Sevcon Gen 4 Errors and Solutions thread

methods

1 GW
Joined
Aug 8, 2008
Messages
5,555
Location
Santa Cruz CA
SOLVED: 24V contactor not compatible with voltages over ~110V
https://endless-sphere.com/forums/viewtopic.php?f=38&t=88679&p=1293374#p1293374


Error list that I reference: http://jan.peman.com/hacks/EmergencyErrorCodes.txt

Error I am seeing on a new setup: 0x5185 Alg Out Off with FS

This setup has a pot input on Foot Brake, no foot switch, and my intent is to regen with it. I suspect the error revolves around this domain.


Error in context from above list:
0x5181 Dig In Wire Off
0x5182 Alg In Wire Off
0x5183 Alg Out Over I
0x5184 Alg Out On with No FS
0x5185 Alg Out Off with FS

So... Dig means digital
Then Alg must mean analog? May be wrong on that.

guesses
5182 detected when a signal rails
5183 detected when measured I over acceptable for longer than time X
5184 Something to do with analog output being on... while there is no Foot Switch ? (FS means Foot Switch I am guessing)
5185 Something to do with analog output being off with a FS

From that I surmise that whatever "Alg Out" is.... it should be OFF with no FS and ON with a FS

The error indicates that I HAVE a FS which I do not believe I do... so off to check the settings.

FS could refer to FS1 I suppose... which I do have wired... and is acting on the system.
Ah... thinking back to something the master said...

The master said that one can not use a Foot Switch at the same time as a Foot Lever (potentiometer) as the Foot Switch will override the analog input.

Ok - removing FS1 from the software configuration and retesting.
I had mapped it to one of the first digital inputs for use as an interlock
We will rely on FWD/NEU/REV as this interlock if possible or otherwise work around

I prefer there to be a LOCKOUT so that a motor can not blast off without the user driving the throttle all the way to zero and ALSO a switch going from OPEN to CLOSED>.... i.e. if it starts up with switch CLOSED, then it must OPEN and CLOSE again to proceed.

Interlocks... them's is there's for a Reasons's ... :wink:

-methods
 
FS...

Forward Switch
Foot Switch
frocking Sevcon...

Lol...

Verbose errors are great.
As a dealer... having to go back to the manufacturer to have my hand held for every little thing is not great. Thats not how I roll.
Somewhere there is a description of all these errors in excruciating detail.
My engineering hours are being wasted because it is not publicly posted.
These are not billable hours.... and that affects my attitude.

-methods
 
Hmmm... FS1 is not mapped to the Foot Brake Switch, as that is a separate option in the digital in pull down menu.
Put it back in...

Sigh - I broke my rule of "Change One Thing At A Time" <COTAAT> while setting this controller up. So many features on this one... User has big plans.

-methods
 
Solution Determined.

Master mentioned that often the error indicates a blown fet or otherwise.
This was a size 6... looks pretty new.

Swapped out with another - same error (but then my "other" pile is of unknown pedigree for the most part)

Used my shorty-harness to disconnect all signals... did not go away.
Ahh...

But forgot to disconnect the CONTACTOR.
Removed it from the circuit - error gone.

Turns out the Sevcon is so smart that even while NOT DRIVING the contactor it decided that was a problem that it could not manage.
Contactor is a 24V with economizer.

There are two errors I have seen now with 24V contactors:

Gigavac 24V... Internal diodes short out when using a controller on 116V
Kilovac 24V with econo... apparently it is too low of impedance.

Onward.

Moral of the story - dont use 24V contactors with a 116V Sevcon Gen 4
Maybe at 80V... or some other voltage... but not at 28S
Leakage currents get too high - flip triggers.

Remember... this is an 80V controller that we are running at 116V... and the 24V and other voltage controllers have TUNED INPUTS AND OUTPUTS - (listed in the manual) so its clear that we are on to a fringe problem here.

Implementation problem, not Sevcon problem.


-methods
 
NOTE on Contactors:

Kilovac 12/24V
To use with 116V...

Remove Economizer
Set Sevcon to drive at 12V and hold at 6V (6 may turn to 8)

It works, no errors.
24V throws an overcurrent error
Economizer throws a weird error

Best bet - dont use 24V contactors on >100V bikes

On to testing Regen and 3 profiles...

-methods
 
Good stuff Methods, I am not sure if you have downloaded the manual for these controllers but it does state that economisers and diodes are not required with these Sevcons. Sevcons deal with this internally.

Bernel

Sent from my SM-T310 using Tapatalk
 
I have spent many hours with the manuals.

Admittedly I have not read them end-to-end like I do with microController docs... as there is a lot of irrelevant info in there.

I think you may have understated it:
Contactors with Economizers and bypass diodes actually trigger the protection circuits, result in lockups, and otherwise fail hard.

I do see that the guts of the Sevcon are "strapped hard" with sacrificial diodes.
This is great since it is an un-potted assembly.

We are seeing more and more that comercial components are "broken" simply due to shorted strapping diodes... which can be removed and/or replaced... bringing units back online even in the field. Perfect!

Still pondering what is right with a BMS...
To strap the channels hard... which in a noise event may result in cells being discharged relatively rapidly...
Or leave bald componets which will "fuse out" quickly but can be detected in self test.

We are going to be potting all of our BMS units so these things become VERY important.
In a situation where a BMS is attached only once and can not be touched again...
I am tempted to skip the big diodes strapping the inputs... run some large inline resistances... and accept the slower settling times and accuracy loss... to help protect the battery.

Hmmm... needs to be a T-Handle on a BMS connection which is potted so as to free it from its battery in the case of damage - thereby allowing battery to persist
Yea... dont want my BMS potted.
I want it external and IP67 rated...

(sorry - totally off topic. Spent 7 hours tossing and turning last night and one our having nightmares.)

-methods
 
Another (or perhaps the same)

0x5185 Alg Out Off with FS

sigh....

-methods
 
Common Errors:

* Programming a new features but forgetting to expand the channel count to the correct count... i.e. adding a digital line but not telling it there is an additional digital line. This can be especially tricky because you can see your input working in the Raw Digital... but no change in performance occurs when you toggle. Make sure you change the count setting over the input line selection.

* Fooling with anything you dont have business with. Listen here son... the LAST THING you want to do is go into your Sevcon DCF and "clean up". Believe you me... I dont know why or how... but there are some bugs and they are nasty. If there is some weird number in some weird spot... like "throttle 2"... something you dont use... DONT TOUCH IT. Just leave it. If it says 0.9999933456 great. Thats exactly what you want it to say... changing it may brick you for hours.

* BACK UP EVERY TIME YOU CHANGE ANYTHING. Get a really good routine going for saving off files with an order. As Delivered, As delivered mod01, As modified 01, As modified 02, As modified 03, As shipped 01... whatever... but save what comes in, what changes you make, and what goes out. Trust me... if in doubt revert to know good.

* When flashing large DLL - like 66... or doing anything... I set the CAN speed to 500KB if it was at 125KB. Dont know if it speeds things up... but DLD flashing takes FOREVER and doing it 1/5th as fast ... even the threat of that... makes my eyeballs cross.

and then De-Bricking when the F*&KING DLL transfer hangs up for 20 minutes...

:mrgreen:

-methods
 
And again... when you brick up (solid black) due to a failed DLD upload:

Turn off controller
Type bd into the DVT interface and press enter (just the two letters bd)
Turn your controller back on... now it should be in BootLoader mode (apparently this holds it in bootloader)
You can then go to Software/Program to try and flash again.

Watch your CAN light closely... this is really your only indicator of DLD progress.
DONT TOUCH ANYTHING till it goes black for a few seconds.

Remember to get out of boot loader and to power cycle before trying anything else.

66 DLD takes a lot longer to load than previous versions due to the lookup table.

-methods

P.S. If you cant type bd... then send this over CAN: 0x7fa 0x48 0x65 0x79 0x73 0x6c 0x69 0x70
 
Error 2911 Sub 1 or sub 3

While trying to change Footbrake priority over drive (to implement variable regen and full power...) you will see this fail error if you dont first go into PreOp.
To diagnose random errors from Sevcon:

Grab the first number... like 2911
Grab the sub...
Save off your DCF file

Open your DCF with WordPad or anything that will carriage return line feed

Search for 2911
First hit tells you how many subs... in mine... 15

Scan to the sub 1 (in this case) and it can be seen that the controller MUST be in PreOp in order to change this setting.
Common error.
Make you scratch your ass... then scratch your head... then scratch your ass... then while eating a sandwich... it all comes back to you :shock:

-methods
 
While testing on the bench you will see all sorts of hell coming out of the motor... this is residual PID crap.
Ignore it.

Like... if you are freewheeling to 5000 RPM and hit variable regen... it will stop short and run backwards.
Meh... artifacts

Load the motor down by testing it on a dyno or GoKart (NOT on a motorcycle... and WITH proper E-Stops...)
I think most of the buggyness will go away.

This may be true for Reverse settings as well... in FreeWheel when I set reverse as low as 100RPM I get very poor behavior.
I set it up closer to 600RPM... which I figure is about 100mph @ 6000RPM... so should be ballpark 10mph... which is reasonable enough.

I can ride a motorcycle backwards at 10mph... bet you cant. :mrgreen:

-methods
 
If you have a ClearView... eh... chuck it out the window :evil:

Ok... they are sort of awesome in that you can fiddle a few settings on the road... but tough to work with.
Really... like the DVT software... just not ready for prime time. Its like they WANT it to be tough to deal with.

Right now I am physically remapping wires because it is so difficult to re-map in software.

Tell me there is a magic GUI someplace for programming these things ?!?!?
If so - I dont hate them.

EDIT: Got DVT working for GUI editing by setting NODE to 11 so ignore rant below... simple frustration around working with "OEM ONLY" stuff... :roll:

If in fact this is the real deal.... like... hex codes with an up/down button... come on.
My senior project in undergrad was more sophisticated than that.
G
U
I

GRAPHICAL USER INTERFACE

Like... at a PC... with a keyboard and stuff.
Even a DOS prompt.
How long does it take?
15 minutes? Maybe 15 hours?

grrrrrrrr

-methods
 
Drive Wizard – PC based configuration tool

Is that the configuration tool for programming ClearView?
Tell me it is so...



-methods
 
Ok.... found a spot in DVT that allows changes to the ClearView... sorta...
Its not clear in the documentation... and it is not accurately reading my current display... but it is giving me choices like Arch... which is on the right path.

I was just looking for a way to download the config for the screen... like... downloading a DCF
Then at least I can text edit...

No luck so far.

-methods
 
I made some inquiries.

One can "save off" a Clearview configuration.
I need to put in some hours to figure out how to do this.... or wait to hear back from Sevcon.

Currently I am communicating with the ClearView through a Sevcon (on the same CAN bus). I may need to remove the Gen 4 from the buss and have DVT target only the ClearView... either way I was warned that they can be bricked so I am moving slowly.

If this pans out.. folks will be able to get in and out of PreOp, clear and view errors, tune basic parameters... all without having to have a proper DVT install + IXXAT. I will share anything critical that I learn in regards to up-cycling legacy gear that is out in the wild.

Electric Motorsports was not at all helpful. We have 2 units that they originally sold... they blew us off. I will not pursue further partnering with that company as things sit.

Thunderstruck-EV was actually helpful and even accepted a phone call. That is a company I respect for their efforts. The took the time to post manuals and do a good job of supporting advanced enthusiasts. Yes their prices are a bit high... but your effectively paying for Engineering Support.

I will have this thing dialed in a few days. Just not billable hours so... UGH.

-methods
 
Got it... I was not properly addressing the node ID... which for ClearView in my case is node 11.
Once doing so - I can see the example screen (which is skewed but close)

Clear001.jpg



This kind of frustration and grumbling boils down to not having access early on to Sevcon Materials... They push a heavy "OEM ONLY" attitude.. so in the beginning I had to learn the hard way. I have still never called them a single time. Causes a bunch of confusion and hostility :idea: Being supportive of people like me helps the cause...

A simple 1 page official cheat sheet would alleviate a lot of unnecessary customer support, complaining, and issues.
Looks like I need to go search all my rant threads and write it up :roll:
Sweet... more non-billable hours.

And Sevcon COULD just make all of their documentation Public.
Its one or the other... Keep it secret and deal with endless support issues... or make it public and let us solve it on our own.

So simpo...

Now to program the display to our liking.
Step one is clearing off the splash screen of the company who originally sold it then gave us the cold shoulder on support. :|

-methods
 
Procedure: Bring up a ClearView device so you can control your Sevcon. You can get in and out of Preop, acknowledge errors, trim settings, check digital switches... lots of cool stuff. Read voltage, curent, speed... fun toy... but it takes some setup. If we find a paying customer we will dial one it. Un-dialed... they are hit and miss based on how they were previously programmed!!!

* Power Clearview with 24V from pin 28 referenced to Batt negative
* Attach CAN pass-thru wires and REMOVE can terminating loopback

Note: Once you remove the loopback.. your CAN will not work without the display in place... so make a jumper plug that goes in place of the display with a 120ohm resistor.

* While attaching the display... there is a wire that needs to be shorted. This enables internal CAN termination.

Note: RESPECT PROPER CAN BUSS ARCHITECTURE. There is no such thing as a "Y BUS"... or "W BUS"... only a daisy chain. Dont do a star configuration... dont do anything lame. Just daisy chain plain and simple! Terminate both ends.

* Once display is attached it should power on with your normal Sevcon key-switch procedure.

* Jump into Menu and confirm your node ID

* Open DVT, Open H (Helper). Navigate to CAN tab.

* Click on Setup Display (under CAN tab in Helper)

* If you did it right... useful things should pop up.

Then navigate over to Thunderstruck-EV to download the User Manual and Quick Start Manual. Remember.... if you are using ThunderStruck-EV resources... BUY SOMETHING FROM THEM.

Shit aint free. frock off if you think it is :wink:

-methods
 
Looks like you can configure 10 unique objects.

In our case:

MPH gauge
Bar for amps
Text for amps
Bar for volts
text for volts
Error indicator
F/N/R indicator
Motor temperature
Odometer is on there too...
BItmap logo (for parking brake??)

Whatever - cleaning up the mess that is on this thing.
Old school array programming

-methods
 
Very interesting with the display programming. I did it all by hand on my unit using the display a couple years ago. Might give this a try sometime soon.

If I'm understanding correctly, you are using the DVT software to program what displays and where/how big it displays on the screen?


Sent from my iPhone using Tapatalk
 
Not successfully yet... :mrgreen:

Working on Variable Regen and Reverse mods today.
But yes... thats the idea.

The buttons and debounce on the buttons is dismal
Operating the display by hand makes my temples thump

Saved off the DCF yesterday... lemme have a lot at what I got...

ACK!
Yea... I got it... here is a snapshot or two

Its the Clearview DCF

[FileInfo]
FileName=C:/Users/methods/Desktop/ClearView01.dcf
FileVersion=1
FileRevision=0
EDSVersion=4.0
Description=Clearview Display
CreationTime=03:37PM
CreationDate=07-20-2017
CreatedBy=Tech/Ops SEVCON Ltd
ModificationTime=03:37PM
ModificationDate=07-20-2017
ModifiedBy=Tech/Ops SEVCON Ltd

Example of FS1 entry

[2123]
ParameterName=FS1 switch
ParameterValue=0x00
ObjectType=7
DataType=1
AccessType=rw
LowLimit=0
HighLimit=1
PDOMapping=1
ObjFlags=0x1
;SEVCONFIELD OBJECT_VERSION=1
;SEVCONFIELD OBJECT=DIGITAL_SIGNAL_IN
;SEVCONFIELD SECTION=TractionDI
;SEVCONFIELD NUMBER_FORMAT=2
;SEVCONFIELD 0=Inactive
;SEVCONFIELD 1=Active
;SEVCONFIELD MASTER=TRUE
;SEVCONFIELD CATEGORY=MONITORING


As to specifics

[5000sub1]
ParameterName=Access Level Indication
ParameterValue=0x04
ObjectType=7
DataType=5
AccessType=ro
LowLimit=0
HighLimit=5
PDOMapping=0
DefaultValue=0
;SEVCONFIELD OBJECT_VERSION=1
;SEVCONFIELD NUMBER_FORMAT=6
;SEVCONFIELD 0=Not logged in
;SEVCONFIELD 1=User
;SEVCONFIELD 2=Dealer
;SEVCONFIELD 3=Production
;SEVCONFIELD 4=Engineering
;SEVCONFIELD 5=SEVCON

No thanks on setting that to 5... Level 4 is already enough to ruin your day real quick.
Pandora's box guys. Live and let live :shock:
I wouldn't fuss around in there. Just be happy it is reliable and configurable.

-methods
 
Nuts&Volts said:
...If I'm understanding correctly, you are using the DVT software to program what displays and where/how big it displays on the screen?
...

Now I can answer a definitive YES to that question.
Very easy

Once targeted with DVT...

I went to the Speed Indicator
Selected Left Motor RPM
Pressed Write

And... my display now shows RPM instead of Speed.
I can now save off that DCF for future use.

Sweet.
This thing only sucks if you have to deal with it in the dark. Using DVT with it is tolerable and in an 8hr day I could dial one to spec.

-methods
 
Hello fellow Sevcon peeps,
I am working on characterizing my Sevocn, single motor set up. I have one error that is holding me back from entering drive mode.

0x5051 Line Contactor o/c

I am running a 24v Gigivac Gx 23ccv contactor.

Anyone experienced this error and have any ideas on how to approach it?
 
That means over current on your contactor coil

This is because the Sevcon is PWM'ing your pack voltage and sending that to your coil.
If it is a raw coil... it works... as say your pack is 100V and the coil is 24V you send a 25% duty cycle... you get ok voltage

Your contactor has built in supression
http://www.gigavac.com/sites/default/files/catalog/spec_sheet/gx23_0.pdf

That means that it is likely shorting out these noise pulses (PWM pulses) which are coming in
This triggers the (what is it? 7A?) limit for coil drive.

You have many options:

1) Different contactor with a bare coil (no supression, no economizer)

2) Contactor with a coil rating closer to your pack rating (you can sneak in under the supression)

3) Intercept the signal with some sort of DC-DC - so the chopping coming out of the Sevcon gets buffered - then drives the coil

You may be able to clever your way out of it... but I would put the contactor on Ebay and buy a new one.

I would also stick to 24V since they are one of the most common contactors available.
They are no more or less efficient.. for the most part

Yea - you can go with a coil rating that is the same as your battery - but those are often special order or otherwise hard to find.

So
There you go

A 4th option would be to "catch" this signal coming out of the Sevcon and then Drive your contactor*

You could do that with a common relay - then have the relay drive your actual contactor coil
I do not advise doing silly things like this as it can result in lower reliability

Your contactor MUST open when commanded
Even more important... your contactor MUST stay closed when it is supposed to
It is life critical

Good luck brother

Kilovac is your friend - ebay is your friend, look or "new old stock"
You can tell by the leads and packaging.

-methods
 
To be clear
If nit picking
In this very particular case
It is more reliable to run a coil that is of a voltage that is close to your pack

This lowers the stress on the PWM fets (one top that stays on, but the bottom nchannel switches)
They are small fets and ... if they fail... your system will fail.

The reason I say it is more reliable to run a coil the same size as your pack FOR THIS ONE CASE....
Is because that architecture was intended to max out at something like 96V
Zero helped push them up to 120V and ... the design is falling out of favor at this high of volage

It was really intended for something like 48V or 72V

Zero runs coils at pack voltage
It appears that they switch the coil with the MBB using an Nchannel

Ballsy move if you ask me...
As the Sevcon is under TIGHT CONTROL

Lets hope Zero did their due diligence in hardware validation, system validation, firmware change control...

When I was there it was fast and loose.

(Yes I am going to keep hammering on them till I see proof that they take responsibility for past risks taken)

(this is not fight club - we dont do the math and decide not to recall. We recall and fix.)

-methods
 
Back
Top