Speaking as an electronics engineer, I think it's unlikely to be a software failure as you would probably have been able to replicate it. It could be any number of random electronic failures that just happened to occur when you either unplugged the front derailleur or reconnected it. You can think of electronics a bit like an old light bulb - it's only guaranteed working immediately prior to switching it off, there's no guarantee that it will be working the next time you turn it back on! Interestingly, component failures tend to happen at the very start of life (within minutes of first use) or at the end of life (after many years). Once something has been working for several hours then it tends to work for a very long time. The failure probability plotted against time is what's referred to as a bathtub curve. Anyway, it's hard to say looking at just one side of the PCB, but chances are that the vast majority of the functionality (not that there is much tbh) will probably be done by a dedicated ASIC (application specific integrated circuit) taking care of processing, motor drive, bluetooth communication and wired communication. Alternatively it could be a microcontroller and a few support components which look after each function separately. Can you read the part number on the square black chip with the multiple pins? Judging by the 4 conductors in the interface cables the system uses separate power and probably some implementation of CANbus or maybe some L-Twoo proprietary 2-wire communication network, as opposed to Shimano's power over CAN that uses just a single pair of wires for both power and comms. Given that the only wired communication going on is between the front & rear derailleurs, my gut feeling is a failure in the communication. Whether that's part of the ASIC or a separate comms controller who knows? You say that the FD buzzes when you plug it in, so that suggests that power is probably OK, so comms is the likely culprit. If the communication is by standard CANbus then the next step is to put a CAN sniffer on the data lines, or just look for signal activity using an oscilloscope. If the system is CAN based then you should be able to measure about 60 ohms resistance between the comms wires when everything is connected. If it does measure 60 ohms then this is a strong indicator that it is indeed CAN based. If it is CAN based and the system uses a microcontroller as opposed to an ASIC then a repair might be possible because there will be a separate CAN transceiver, but If everything is done on an ASIC then forget it unless L-Twoo are willing to send you the part (unlikely). Diagnosing and fixing small embedded devices like this is notoriously difficult. If it's a failure of either an ASIC or a microcontroller then you're onto a loser because either device contains the embedded software, you can't just replace a microcontroller and expect it to function without first 'flashing' the software to it.
@@jaro6985 How do you know it's a NRF53832? If it is then I guess they could be using I2C to talk to the FD but I doubt that it would be very robust compared to CAN.
@virtouso47lol73 I'll be adding a new ratchet to add a gear to my mechanical SRAM shifters for the foreseeable future. But that's because I don't like how ANY hydraulic hoods feel as much as my mechanical shifters. Electronic works undeniably better though and with the exception of this VERY new product, electronic is definitely more reliable.
I love the fact that despite the 100k subs it still feels to be watching a long time friend nerding about bikes, continue like that Luke, thanks and congrats for your achievements and tha upcoming baby
This is such a lovely comment, thankyou so much!! This is exactly what I am aiming for. I LOVE geeking out on bike stuff like this, and if the camera wasnt rolling, it is EXACTLY what I would be doing with my time anyway 😅
My bet, ESD killed an output on the controller in the rear derailleur that goes to the front one. It could have been back emf, or similar as well. I agree that weather proofing looks pretty good.
Yeah after reading the comments, i think you could be right. I suspect that I accidentaly triggered a shift when the FD was unplugged, which may have fried something on the integrated chip. Would be great to confirm this 100% however!!!
This kind of controllers have usually inputs guarded very well against that. Unlikely. More likely is that some already going bad solder point just gave up due to the handling forces.
Something simple to try is to make sure the 4-pin conncter on the RD is 100% passing signal. Take a digital meter that can do continuity test. Test the connector if the signal goes from outside the sealed connector to inside on the solder traces.
the connector on the RD looks to be shorted as the wires are exposed where it goes in the epoxied part. Check for continuity + check for shorts between the 4 pins.
@@wearbabe visually there is no short, sure he can test it though, as they had designed in a connector but then soldered by hand to save time or money.
So I have already tested the continuity of all the pins, and the correct 8V are being sent over to the FD, hence the reason it buzzes when being plugged in. I did attempt to read the signal being sent over, by reading a change in voltage, and these was a very small change, but i need to dive into this a little more. I really need to compare the signals being sent when shifting on both RD's. From reading the comments, I suspect the faulty RD doesnt pass the correct signal due to a problem with the ASIC chip, but it would be great to confirm this 100%
I feel like I just watched the conclusion of an ongoing murder mystery, only to find out the least likely culprit was the rear derailleur! Anyways, I'm glad you're back up-n-running on your bike! I definitely feel like this issue could have happened to any of us who partake in the AliExpress bike journey! Anyways, if you get a chance I'd love to see a review of the finalized Sensah HRD Empire and SRX groupsets which are now available. They seem to have been released under the radar without much fanfare. Thanks for the great content! 🥖🥖🥖
Thanks Jonathan! Yeah I woudnt have suspected the RD out of the gate either! As some other comments have suggested, I think I need to get an oscilloscope out, and compare the signals being sent from both RD's.... The investigation continues!
We don't love Luke only for testing the bike and groupset for the entire english winter, we love him because he has jewell eyeglass for us to check the minor possible issues: hats off for the extra effort! 👏🏼👏🏼
You need to do point to point testing on that pcb and trace the fault backwards. Did you put your probes across the resistors and capacitors etc? Put your multimeter on the signal cables and see if you are getting a signal and keep moving backwards on the board until you find the signal, you will see a short rise in voltage when you change gear.
That board looks pretty clean and the board may be coated to reduce oxidation. It's quite rare to get cold joints in SMD boards and it looks to be zero thru hole. It may be dual sided too so Luke need to pop that board out to check the underside.
Yeah this would be a good step. The board is tiny though, and i dont really have the equipment to check all the tiny resistors. After reading the comments, the most likely culprit is the integrated chip that controlls everything, and without a full testing suite, there is no way to check the problem. What I can do however is sniff out the signals being sent when shifting the FD. If i can monitor the signal with an oscilloscope, and compare the busted RD to the working one, I will be one step closer to hunting down the true culprit
So, our baby is 6 weeks now and the take is, you go for a car ride as two people, a crazy whirlwind happens, and three people come home. Its so good, all the best to all three of you. x
Would think LTwoo would want to get that Rear Derailleur sent back to them so they can get their techies to go over it with a metaphorical fine tooth comb, since if it's an odd software or signalling issue they'd presumably want to come back and say "found it..."
1 - I believe Ltwoo will wait to see if you find the origin of the problem (communication issue with board chip), be it from the disconnect or the (re)connect, after all, you happen to be one of their best "beta testers" 2 - With all of this, adding the fact that 105 Di2 is roughly in the same price range (when you consider crankset, cassete, chain), Ltwoo will have to bring the price further down in order to effectively disrupt Shimano and/or Sram electronic groupsets 3 - Last but not least, hope everything goes smoothly regarding the newest member of your family 😉😉🥖🥖🥖🥖🥖🥖🥖🥖@@TraceVelo
Those were some good diagnostics! I have to agree after having ridden eR9 for almost 1000km, when it works, the groupset is fantastic! I have had no issue at all so far. Unfortunately, I had cable issue before I installed the groupset. The FD cable was faulty, so I requested for replacement, which the LTWOO store did promptly. If LTWOO can improve on their QC, I have a big hope that eR9/eRX can challenge the bigger players and give us consumers more choice.
@@M.F.Hafizhan eR9 and eRX are practically identical, except some parts like the shifter paddle replaced with carbon for eRX. So, yes you can do 10 to 12 speed as you can also do on eRX.
Spot on! When it's working it's absolute class! Glad to hear you have had a good experience with your set, and you are right, if they can tighten up their QC, they will really be onto something!
Seriously impressed by the lengths you have gone to to get to the bottom of this. I really want to know if any of the suggestions in the comments will now finally unravel the mystery. Assuming you figure it out, I think you deserve a lifetime of L-twoo products for your work in helping them get a major issue sorted. Nice sign and good luck with your imminent arrival!
HA, thanks!!! Yeah after reading the comments, there are some super knowledgeable people dropping LOADS of good advice! Basically I need a way to read and compare the signals being sent by both RD's. I can do some of that with a multi-meter, but i really need an oscilloscope to get to the bottom of it!
Try bypassing the connector on the rear derailleur. Open it up and go directly from the cable straight to the pins on the circuit board. My money is on a connector failure OR a HW component on the circuit board, perhaps due to a surge or short
To date, I have traveled 800km with the Ltwoo Erx electronic group without problems in the system, but with some observations that need to be improved, even though I installed larger capacity 2800mah batteries, I cannot exceed 160km of autonomy, which is its greatest weak point. Since the electronic groups of Shimano and Sram have around 800km of autonomy, sometimes it does not change immediately, you have to press the button 2 or 3 times for it to make the change. It happens to me very often with the rear derailleur, although remembering if I have a problem Serious but with the front hydraulic brake, no matter how much you bleed it, it works for a few days and stops working.
I've made similar distance on the groupset. Battery holds string for me. At least 250km for me and i don't have larger capacity. I also had problems with my front brake. Bled it twice and i still had not the full brake force... at home i squeezed the brake lever for 20minutes and on the next ride it was gone. Since then i have full brake power!
I hope L-Twoo figures out this glitch. It's good that they honor the warranty, if true. Only thing I can think of is flashing the shifters with the software. This is different from updating the software but I wouldn't know how to do this without specialized applications. Also *maybe* check continuity on all the electrical but I'm not a CS or CE major so I couldn't tell you anything. I do like the water proofing on the derailleur. The screws are on the outside of the body and presumably that seal is in compression when fully assembled. Only thing I could suggest is check continuity from the signal pin from the derailleur. If the pins aren't corroded and you see a spike in the voltage from the single pin to ground, I am also not an electrical engineer either, then that means there should be signal going from the derailleur through the wires. If not then it must be a software thing, presumably. I'm only a mechanical engineer. I can design a bike, not the electricals. Edit: Thinking about it, in addition, do the same from the pcb side of things.
I'm kinda thinking it's the signal wire to the pcb. Like if the pin wasn't seated in the connector and when he put in the cable it pushed the pin out or something. Or cold solder joints broke.
Thanks for the suggestion BeyReaper!!! So i have already checked the continuity of the cables, and everyting seems to be fine. The fact that the FD buzzes suggests it is getting power, and after getting out my multi-meter, i confirmed the correct 8V is getting across. What i am not sure about however is the signal to shift being correctly sent. I think i need to get out an oscilloscope, and sniff out the signals being sent. Still more work to be done i think! 😅
The fact that L-Twoo have not asked for the original to be returned for the fault to be diagnosed suggest that it is something they know about and have already fixed. If you have the know how and the tools to read and write to the electronics, you could probably copy the software on the new one to the old one and see if it will work.
@@mazditzo I'm pretty sure that's not it. The smaller, but still significant, Chinese companies I've worked with are far ahead of the curve when it comes to caring for their reputations. Trace may be well known but we can all pitch a fit on any youtube video that mentions this brand and we can all comment and share every time someone says L-TWO. Reputation at the beginning can make or break a company. It even kills their direct to brand sales if their groupset has a bad reputation.
That's electronics. I had to replace a SRAM red etap rear derailleur once and luckily under warranty. It just stopped working. That was years ago and it's still working perfectly. Hopefully the new RD from L-twoo will also work for you for years and years.
(another electrical engineer here) interesting to see that the wires coming from the connector are just directly soldered onto the board instead of using a connector, which was clearly the original idea since there are soldering pads and component outline for a connector socket (JST style). the soldering itself on them looks quite bad to be honest, hope its improved on the production units, or better yet they don't skimp on those couple cents and actually populate a connector instead of directly soldering wires.
Luke, great diagnostics and I love the new sign! ❤ 🥖 Good luck with the growing family! So the RD is the culprit. 🧐Assuming the shifters talk to the RD (not the battery) and the RD then talks to the FD through the battery, it seems reasonable that the RD entered a protection mode where it simply won’t command a FD movement. Perhaps there was a message exchange going on when you unplugged the FD that triggered the lockout. Shimano has that same Crash mode type function that needs to be reset to restore normal operation. Unfortunately, it would appear the app team forgot to include that option, or there is a bug in the code. Is there a firmware update option that might simply refresh the firmware? I think you mentioned performing a reset to factory defaults already. As you are an early user, some firmware issues are going to be inevitable. Or, as an alternative, if the RD only ever talks to the FD over the cable, perhaps the RD transmit circuit was damaged when the cable was disconnected. Anyway to determine if messages are being sent from the RD to the FD? Put a LED on the TX/RX lines? Keep up the great work!
Thanks so much! So by holding the button on the RD for 5 seconds, you can "recalibrate" the position, so essentially reset everything, but i tried this to no avail. After reading some comments, I'm coming to the conclusion that it is probably the fact the signal just isn't being sent, maybe due to a problem on the integrated chip, or possibly a popped resistor.... I think If i get hold of an oscilloscope, i will be able to sniff the signal being sent from the functioning RD, and compare it to the busted one. Def more work to be done i think 😅
When you show the inside of the electronics there are four wires that are soldered to the circuit board. The furthest looks like it is only just making contact with the solder. Try reflowing this joint. I also love your new sign 👍🌞
At 9:19 in the video you show a close up of the 4 control wires. Green, Blue, Black and Red. The Blue wire looks to have a cold solder joint and needs to be re-soldered. I'd also hit the green wire too as it looks a little poor. I would guess the green and blue wires were for a data line which would explain why the derailleur wouldn't respond but still have power. Also at 9:03 where you show the wires coming in through the epoxy, is that red wire exposed? I wonder why the epoxy didn't come up to the insulation.
3-4 weeks ago thisa was in the comment section as a reply ... Have you tried this out ? Hello, the same thing happened to me in particular but in both pieces, both the derailleur and the tensioner, this happens when there is an unusual effort in the system,It generally occurs when something obstructs the path of movement, thus producing a protection that remains completely immobile,The solution is to constantly hold down the tensioner button, which resets and sets both pieces to 0, It happened to me several times and I performed this operation every time and it was solved without problems.
@@abedfo88 i was hoping he saw that way earlier now he made 2 more videos after the incident but i guess he knew that the fixxed one will be a giveaway ^^
It can be seen from your video that the values of the two gears of your FD are 10 and 94, which may be because they exceed the range of movement of the FD. You can try to use the APP to connect to the RD, restore the factory settings, and then press the function button on the RD to reset. Wish you success.
My bestie just built a new bike using all L-Twoo components. I’m his bike guy so I’m just learning the ropes on all L-Twoo. Thank you for this deep dive into what can go wrong with the e-shifting should something ever go wrong. So therefor, I’m a new subscriber and follower of your great channel. 😇
I am using L-TWOO eRX too, not sure if you all have the problem on shifting the FD when you are on a very very very slight uphill slope. I am using a 50-34T crankset. Couldn't shift to 50T from 34T when I am on a super slight uphill slope. It shifts perfectly fine if I am on a flat road or downhill. But uphill is the problem. Can see the FD is moving up and down but it couldn't push the chain.
Thanks for your videos about this groupset. I was considering it for an upgrade to my road bike (currently running mechanical Ultegra R8000) but given your issues and the fact that 105 Di2 isn't that much more expensive, I think I'll be going to Shimano route when I go to upgrade.
Yeah right now i cant reccomend dropping £550 on this, plus whatever else HMRC adds on top (it was an extra £140 in my case). 105 Di2 is a much more sensible choice. It's only a few hundred quid more at the end of the day, and it's _pretty much_ guaranteed to work.
If there's the chance do flash the firmware and hard reset the internal memory, this might fix the issue if it's the software that entered a bad status. Otherwise I think the only chance is to send it back to Ltwoo and ask them to do the diagnostic - unless you're an expert in reverse engineering of embedded electronics...
probably the messages from the controls/apps are all received by the rear derailleur which then sends the commands to the front. I see that the connector has 4 contacts, probably 2 are power and two are for data transmission. You should sniff out the signals that come through and then try to interpret them. Going backwards perhaps it could be a simple false contact of a weld
Nice wife's gift and Congrats for the little baby girl! as a father I am, blessings!🤜🤛🙏 About the FR dead, I think the problem solves resenting/upgrading the software, the thing is how! and maybe changing the green electronic card if you find it
Congratulations on the pending arrival of a Baby Velo ,i would think that L-TWOO eRX company are learning from watching you Luke sorting there group set .
thanks for the deep dive into this issue, nobody else is looking this closely at chinese groupsets. waiting on your results to see if it will work for me.🥖🥖 congrats on the family!
You need to check everything on board. With a multimeter. Not just look at it. You can not tell just by looking. You can have a bad connect but everything looks good
This reminds me of the old game Cluedo - "it was Colonel Mustard in the kitchen with the candlestick" etc. But this time the original murder victim turns out to be alive and the real victim was only pretending to be alive. Nice detective work, Luke.
It seems to me that battery is just a battery and only provides juice to both front and rear derailleurs, and the rear transmits command to front with some kinda CAN bus or RS232. So I would say it's not software at fault, but probably something like driver dead - explains why rear still works fine but not communicating with front. Also explains why front buzzez when you plug it in - just regular initiation happens. So would need to diagnose it with oscilloscope and similar debug tools. May be try to collab with someone who makes video about electronics to figure the real issue.
Exactly. Probably some implementation of CAN, and probably an ASIC chip doing everything, so basically unrepairable other than by the manufacturer. However, given that they will have no more than $5 cost in the whole electronics they wouldn't bother warming up their soldering iron LOL.
@@thegearboxman I don't think it's asic, more probably some general purpose arm based controller. Repairable if it's a driver (since it's always a separate ic) or firmware is not locked.
It would be cool if you could use this as an opportunity to show us what navigating the warranty process for one of the more reputable chinese manufacturers is like
If the old components are operating with the new shifters but not the old ones it is quite obviously the front derailleur shifter not talking to the component. I think you will find that it is in the shifter, bluetooth not communicating with derailleur, maybe try putting new battery in shifter.
Great video Luke! Great content. It's frustrating that you couldn't get to the bottom of the issue. That rear derailleur really threw a baguette 🥖🥖 into the works. Good luck with the new arrival! 😃
Yeah after reading through some of these comments, i really need to spend a bit more time on this. I recon i can sniff out the signal being sent by the working RD, and compare it to the faulty one... There is more work to be done 🤣 And thankyou for the well wishes, she is 6 days late now 😅
There's a "Protection Mode" in the App that when turned on prevents the FD from working under certain circumstances. In my case, The FD would not shift when I'm riding outside, but when I'm home and trying to replicate the issue, it shifts normally. So if anyone else is running into the same issue, go to the app and make sure that Protection Mode is off in the settings.
As a software engineer I think this is unlikely a software problem. It's firmware is usually on read only memory that can only be updated with a special process. It doesn't take much water at all to short the circuit board, it may not be visible to you..
Another software engineer here agreeing with this. I am also confused why he says he didn’t figure it out, seems pretty clear to me that the rear derailleur went bad. He just couldn’t see the issue, but it’s definitely there. Also kind of surprised the LTWOO doesn’t want it back to diagnose it and make sure it wasn’t any issue on their end that they could improve on.
At least they could have the derailleur sent back, faked a diagnosis, then said something about the origin of the issue. In terms of image, would have been better with no additional cost.
To me what screams “software” is that many others across the globe had the similar issue at the exact same time, within days. I wonder if LTwoo pushed out an update and something in that update caused a irreparable malfunction. Just my $0.02. I am not a engineer of any kind though, so just a finance guy making speculations. Lol.
@@dbk81 In order for an update to be pushed, the deraileur would have to have be able to receive a cell signal or connect to the internet via wifi. Also if it's a cell signal someone would need to pay for the cell service access. Once could manually update the firmware but it sounds like Trace didn't do that. I still don't think this is a software issue.
@@mikemurphy1856 Good point. I thought it might be possible that when he connects the companion app to them that they could access any push updates. But again, unlikely.
After watching your videos I've decided... Shimano 105 groupset from ali for my Specialized Allez, lol. But I have to applaud your skills and its fun learning how all this works. Cheers.
I just installed my GR9 hydraulic groupset and I am disappointed. The shifting resistance is appauling. I used ltwoo's supplied cable and a jagwire housing. The cable looks good quality. Copper-colored so maybe it's coated with something. It's definitely polished and smooth so maybe it's the cheap housing or the fact that I routed it on the front side of the handlebar where the horizontal bent is bigger forcing the housing to bent more. The levers have another groove to route the housing on the back side of the handle bar for a less bent of the housing so I must give it a try. But the worst thing is that the left lever bleed port screw is tightened so hard that it won't budge. I stripped and rounded its head trying to loosen it and now it's stuck forever. Suprisingly I was able to bleed the system using just one syringe plugged into the caliper so from top to bottom. But anyway I am waiting for the seller's response and his solution becasue this lever qualifies for a replacement. Other drawbacks: 1. Rubber of the hoods is slippery although it has a matte feel to it and has a texture. Shimano's hoods are way more grippy. 2. To save manufacturing costs, for the left lever they use a lever from a 2x groupset (maybe rx) because it has guiding groves for the shifter housing under the hood. They simply just don't put the cable ratchet pulley and shifitng paddle and trigger. That's ok but without this ratchert, there's a gaping hole underneath the lever where all the gunk and dust may build up. Also while on the hoods having your index and middle fimgers wrapped around the hood, you can feel that hole and it's edges under the fingers tips which is uncomfortable. Especially when you grasp firmly while pedalling hard out of the saddle uphill. How much would it cost them to desing a simple plate cover with an i.e. elastic latch similar to those of battery compartments in tv remote controllers? 3. Downshift trigger is in awkward position. Its position and overall lever shape is heavily inspired on Campagnolo Ekar levers which apparently have the same problem which Paul from Mapdec Cycles reported in one of his reels. On the hoods to reach it you have to sort of lift your hand and twist it inwards. also when pressing down the trigger, if you have your fingers wrapped underneath the lever, they are right under the trigger and obstruct its downward travel so you also have to move them away. I would expect the trigger to be place more towards the front of the lever just like it is in ltwoo's mechanical road levers like r9, r7, r5... Also, it offers up to 3-4 downshift with single push. I think this is unnecessary, because it causes excessive unwanted downshifts. It could be justified in the road RX groupset in finish line sprint situations, but this is a gravel groupset so I think it's pointless and useless. On the plus side, the trigger is ergonomic and right were you'd look for it while on the drops. 4. Lever instllation is troublesome. The bolt clamp is located in the center of the lever body so the vertical part with the master cylinder obstructs access to it to an extent. Also, the bolt is very hard to reach becuse the rubber of the hood is very tight and does not peel off enough to expose the bolt. You need to have an extended hex wrench with a ball-head end and prefferably with an attachable handle. With a standard-sized wrench it will be very hard to reach the bolt. For comparison, in Shimano levers the bolt is on the outer side of the lever body and it only takes to slide the wrench under the hood to reach it. 5. The rear derailleur isa bit too big. I've installed it on an Accent Furious Pro frame and noticed that if I loosen the b-tention screw too much, the front of the derailleur where the clutch is located touches the chainstay and wedges on it preventing normal operation. Also, the part of the derailleur with this clutch is visibally made out of some kind of composite material, probably based on nylon. It's definitely not metal. The rest of the parts, like the pantograph, pulley cage are indeed metal. Unless you buy a version with a carbon pulley cage of course. Pros: 1. Overall quality is good. The logo imprints are good. They are thick and have a nice sort of brushed aluminum texture. 2. The shape of the hoods and lever blades is ergonomic. They feel good under the fingers both on the hoods and on the drops. A this point I honestly do not recommend GR9. It's not worth the risk. I can recommend R5 because I have had these levers for two years now and they have not failed so far although the shifting resistance is also quite strong but not as strong as with these GR9 eventough I have the shifting cable routes along the front of the handlebar as well . The overall build quality and shifting is comparable to Shimano Sora..
Plugging it back in may have put a bit of a surge through it. You could try and see if there is a short on the board. You need to look at the back of the board really as well
I thinking that with the ability to support 10-11 rear cassettes one could setup an awesome drive train with reduced cross chaining, lower cost rear cassettes, less stress in the chain, greater tolerances for the rear cassette spacing for the chain than a 12speed rear cassette. Over all the drtive train would cost less and last longer while maintaining a very good range of gear ratios. Looks like I just designed my next bike drive train. FYI MTB rider.
Argh frustrating. Sometimes electronics *shrugs shoulders* I have literally no authority in the matter but something may have fried when you unplugged the cable. My boy was a week late, then it all happened super quickly! Good luck to you and your wife luke.
Hey like could be bad ESD protection on the output motor driver in the rear derailleur. If you want me to do a diagnose with my oscilloscope or logic analyzers... let me know I do need a working and not operational specimen to do some comparisons.
Those solder joints on the wire to the PCB don't look great. A bit too much solder on some of them. I'd reheat those and see if it fixes the problem. There's a bit more wire exposed than is strictly necessary too so it could be that they're shorting when the casing is in place but they're happily separated when you pull the housing apart. Using some hot melt glue over the top of the solder joint before re-assembly would sort this out
