledger by both the debug interface and the USB connector at the same time is a good idea (but may be harmless - I don't know if it has protection diodes or not).
There is actually a nice strategic pick you can make there: Can you say you're a passionate of one of the two? Then pick the other. If you're passionate with algorithmics, system design and general computing stuff then picking EE will allow you to expand the knowledge in areas you would not pick up by yourself, that will often seem tedious redundant or unnecessary, but in the end provide you with a much better general picture.
At my University in germany, you can choose between Computer science and (maybe a bad translation) technical Computer science. One is more focused in Software, the other one more on Hardware.
My trick to make a path with solder : when you put it on the first point, don't remove the soldering iron and keep it barely attached to that drop of solder (try not to burn the boars) this way the flux inside it burns (or flows down, not sure) and it will follow the soldering iron that you use like brush to draw
ST nucleo board is more expensive than programmer here in Turkey. not too much. Thanks for the videos btw. Really interesting ones. I admire your works. Regards
Yeah. But currently I can still offer it for free and don’t need the money. Ideally I want it all to be free. Though can’t promise this forever in case life changes and need money. Also maybe I can make the patreon and ads model work instead ;)
Pin change happens and it is bad and terrifying!!😌😌😌 Just the day before yesterday I was working in addition for our school project, and took off the line from the ultrasonic module to pass through a hole in wood After reconnecting, horror of horror , the arduino turns on for a second but the ON led fades (AAAAA!!!) a lot of panic later, I realized the pin arrangement of the module was VCC. Trig Echo. GND (module pins) ¦ ¦ ¦ ¦ GND Echo. Trig. VCC ( my connection) SO GLAD to find it still worked! Thought I had fried the module 😌😌😌
2:18 Ehhh, +5V? Yikes no! That would kill your ST-LINK/V2! Its supported voltage range is 1.65 V - 3.6 V. Fortunately VDD of course wasn't 5V, since that would be deadly for your target as well ;-)
After using these demo kit ST_LINKs for several years, I just realised that on the two boards I've examined, VDD_TARGET is not connected to anything other than a 10k pulldown on the devboard. The resistor which would connect it to the programmer is designated as not fitted, and when I checked a real board, yes, it's not fitted. Might as well not connect that wire.. Hopefully the 'real' programmers are better than that. (edit) Ah, I see from Rafał Mróz's comment that you must fit that resistor if you snap off the ST-LINK PCB. The ones I've been using don't have the snap-off feature.
when the display board is connected to st-link and if st-link is not powered the program wont run (maybe its due to nreset being tied to the stlink).Power on the stlink and it will resume running.
You can buy $2 USB-dongle ST-LINK v2 from China. I bought bunch of them. Unfortunately they sometimes lose them HSE generator and stops working, but they very cheap.
Ah yeah! This video is perfect for a project I’ve been working on. Gave me some great ideas and also some insight into the debugging. Now that you dumped the firmware “original.bin” you could reverse engineer it correct?
Not sure the difference between the stlinkv2 and v3.. but how does the $40 stlink-v3/mini debugger differ from the $700 ulinkplus debugger? The latter is said to use coresight and swj (swd+jtag)... but the stlink also uses swj/jtag! So, for example, could the stlink-v3 hardware be used with openCSD to analyze a trace stream?
@LiveOverflow You was wrong about the VDD voltage, it's 3.3V not 5V, cause the stm32 not 5v tolerant, also I think you should not connect the second usb cable (to the ledger), cause it's already got the voltage form the VDD pin, but I'm not sure about that.
I use STM32 discovery board for STM32 bluepill. Need to get blackpill for testing. It seems almost all debugger prices are going down. microchip have some 10-15e debuggers (mplab snap) and Ti also have pretty cheap launchpad
please checkout the threat model documentation and the architecture/design of the ledger. This is the firmware dump of the "insecure processor". NOT the "secure element". It's not so simple and black:white.
@@LiveOverflow Hi, but if you got hands on a ledger, could you write a program on the insecure chip that when asked for authentication by the secure chip could send that to a computer, the computer calculates the answer from the firmware dump, sends it to the insecure chip, that sends it to the secure? It bypasses the size limit of the flash in the insecure chip (Inability to store all the responses to secure chip's authentication), allowing the key extraction if you, for example stole it and opened it up.
The chip reset because the reset pin got pulled up by the development board. I've seen this happen with Arduino's, ESP8266's and teensy's It's usually for communication or to activate programming mode. The reset pin has to be pulled up to synchronize communication and reset current running script which sometimes uses communication pins for your script to be able to upload. I'm going in to cyber security and I'm so glad I already have a background in electronics.
@@LiveOverflow I just checked and It looks like it was the ground and SWD clock you swapped so it wasn't even the reset pin strange. I suppose the ground pin must have pulled down the signals of the clock or something not really sure why that would happen but it shouldn't damage it at least.
What? You have pin headers. Why not use them to connect the ledger with the perfboard? Also you can cut the dupont cables and solder them directly to the ledger.
8:25 Oh yeah, you should have powered the ST-LINK/V2 _before_ powering on the target. As you noticed, doing it the other way around can cause problems, and I think it may also be bad for its long-term health. Hotplugging into a running system is also a headache since in my experience it tends to glitch the reset line. Basically, the ST-LINK/V2 just sucks :) I ended up ditching it and used my XDS100V2 (from Texas Instruments) instead, which doesn't care about the order in which you power or connect things, and keeps its I/O properly high-impedance until both it and the target are powered and the debug software clears the powerfail flag. It only does JTAG though, no SWD, but fortunately that was all I needed.
I guess its only for the main controller and not the secure element, though it stills gives way to reverse engineering and potentially malicious firmware that could, for e.g., record the PIN code or something like that. Since you would need physical access for it anyway, its not something that I think ledger would care too much about. It's still surprising how they actually gave you solder pads for debugging it though - i wonder if that's going to do more harm than good...
Why are you doing simple connection in so difficult way? Nothing criminal to solder wires to spots on the PCB immediately. Wires can be easily unsoldered back one by one but unsolder hard plate from all pads simultaneously is much harder and there is danger to damage the pads.
Seriously, why would you use both the link usb and pcd usb together? Then you wondered why the board did a reset? The soldering part was a complete waste of time.
@@LiveOverflow Guess someone didn't understand what the whole video was about. On another note I think you can use this with VSCode + PlatformIO, so you can actually set breakpoints and such in your source code.
There is absolutely no difference between those and the onboard one, it's just a STM32F103CB, except it has a different header shape. Except you have a nucleo board as an added bonus, and they come in handy.
Why wouldn't you get a male/female header for 25 cents and connect them that way? Then you can disconnect it as well. Great analysis of the datasheets, but the hardware makeup is horrible!
@@ilya_mzp I know, but those bridges can be done using wires - see the video. To be honest, he should've just soldered the coloured wires straight to the board instead of messing with the additional PCB.
*screaming* “use a pin header!”
*think* with all this trouble he must not have any
*opens drawer full of pin headers*
DEAD
hahahahahahaha soo true!
Seriously what the fuck
breadboard, some pin headers and fit the pcb on top = done. :D
solder needed = zero
SAME xD
I love that you show the whole thought process, including the failed steps as well.
You should either power it via its usb port or via the SWD, not both. Especially if you are powering them from different sources. Keep just the ground connection, two data lines and the reset if you want to power it via the ledger's usb.
Later edit: although a lot of serial programmers (microchip, atmel, arduino, etc.) provide VDD to the target, it seems that this is not the case here (ST link nucleo). See the reply from Repligon below.
ST-link on nucleo doesn't provide power to the target. VDD_TARGET pin is there to read target's operating voltage. It won't even connet if voltage on this pin is too low.
@@Repligon oh, didn't know that. I thought that it also provides power.
in that case i guess it was resetting cause vcc and ground shorted(since it was swapped the vcc on the ledger was connected to the usb chassi on the st link)
Oh well, that's some quite resistant hardware XD
@@Repligon I was pretty sure that was correct, but when I looked at a couple of ST devkit schematics, and checked a devkit I have in front of me, the resistor connecting VDD_TARGET to the programmer isn't even fitted! I've been connecting it religiously, but apparently on my devkit it's a do-nothing pin. A 10k pulldown and nothing else..
So you basically made a PCB so that the pins are at a right angle? Instead of just poking them through the PCB of the ledger? Why is that? Why not just solder the pins directly to the PCB and, looking at the pinout, directly put the ledger with it's pins into the debugger?
Looks cool LOL
I thought it looks cool 🤷♀️
@@LiveOverflow lol agreed
so its more rigid to solder more hacks on :D add i2c sniffer pins, and to add blue status leds :D .... or it looks cool
@@LiveOverflow lol 😂
3:31 - Technically, the dot (or beveled corner) indicates pin-1.
Tip: you can flash the ST-Link with the firmware of the Segger J-Link, enabling more features and compatibility. IDK if it works with non ST devices, but they say it would be "illegal"
You could've used the pin headers as a link between your perf-board and the PCB, or directly hook up the female-to-male connectors to the pin-headers.
Hey that's pretty cool
Gosh, I love your videos especially the ones that talk about niche subjects in computer security such as Debugging Hardware. we need more.
Love & respect
BTW Mr Overflow, that copper mesh is called solder wick.
ENTLÖTLITZE in German ^^
@@RoiEXLab Yes I also like to use ENTLÖTLITZE. For any soldering job ENTLÖTLITZE is a must have. The greatness of ENTLÖTLITZE must be capitalized here.
@@RoiEXLab Huh, TIL. I'm not even going to try to pronounce that but thanks for including it - I wonder if it translates to "copper mesh" by any chance. Google translates it to "braid" which i guess is correct, its also called solder braid too.
@@QuickishFM Entlötlitze means Desoldering braid
@@QuickishFM Huh, I don't know TIL. What does that acronym mean?
I am both simulatenously attracted to eletronics because of how cool it is and repulsed by it because how annoying it can be
Same. Exactly.
I once had to implement something on a FPGA for university. I hated it because the hardware world is so different than the software world I'm so used to :(
Next time just solder jumper directly to pad and don't apply any force. You dont want pad to lift up. Trust me its simple as that.
Why haven't you soldered the header directly on the pcb?
PCB have no holes, those debug dots are placed right on top layer of PCB, so it's kinda hard to solder a header there
@@V4ker No need holes, it is easy to solder the header to the pads.
@@V4ker You can easily solder the header sideways if you know what i mean
@@gabiold Yeah, but it's kinda way too easy to rip them off
That soldering hurt my brain, but everything else was super interesting. Appreciate all you're doing :)
Great series! Thanks for putting so much effort into these videos.
I'd love to see a collab with GreatScott, whatever that would be!
Shouldn't even be too hard as both of them are german (afaik). Would be interesting to see what a hardware guy and a software guy could to together. The soldering would still be shit though xD
Your intro on the ledger was pretty cool. I don't understand much of this, but your videos are really entertaining nonetheless 🙃
That intro was AWESOME
as a non native english speaker your subtitles are a bless, regards from spain :D
Nice video! Can't wait for the next episode of Hardware Wallet Research#4!
The dot on the chip designates PIN 1 and NOTHING ELSE!
Soso, das Bändchen vom 35c3 fehlt schon ... :D
Tolles Video mal wieder - Hardware Debugging macht richtig Laune! :)
this video was recorded ~6months ago
Never trust the upload-date .. point taken. :D
Anyway - interesting stuff, waiting for the next part! :)
When you want a soldering alloy to leak somewhere deep and narrow, use much flux (you really can't overshoot with it) and slightly (just slightly) overheated soldering iron. You touching your place with the tip and holding a little and then you can literally see how the alloy suddenly seatles down. Then you stop heating but let it cool down before you wil move it any kind.
And in general use more flux and use a good one like FluxPlus, that can be leaved without washing and don't worry that it can corrode the copper.
I'm not sure trying to power the Ledger by both the debug interface and the USB connector at the same time is a good idea (but may be harmless - I don't know if it has protection diodes or not).
Sticky tack works awesome for holding components on the board while you solder.
That intro is a nice touch
What do you think is better to study to learn something like this? Electrical Engineering or Computer Science?
Does it matter? Just start right now ;)
@@LiveOverflow Well I can't decide between these two and hope you can help me :)
@@maximood-tired Electrical Engineering, its closer to the hardware, low level stuff :)
There is actually a nice strategic pick you can make there:
Can you say you're a passionate of one of the two?
Then pick the other.
If you're passionate with algorithmics, system design and general computing stuff then picking EE will allow you to expand the knowledge in areas you would not pick up by yourself, that will often seem tedious redundant or unnecessary, but in the end provide you with a much better general picture.
At my University in germany, you can choose between Computer science and (maybe a bad translation) technical Computer science. One is more focused in Software, the other one more on Hardware.
Endlich mal ein Guter Deutscher RUclips Kanal!!
LG
But it’s an english channel
A really Good english german channel ;)
My trick to make a path with solder : when you put it on the first point, don't remove the soldering iron and keep it barely attached to that drop of solder (try not to burn the boars) this way the flux inside it burns (or flows down, not sure) and it will follow the soldering iron that you use like brush to draw
Now this is hardcore. Im in awe
2:25 the target isn't generally +5V and a stm32 is never! it isn'T 5v tollerant either so take care!
ST nucleo board is more expensive than programmer here in Turkey. not too much. Thanks for the videos btw. Really interesting ones. I admire your works. Regards
Have you ever thought about making professional learning course about reverse engineering for Pearson, Lynda, Udemy, Pluralsight...etc ?
Yeah. But currently I can still offer it for free and don’t need the money. Ideally I want it all to be free.
Though can’t promise this forever in case life changes and need money.
Also maybe I can make the patreon and ads model work instead ;)
Pin change happens and it is bad and terrifying!!😌😌😌
Just the day before yesterday I was working in addition for our school project, and took off the line from the ultrasonic module to pass through a hole in wood
After reconnecting, horror of horror , the arduino turns on for a second but the ON led fades (AAAAA!!!)
a lot of panic later, I realized the pin arrangement of the module was
VCC. Trig Echo. GND (module pins)
¦ ¦ ¦ ¦
GND Echo. Trig. VCC ( my connection)
SO GLAD to find it still worked! Thought I had fried the module 😌😌😌
It works fine, if you first connect two spots. Let it cool down an connect the third spot. In this way you can do it like you need it.
you could try to use golden tape in order to secure the ledger, not sure if it's better than electrical
2:18 Ehhh, +5V? Yikes no! That would kill your ST-LINK/V2! Its supported voltage range is 1.65 V - 3.6 V. Fortunately VDD of course wasn't 5V, since that would be deadly for your target as well ;-)
A great work. Thanks for sharing.
Just a small correction: Modern ARM based MCUs almost always have a nominal VDD of 3.3V.
After using these demo kit ST_LINKs for several years, I just realised that on the two boards I've examined, VDD_TARGET is not connected to anything other than a 10k pulldown on the devboard. The resistor which would connect it to the programmer is designated as not fitted, and when I checked a real board, yes, it's not fitted. Might as well not connect that wire..
Hopefully the 'real' programmers are better than that.
(edit) Ah, I see from Rafał Mróz's comment that you must fit that resistor if you snap off the ST-LINK PCB. The ones I've been using don't have the snap-off feature.
when the display board is connected to st-link and if st-link is not powered the program wont run (maybe its due to nreset being tied to the stlink).Power on the stlink and it will resume running.
You can buy $2 USB-dongle ST-LINK v2 from China. I bought bunch of them. Unfortunately they sometimes lose them HSE generator and stops working, but they very cheap.
How did you get the confidence in the beggining to know that its possible to do this in the autodidactic way that you do?
Really good informations 👍
nice vid, fabian!
Ah yeah! This video is perfect for a project I’ve been working on. Gave me some great ideas and also some insight into the debugging. Now that you dumped the firmware “original.bin” you could reverse engineer it correct?
Not sure the difference between the stlinkv2 and v3.. but how does the $40 stlink-v3/mini debugger differ from the $700 ulinkplus debugger? The latter is said to use coresight and swj (swd+jtag)... but the stlink also uses swj/jtag! So, for example, could the stlink-v3 hardware be used with openCSD to analyze a trace stream?
@LiveOverflow You was wrong about the VDD voltage, it's 3.3V not 5V, cause the stm32 not 5v tolerant, also I think you should not connect the second usb cable (to the ledger), cause it's already got the voltage form the VDD pin, but I'm not sure about that.
im not sure but isnt armV6 is mix of thumb/thumb2 (16-32) bit instructions?
I use STM32 discovery board for STM32 bluepill. Need to get blackpill for testing. It seems almost all debugger prices are going down. microchip have some 10-15e debuggers (mplab snap) and Ti also have pretty cheap launchpad
It's nice to see, that i'm not the only one making stupid mistakes like that 7:39
8:35 Literally anytime I come across a bug, and I ask my teacher to come look at it, only for my code to work perfectly fine the second time I run it
Thanks for the video!
Do the start of the ARM flash is written in the Datasheet?
Super duper Video!
Full Firmware dump? And this is secure?
Are you serious???
please checkout the threat model documentation and the architecture/design of the ledger. This is the firmware dump of the "insecure processor". NOT the "secure element". It's not so simple and black:white.
@@LiveOverflow Hi, but if you got hands on a ledger, could you write a program on the insecure chip that when asked for authentication by the secure chip could send that to a computer, the computer calculates the answer from the firmware dump, sends it to the insecure chip, that sends it to the secure?
It bypasses the size limit of the flash in the insecure chip (Inability to store all the responses to secure chip's authentication), allowing the key extraction if you, for example stole it and opened it up.
Afaik the secure chip requires tight timings that can’t be (easily) met with a proxy like that. But definitely one of the ideas to try!
Great video!
I, subcribed and like and watched the whole video, almost all of your videos. Even if I dont understand a thing. 😂🤣
The chip reset because the reset pin got pulled up by the development board. I've seen this happen with Arduino's, ESP8266's and teensy's It's usually for communication or to activate programming mode. The reset pin has to be pulled up to synchronize communication and reset current running script which sometimes uses communication pins for your script to be able to upload.
I'm going in to cyber security and I'm so glad I already have a background in electronics.
Yeah of course. But here it happened because I reversed the cables and it was accidentally triggering the Reset - not the reset sent when debugging.
@@LiveOverflow I just checked and It looks like it was the ground and SWD clock you swapped so it wasn't even the reset pin strange. I suppose the ground pin must have pulled down the signals of the clock or something not really sure why that would happen but it shouldn't damage it at least.
Nice work
What? You have pin headers. Why not use them to connect the ledger with the perfboard?
Also you can cut the dupont cables and solder them directly to the ledger.
Be carful. You said vdd = 5V when Vdd for stm32 is always around 3.3 V (2-4V)
8:25 Oh yeah, you should have powered the ST-LINK/V2 _before_ powering on the target. As you noticed, doing it the other way around can cause problems, and I think it may also be bad for its long-term health. Hotplugging into a running system is also a headache since in my experience it tends to glitch the reset line. Basically, the ST-LINK/V2 just sucks :)
I ended up ditching it and used my XDS100V2 (from Texas Instruments) instead, which doesn't care about the order in which you power or connect things, and keeps its I/O properly high-impedance until both it and the target are powered and the debug software clears the powerfail flag. It only does JTAG though, no SWD, but fortunately that was all I needed.
It is weird that the memory protection is not enabled on such device (chapter 3.3 of RM0091).
I guess its only for the main controller and not the secure element, though it stills gives way to reverse engineering and potentially malicious firmware that could, for e.g., record the PIN code or something like that. Since you would need physical access for it anyway, its not something that I think ledger would care too much about. It's still surprising how they actually gave you solder pads for debugging it though - i wonder if that's going to do more harm than good...
extremely informative!!!1!
Why are you doing simple connection in so difficult way? Nothing criminal to solder wires to spots on the PCB immediately. Wires can be easily unsoldered back one by one but unsolder hard plate from all pads simultaneously is much harder and there is danger to damage the pads.
can you post link of dumped firmware?
Flux is your friend. Also could have just soldered wires straight to it instead of a pin header
You should try openOCD. It is way more reliable than st-util and works with mcu's other than stm.
we back bois!
Seriously, why would you use both the link usb and pcd usb together? Then you wondered why the board did a reset? The soldering part was a complete waste of time.
ok boomer
@@LiveOverflow Guess someone didn't understand what the whole video was about.
On another note I think you can use this with VSCode + PlatformIO, so you can actually set breakpoints and such in your source code.
Hello, does anybody knows how to send comannds to stm32f042 with windows ?
When i think of a JTAG, i think of that old modded xbox 360 i had
Lol, any time i watch videos of binary reverse engineering i'm taken back to modding the BIOS of my t430 to accept a different wireless card.
Omg. You need some training on how to solder ;D
Pure awesomeness
So, are you an electrical engineer or a computer scientist?
Why do you need boxes? I don’t know what I am. I’m just enthusiastic about how stuff works ;)
I know this feeling of connecting the wrong wires ;) Destroyed one ATMega because of it ^^ So, see it this way: Education is expensive :D
So 1337 much h4x0r
In all seriousness I love these hardware videos they are a nice variation from CTF vids and really informative
wow so impressed
Everybody complaining about how he soildered it and then I don't see anything wrong with it lol
just buy some flux if you plan on doing something like this again it makes it a lot easier
you don't even need development board !
you can order cheap ST-link on aliexpress or ebay for few $
And?
The soldering was painful to watch
Electronics engineers be like ......owch :)
Naa im EE and I liiiiike. 😉
So cool
Why didn't you just buy these cheap chinese ST Links for just 3 USD?
Why should I?
There is absolutely no difference between those and the onboard one, it's just a STM32F103CB, except it has a different header shape.
Except you have a nucleo board as an added bonus, and they come in handy.
@@LiveOverflow Cheaper, has a case. A Bluepill is also a good option :)
you could have just got an ST link from eBay for just $3
my brain hurting lol
Dude, connect your grounds together (:
Remote??
It was awosome
Why wouldn't you get a male/female header for 25 cents and connect them that way? Then you can disconnect it as well. Great analysis of the datasheets, but the hardware makeup is horrible!
Dude, use flux! 6:40
But flux works the other way around. It makes solder liquidy to roll into balls. Without flux solder is sticky and make spikes and bridges.
@@ilya_mzp I know, but those bridges can be done using wires - see the video. To be honest, he should've just soldered the coloured wires straight to the board instead of messing with the additional PCB.
@LiveOverflow Ich hab das gleiche Multimeter wie du
Ground issue
What a cheap ads!! Shame on you...
no offense but this was hard to watch. Just one bad idea after another...