General thoughts and suggestions on HEX-NET

   #41  

jyoung8607

FoRT
Verified
Joined
Feb 25, 2014
Messages
2,780
Reaction score
4,490
Location
Cincinnati, OH
VCDS Serial number
C?ID=25607
I think so, but what are you testing it on? It's never been available universally.
It definitely failed across the board on my 2006 Phaeton W12, on which my HEX-USB+CAN worked fine. I can't recall if I tested it on my 2004 Passat GLX or not. I'm travelling on business this week but I'll throw it on my to-do list for when I get back.

Jason
 
   #42  

PanEuropean

Verified VCDS User
Verified
Joined
Apr 23, 2014
Messages
111
Reaction score
118
Location
Toronto, Canada
VCDS Serial number
C?ID=11844
I have not been able to use my HEX-NET for the past week, because I forgot to unplug it from the OBD connector of the Audi RS7 I was working on, and the owner of the Audi headed off to Austria to spend the week at some kind of VW get-together being held at Wörthersee. He arrived back home today, that enabled me to fetch the device out of the Audi. If that HEX-NET thing has a battery inside it, it will certainly be well charged up by now.

Returning to a subject I mentioned on page 1 of this discussion, I think it would be in Ross-Tech's best interests to do something to make the device more conspicuous when it is inserted in the OBD connector. After being without it for the past week, the first thing I did when I got it back was to attach a ribbon to the little clip on the corner of the device. If nothing else, the presence of the ribbon will prevent me (or someone else) from driving away with the device connected.

Below are photos showing the ribbon.





I have a small, cheap sunglasses case that I picked up as a trade show premium some years ago. I'm sure it is very inexpensive, because the trade show exhibitor ordered 40,000 of them to use as booth giveaways at a 3 day trade show. It so happens that this little fabric sunglass case is just the perfect size to hold the HEX-NET device (with a ribbon attached). I realize that you folks have already gone to the trouble of designing a very elegant corrugated package for the HEX-NET, however, it might be a nice touch to include a little soft fabric case like this. My rationale is that because the case has a clip on it (to attach to a belt or pocket), it is less likely that the device will go missing - either in a car, or somewhere on a cluttered workbench - if the owner has a slip-case to hold the device in, and if the slip-case can be easily attached to clothing.

I don't know how many of the hard plastic shells you have already made for this device, but if you need to do another production run of shells, it might also improve conspicuity (and minimize lost and misplaced devices) if you had the plastic shells made in bright red or bright yellow, rather than flat black.

The Soft Case




 
  • Like
Reactions: Uwe
   #43  

jakematic

Verified VCDS User
Verified
Joined
Feb 13, 2014
Messages
812
Reaction score
632
Location
NC USA
VCDS Serial number
C?ID=129085
If that HEX-NET thing has a battery inside it, it will certainly be well charged up by now.

No battery.

But.... have him check the battery if you hadn't set it to sleep per 6.10 of the FAQ

After your last post I bought a bunch of these. :cool:
 
   #44  

jyoung8607

FoRT
Verified
Joined
Feb 25, 2014
Messages
2,780
Reaction score
4,490
Location
Cincinnati, OH
VCDS Serial number
C?ID=25607
I did get "Update Failed" on the CB, but that is indeed cosmetic, 'cause I've seen that more often than I'd like but never had one that actually failed.
I just upgraded from CB 4104 to CB 4105 and got the "Update Failed" message again. I think it's just a timing problem. This time I noticed the message while the HEX-NET WiFi LED was still flashing red, indicating it wasn't associated with my home WLAN yet and therefore my browser couldn't possibly have polled it for success yet. The difference was a small single-digit number of seconds, but it was there.

If I'm reading the browser-side JavaScript correctly, the browser polls for status in a loop every 2.4 seconds to update the progress bars. When it learns a reboot is imminent, it waits exactly 7.0 seconds and polls the interface for status exactly one more time. It then compares before/after version variables without first checking to see if it successfully fetched and parsed the status page. I fetched a copy of that status page by hand and the versions were okay. It appears to need another loop where it checks for any valid response (with a more generous total timeout) first, to wait for the interface to reboot and reassociate, and only then compare versions.

Jason

Manual browser fetch of "fw_update_status.shtml" right after the upgrade, CRLFs added for readability

Code:
<div id="main_percent">0</div>
<div id="sub_percent">0</div>
<div id="ld_mainstats">0% </div>
<div id="ld_subdesc">Preparing to update...</div>
<div id="ld_substats"></div>
<div id="ld_cur_status"></div>
<div id="ld_up_done">0</div>
<div id="cur_fw_version">0.184</div>
<div id="cur_cb_version">0.4105</div>
 
   #45  

Shaun

Active Member
Ross-Tech Employee
Joined
Jan 29, 2014
Messages
1,616
Reaction score
2,527
Location
USA
You are absolutely right that this is most likely a timing issue. We are actually well aware of this, and a fix for that timer was made long ago that will hopefully be slow enough to prevent this from being a problem. However, since it is currently the only change to the datapack, it has not been released on its own. This fix is in the queue, waiting for some future changes to go along with it.

Thanks for the feedback though.

--Shaun
 
   #47  

jyoung8607

FoRT
Verified
Joined
Feb 25, 2014
Messages
2,780
Reaction score
4,490
Location
Cincinnati, OH
VCDS Serial number
C?ID=25607
Following up to myself --

I'd love to see some default presets made available for common troubleshooting scenarios. Some random examples:

  • Troubleshooting misfires: Select all individual cylinder misfire counters
  • Troubleshooting A/F ratios: Select mass air-flow specified and actual, all fuel trim data (turns out to be widely scattered now!), etc
  • Troubleshooting transmission slippage: Select current gear, engine RPM, transmission input RPM (TC slip), transmission output RPM, etc
Both Desktop and Mobile do offer something of a way to craft our own presets, but with quite a bit of effort. Some default preset guidance would be helpful, especially on Mobile. If your eventual direction for Cloud is "GFF that doesn't suck", this would be a nice foundation piece to have.
It turns out this already exists on VCDS Desktop, in the form of some pre-built save files for Advanced Measuring Blocks. I just stumbled across them yesterday. I like it! I hope the idea grows further than the limited set available today, perhaps with community contributions. Also, I hope that some similar feature makes its way into VCDS Mobile.

Jason
 
  • Like
Reactions: Uwe
   #48  

jyoung8607

FoRT
Verified
Joined
Feb 25, 2014
Messages
2,780
Reaction score
4,490
Location
Cincinnati, OH
VCDS Serial number
C?ID=25607
You are absolutely right that this is most likely a timing issue. We are actually well aware of this, and a fix for that timer was made long ago that will hopefully be slow enough to prevent this from being a problem. However, since it is currently the only change to the datapack, it has not been released on its own. This fix is in the queue, waiting for some future changes to go along with it.

Thanks for the feedback though.
When I upgraded yesterday, I got a success message for the DS upgrade and a failure message for the CB upgrade. I looked at the page source after the CB upgrade failure message, and I saw the timer has been changed from 7000ms to 8200ms. That still seems really aggressive. Device reboot, wireless re-association, DHCP/ARP and friends, and HTTP service availability are going to take some time. That time won't necessarily be consistent in the same environment, much less different environments (mostly wireless and DHCP). For a one time no-retry check, something in the 45 to 60 second range would be a lot safer.

Anyway, just a data point. I understand now that it's cosmetic and I'll ignore it going forward unless you're looking for more feedback.

Jason
 
   #49  

jyoung8607

FoRT
Verified
Joined
Feb 25, 2014
Messages
2,780
Reaction score
4,490
Location
Cincinnati, OH
VCDS Serial number
C?ID=25607
Suggestion: Consider adding some protocol info in the Auto-Scan output. For each address successfully scanned, call out the physical layer type (K-Line, K1 vs. K2 if VCDS can see that, CAN), the K-Line or CAN bitrate, and the protocol layer (KW1281, KWP2000, KWP2000 over CAN(??), UDS). It would help with troubleshooting, with retrofits, and it would help people trying to understand what VCDS interfaces are compatible with their car.

Question: What the heck is a VCID? A VCID value is returned for every module in a scan, and they're not the same every time. Having collected a bunch of scan types of the same vehicles, I can see it seems to vary somewhat by module, VCDS version, interface model, and interface method (USB, WiFi, Mobile in the case of HEX-CAN). Is this something just for Ross-Tech support?

Jason
 
Last edited:
   #50  

Uwe

Benevolent Dictator
Administrator
Joined
Jan 29, 2014
Messages
48,925
Reaction score
33,652
Location
USA
VCDS Serial number
HC100001
Question: What the heck is a VCID? [...] Is this something just for Ross-Tech support?
Yup. And it actually contains everything you're asking about. Frankly, we prefer not to put all that stuff out in the clear for every wannabe competitor to see.

-Uwe-
 
   #51  

jyoung8607

FoRT
Verified
Joined
Feb 25, 2014
Messages
2,780
Reaction score
4,490
Location
Cincinnati, OH
VCDS Serial number
C?ID=25607
Yup. And it actually contains everything you're asking about. Frankly, we prefer not to put all that stuff out in the clear for every wannabe competitor to see.

-Uwe-
Heh, sorry. I didn't realize those two things were connected. I do have this relentless habit of looking for the Man Behind The Curtain that's served me well in my day job. My suggestion for PHY and logical layer info arose from a thread on the Phaeton forum, where someone was trying to retrofit an unusual Bluetooth module. I was trying to answer questions from him and Michael and wasn't fully confident in my answers.

Jason
 
   #52  

Uwe

Benevolent Dictator
Administrator
Joined
Jan 29, 2014
Messages
48,925
Reaction score
33,652
Location
USA
VCDS Serial number
HC100001
Your answer in the post linked above was pretty much spot-on. I'll elaborate on the questions you had though:

Where I start to have trouble is how far these signals get into the car from the diagnostic port, electrically speaking. I think K-Line is fully electrically connected to everything but I'm not sure.
On most cars, this is mostly the case. The Phaeton (and corresponding Bentleys) and the Touareg are exceptions. The first generation of these talk to a scan tool exclusively via K-Line, but traffic to some modules is then routed from the gateway (in the instrument cluster) via virtual K-Line(s) over the CAN buses inside the car. IIRC, there's a diagram that illustrates this in the SSPs (Self Study Programs) for these cars. On the second (facelifted) generation of these cars, there are indeed some modules that require communications directly via CAN, and there are other modules that can communicate both ways.

Another exception is that most K-line cars that are immobilizer equipped do not have the K-line going directly to the ECU; it gets there through the immobilizer (which may be integrated in the instrument cluster). This is done to prevent snooping/spoofing of the negotiations between the immo and the ECU at the DLC (Data Link Connector -- the OBD-II connector).

For CAN I think the DLC CAN bus electrically stops at the CAN Gateway (address 19), which is often built into the instrument cluster. It acts as a router, filtering and forwarding messages that need to be passed on.
Well, yes and no. The Gateway does indeed act as you describe, but most of the newer "CAN only" cars don't have the gateway in the cluster; it's a separate, stand-alone box.

And then there are cars like the B7 A4 -- it has two K-lines, but still needs direct CAN access for some modules, but has no CAN gateway; it exposes its powertrain CAN directly on the DLC. :eek:

There are a few other "Lets throw some stuff from the parts bins together" cars that require a mix of CAN and K; the PQ25 platform is a prime example. These do have gateways (but not proper stand-alone gateways), so at least they don't expose the powertrain CAN directly.

Oh, and I don't think any A8 ever had dual K-lines; dual K-lines were temporary kludge that were only used in the B6 and B7 A4 and the C5 A6 (from 2002 on). I suppose the Seat B7 derivative also had 'em.

-Uwe-
 
   #53  

jyoung8607

FoRT
Verified
Joined
Feb 25, 2014
Messages
2,780
Reaction score
4,490
Location
Cincinnati, OH
VCDS Serial number
C?ID=25607
On most cars, this is mostly the case. The Phaeton (and corresponding Bentleys) and the Touareg are exceptions. The first generation of these talk to a scan tool exclusively via K-Line, but traffic to some modules is then routed from the gateway (in the instrument cluster) via virtual K-Line(s) over the CAN buses inside the car. IIRC, there's a diagram that illustrates this in the SSPs (Self Study Programs) for these cars. On the second (facelifted) generation of these cars, there are indeed some modules that require communications directly via CAN, and there are other modules that can communicate both ways.

Another exception is that most K-line cars that are immobilizer equipped do not have the K-line going directly to the ECU; it gets there through the immobilizer (which may be integrated in the instrument cluster). This is done to prevent snooping/spoofing of the negotiations between the immo and the ECU at the DLC (Data Link Connector -- the OBD-II connector).
Thanks for the clues. Now I know why I couldn't find a straight answer on how that worked: there is no straight answer. I have read over the Phaeton SSPs pretty thoroughly, but I didn't realize that part was particularly unique.

And then there are cars like the B7 A4 -- it has two K-lines, but still needs direct CAN access for some modules, but has no CAN gateway; it exposes its powertrain CAN directly on the DLC. :eek:

There are a few other "Lets throw some stuff from the parts bins together" cars that require a mix of CAN and K; the PQ25 platform is a prime example. These do have gateways (but not proper stand-alone gateways), so at least they don't expose the powertrain CAN directly.
It's fascinating how many times they seem to have reinvented the wheel. It must drive the dealer diag tools vendor (whoever that is) to distraction. I would assume they are moving closer to consistency with MQB/MLB and friends. I see now why there's competitive value in the low-level details though.

Oh, and I don't think any A8 ever had dual K-lines; dual K-lines were temporary kludge that were only used in the B6 and B7 A4 and the C5 A6 (from 2002 on). I suppose the Seat B7 derivative also had 'em.
What was the kludge for? I had read somewhere it was too many controllers for one bus, which I thought was an A8 problem, and I would think the A8 would trip over that problem first.

Jason
 
   #54  

Uwe

Benevolent Dictator
Administrator
Joined
Jan 29, 2014
Messages
48,925
Reaction score
33,652
Location
USA
VCDS Serial number
HC100001
Too many controllers (too much capacitive loading?) on one bus is my theory as well, but of course Audi doesn't share the rationale as to why they do things with us. Yeah, you might think the A8 would run into that first, but the D2 didn't really have THAT much stuff it in it (and was replaced by the D3 in 2003 in Europe).

Of course the factory scan tools back then didn't support dual K-lines natively. So another real kludge was the Dual K <-> Single K "adapter" they came up with:

vas6017.jpg


I bought one as soon as they were available -- it cost more than I was selling a complete VAG-COM system for at the time. A few months later, we introduced the TWIN-COM intertface. We solved the problem with an AND gate. :cool:
 
Last edited:
   #55  

Bruce

Active Member
Staff member
Ross-Tech Employee
Joined
Jan 30, 2014
Messages
3,169
Reaction score
5,145
Location
Near Philadelphia, PA, USA
VCDS Serial number
--------
Sorry but now you have done it... an AND gate? You is really showing that birthday that happened the other day! Next thing you'll be speaking about 2N2222A NPN transistors! Relishing the old days are we?

How many in this world understand what an AND gate is?

Yep, just what I thought: limited to you and me! :D
 
   #56  

Uwe

Benevolent Dictator
Administrator
Joined
Jan 29, 2014
Messages
48,925
Reaction score
33,652
Location
USA
VCDS Serial number
HC100001
   #58  

jakematic

Verified VCDS User
Verified
Joined
Feb 13, 2014
Messages
812
Reaction score
632
Location
NC USA
VCDS Serial number
C?ID=129085
Next thing you'll be speaking about 2N2222A NPN transistors! Relishing the old days are we?
These aren't commonly used anymore ?
I'm appalled and apparently not going to retire on my stash of them.


How many in this world understand what an AND gate is?

Heh. Try asking about a triode, klystron, KS-15656, or even a rotary telephone...
Great now I feel old.
 
   #60  

romad

Verified VCDS User
Verified
Joined
Feb 10, 2014
Messages
207
Reaction score
80
Location
Prescott, AZ
VCDS Serial number
C?ID=208765
These aren't commonly used anymore ?
I'm appalled and apparently not going to retire on my stash of them.




Heh. Try asking about a triode, klystron, KS-15656, or even a rotary telephone...
Great now I feel old.

Must be all that selenium rectifier smoke! Oops, I forgot I'm on a party line and need to listen before calling. However my new phone will solve that problem: http://tinyurl.com/o3bhftf
 
Back
Top