Cycle Analyst Repair

amberwolf

Administrator
Staff member
Joined
Aug 17, 2009
Messages
40,479
Location
Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion
file.php


I didn't find any single good thread with CA repair info (just a few here and there discussing mods/etc, with a little component info and such), so here's one to combine whatever anyone has that's useful for those wishing to fix their own.


In my case, it's a donated one from Dogman and Justin of Grin/ebikes.ca/Renaissance, which via some mixups at the board-build-factory on component specs didn't do so well at the 80-something-volt level, when Dogman got it and hooked it up. :(
DSC02961.JPG
That's the back of it, next to the mount.
This is the inside of the back cover:
DSC02959.JPG

Justin allowed Dogman to send me the defective one, while sending him a new one, so that I can try to fix it and then use it for my projects if I can make it live again. :) The box it was shipped to me from Dogman was on the porch when I got home tonite, so after taking care of the dogs, of course the first thing I did was open it up. ;)
DSC02946.JPG

Looks like a few toasted parts;
View attachment DSC02947.JPG
two of the resistors on the LCD, (R8 and R9)
View attachment 4
the main cap C2 and transistor Q1 on the main board,
DSC02950.JPG
plus a diode (D2) on the main board.
DSC02948.JPG
Could be more, but those are fried and/or blown apart. All so tiny, gonna be a challenge to remove *just* those parts and replace. :lol:

Q1 is supposed to be NJD35N04G, so I'll find that or a matching-specs one. Darlington NPN 4A 350V, 45W DPak form factor.

C2 is a tantalum capacitor, which if I can remember how to read the markings correctly is a 1.5uF 10V (157A), based on this image from Justin:
file.php


I don't yet have any idea what D2 is, or the LCD's R8 and R9. I'll be checking a couple of other threads and then the ebikes.ca site first, then if I can't find it on my own I'll ask Justin what the exact part values are so I can get them off other things I have around here (I probably have them already), and see if it still works. If not, I'll ask him what else might be cooked and start replacing as needed. :)
 
amberwolf said:
Could be more, but those are fried and/or blown apart. All so tiny, gonna be a challenge to remove *just* those parts and replace. :lol:

It's just those parts. The sequence of events is that the knock off transistors allows a higher than spec'd voltage to get onto the tantalum capacitor, that causes the cap to fail and draw excessive current, and the excessive current burns up the resistors in series with the LCD LED which eventually act like a fuse and fry.

Q1 is supposed to be NJD35N04G, so I'll find that or a matching-specs one. Darlington NPN 4A 350V, 45W DPak form factor.
Pretty much any NPN transistor rated for over 100V in a DPack form factor will work, doesn't have to be darlington so long as the Beta gain constant is over 100. The ones that caused failure here actually had a measured Beta value of "1" on my multimeter, so were basically non-functional it would seem.

C2 is a tantalum capacitor, which if I can remember how to read the markings correctly is a 1.5uF 10V (157A),

It's 150uF. You'll need at least 120uF or so for all the data to save when the power is shut off. You can run with no capacitance here, but when you turn power off and on, nothing on the trip stats will be remembered.

I don't yet have any idea what D2 is,

D2 is just to protect against reverse polarity. If you don't plan to plug it in backwards, you can short across it and everything will work, but if you ever plug in power reversed things would smoke.

or the LCD's R8 and R9.

These are just in series with the backlighting LED for the LCD screen, and hence are in series with the power input to the linear regulator. They serve to limit the inrush current when power is first applied. R8 is probably still fine, R9 you should replace with something around 300-400 ohms for regular operation, or 1kOhm if you plan to run mostly over 72V.

I'll be checking a couple of other threads and then the ebikes.ca site first, then if I can't find it on my own I'll ask Justin what the exact part values are so I can get them off other things I have around here (I probably have them already), and see if it still works. If not, I'll ask him what else might be cooked and start replacing as needed. :)

Sounds like you got a good handle on things. If you want to have fun, you could try a TO-220 style NPN transistor for the regulator and mount it to a heatsink, and then make a device that can handle the 300-400V applications like Doctorbass is inquiring about. In that case, there are a couple other resistors you'd need to change on the board too (the voltage divider, 220kOhm, will need to be closer to 1mOhm, and the 47K resistor in series with the zener diode will need to be more like 600 kOhm. )

Justin
 
Thank you very much for the info and notes!

Between work and weather, I haven't had time to get back to the CA yet; might be spending my day off tomorrow sick in bed rather than fixing anything, too. :( (been spending tonite on sudden trips to the loo).

I almost certainly have all those parts in scrapped stuff to fix this. I may end up using a TO220 transistor just because it is easier to replace if I toast it somehow, rather than to use for high-voltage (though I might do that mod eventually, too).
 
Regarding the transistor, see the link in my first post to the factory mixup issue. :)

Not sure about the R8 have to go check again. I've been sidetracked on too many other broken things plus horrible tick season for the dogs (at least an hour each day picking them off the dogs and me, and never get all of them).
 
Does this mean i just need to change my transistor to get my cycle analyst work? because i have the NTD version.

ps: my R8 resistor is 4.6Kohm
 
I didn't yet find the right transistor, but along with a prize CA,
http://www.endless-sphere.com/forums/viewtopic.php?f=3&t=21397&start=0
http://www.endless-sphere.com/forums/viewtopic.php?f=4&t=22962&start=0
Justin sent me a couple sets of the parts to fix this problem,
file.php

so I'll be fixing this one (which will probably then go on DayGlo Avenger) as soon as I get the chance.


BTW, anyone that doesn't say Justin and/or Grin / Ebikes.ca rocks must never have dealt with them. :)
 
Yup he rocks. You do too AW.
 
I became totally awake for a little bit, and decided I would try (again) to fix the first Cycle Analyst (from Dogman), but I lost my concentration and didn't get it working.

I carefully removed what was left of the transistor, capacitor and diode (D2), leaving alone D1 and the LCD resistors for now. Unfortunately the pads for the capacitor's and D2's anode both are gone, burned off the PCB.
DSC04032.JPG
Enough of the trace to D2 was there, close enough to the pad, that a little scraping off of conformal coat made a "pad" for it, and I could solder right to it, mounting D2 easily enough (well, except for my shaky hands and my crappy eyes--jeebus these things are tiny!).

The cap was harder. I found what I *think* is the thru-hole that used to feed the trace that went to the pad, but as neither the trace nor pad are there, all I can do is guess at the depressions in the burned conformal coat in that area. The thru-hole is under the cap body normally, so I can't solder the cap directly to it. Instead I used a bit of wire-wrap wire, insulation stripped off, to insert into the TH and then soldered that wire to the TH's outer ring, scraped free of CC first. Bent the wire towards where the pad used to be, then soldered the cap to the wire and the remaining pad on the cathode.
DSC04033.JPG
I can't verify electrical connection, as I don't know where the anode of the cap is supposed to connect to. I didnt' feel confident enough in my skills at the moment to open up the LS-CA to ohm it out, but will have to do that later, unless someone else here already has done this and posted it somewhere.
DSC04034.JPG
Transistor was easy, as the pads are all there with no damage or burning. Hardest part was holding it still while I soldered. :(

But the CA doesnt' appear to turn on. I *hope* that is just because of the unrepaired resistors on the LCD board, but I don't really know yet. Too bad the CA board doesnt' have a status LED to blink during power up. ;)


Anyway, I can't even see the little parts right now, eyes are too tired. Back to this later. :(
 
Not yet. :( worked on this today instead, less fine motor skills required:
http://www.endless-sphere.com/forums/viewtopic.php?f=3&t=16920&p=378025#p378025
 
I still haven't gotten back to the small screen CA; at some point I must have put the CA's internals and the casing/shunt/repair parts in two separate boxes, and have yet to get them in the same room at the same time, apparently. I feel bad for not getting this going, because not only could I really use it, Justin and Dogman specifically went to the troulbe of getting this and the repair parts to me. :oops:


In other news, the rain a few days ago
http://www.endless-sphere.com/forums/viewtopic.php?p=616747#p616747
seems to ahve done in the buttons on my LS CA that I won in one of the Grin Tech contests:
DSC07082.JPG
and neither button works at all now. CA responds to shorting the input pins on the PCB ok, so it jus tneeds new buttons. So I am going to take some front panel buttons from an old dead VCR I salvaged for parts a short while ago, and silicone them into the orginal button places, with a cable to the inside of the unit, then put the rubber button cap strip back over the top.

That should be mroe waterproof than the oriignal has turned out to be, at least for the very little rain we ever get here in Phoenix.
 
ebikes.ca has replacement button membranes available.

there was a problem with the original design of the membrane. which would lead to the problem that you are seeing now...
they issued a new button set to all existing users, im sure if you shoot them an email,
they will send you the new improved membrane.

Jason.
 
Yeah, but I'm not a paying customer and I would rather he spend his money/efforts on those, to keep him in business. :)

Plus I can DIY a solution easily enough, based on an idea someone else already successfully implemented on their LS CA for the same problem.
 
amberwolf said:
Yeah, but I'm not a paying customer and I would rather he spend his money/efforts on those, to keep him in business. :)

That's kind of you Amberwolf and there are lots of ways to hack a custom button fix, including making remote buttons near your handgrips so you can toggle and reset without removing your hands from the bar.

BUT, if anyone else has button membrane issues from the first batches of large screen CA's please don't hesitate to send us an email with your current shipping address and we'll have a new front enclosure fired off in the mail right away. This is really an intrinsic design flaw in the keypad to have the traces cross over each other like that, and eventually with prolonged humidity at the junction electrolysis will corrode the conductor away.

-Justin
 
I finally got the buttons fixed, although I didn't make the wires long enough, as I just used a bit of ribbon off the same board I got the buttons from. Also, I ended up with the buttons l/r swapped, whcih I will fix before I secure them down, when i replace the wires with some longer ones.

I'm tempted to do what Justin suggested with a remote button on the bar, but I think if I do that I will add a jack on the side (like a micro-sized headphone plug) and use a wired remote from an old tape recorder or something to add the buttons, if I should decide I need access to one from the grips.


I haven't siliconed the buttons in yet, to make them weather proof. The switches I used are sealed, but the wiring/connections are exposed, as is the hole the wires go thru.
 

Attachments

  • IMG_6492.JPG
    IMG_6492.JPG
    50.7 KB · Views: 7,171
  • IMG_6494.JPG
    IMG_6494.JPG
    19.9 KB · Views: 7,170
  • IMG_6495.JPG
    IMG_6495.JPG
    32.9 KB · Views: 7,169
  • IMG_6497.JPG
    IMG_6497.JPG
    34 KB · Views: 7,171
  • IMG_6499.JPG
    IMG_6499.JPG
    45.3 KB · Views: 7,172
amberwolf said:
I finally got the buttons fixed,

Ha ha, nicely done Amberwolf! I still think though that if you are going through this trouble, it really is not a whole lot of extra work to make handgrip accessible buttons. You can also have then wired in parallel with the membrane buttons that you've put on the enclosure, so both can still be present and active. There are extra holes in the back of the enclosure that you can use to feed out the cable with your headphone jack as connector.

-Justin
 
That's what I was thinking of doing, although I was going to see about embedding the jack into one of those holes, so I could just plug in the remote on whcihever bike the CA is on at the time, some of which might need the remotes and some might not, but without having to move the remote buttons, which may not fit on the bars in some bikes (or which could be built into a custom grip unit, for instance).

I don't know that I have a jack small enough to fit in even an enlarged hole, but it's worth a look if i have time at some point.
 
amberwolf said:
I don't know that I have a jack small enough to fit in even an enlarged hole, but it's worth a look if i have time at some point.

The jack itself doesn't need to fit into the hole, you can have it dangle out a little bit from the wire and still have the same basic effect. -Justin
 
I know, but I am looking to not have any danglies. :)

I'm also looking at all the other waterresistant connectors (or ones I can make so) that I have on various things, that I could mount back there for the shunt/etc, so I can leave all these things on the bikes and not have to unwire them from each one just to move the CA itself over, especially since some bikes will have controllers with CADP connectors, and some will use an external shunt, speedo, etc.

I've discovered taht I really don't like the JST connectors, because I have had numerous failures of them due to corrosion, even though things only get a teeny tiny damp and only on very rare occasions. Mostly it's been on hall connectors (which I finally started just soldering together), but it has also been on a throttle, an ebrake, and the CA's shunt connector.

So far all the ones i have are too large for teh case but I think I have one military "Canon" type taht has enough pins, and I might even have enough mating male connectors for the bike end. I'd just have to glue a box to the back of the CA case to shield it, and pass the wires thru the CA's rear holes.
 
So the V3 CA with Thun sensor arrived, and today I finally had the chance to try installling it...and I blew it up before I could even ride with it.

(Crossposted from the CrazyBike2 thread)
Having been too brain-dead to even solder wires to a controller here:
http://www.endless-sphere.com/forums/viewtopic.php?p=676199#p676199
and then also get so distracted I even burned dinner, I should have known better than to try to do anything else tonight, but no, I had to try to play with the new toy--the V3 CA.

After some hand, knuckle, elbow, forehead, and knee bangs and scrapes and dents, I managed to get the cranks off, and the old plain BB out (which was very difficult; the weather change to the cold pouring rain presently going on outside is making all my joints ache badly, so it's pretty hard to grip tools tightly enough). Then I cleaned out all the icky sandy grease out of the BB shell with ammonia and a brush, which was easy enough, but time-consuming.

I had a lot of trouble getting the Thun BB installed, though--it took another half hour or more to realize that despite the markings on the BB for "L" and "R", it couldn't possibly go in that way becuase the plastic caps are threaded the other direction from the BB shell on this bike. Maybe Magna welded the BB shell on backwards? I dunno. Anyway, so I installed the BB the only way it'll fit.

To install it's wiring, I had to take the plug off the cable, because the only holes in the BB are barely bigger than the wire. The only one I can actually use is the one in the bottom that's threaded for a screw that normally holds the plastic "brace" for shifter/brake cables, so that's how I ran teh cable out.

file.php

There is a little break in the outer jacket of the cable, with some damaged insulation on the wires, so I cut some heatshrink to slide over it, but when I put it on there I couldn't get the BB in--too tight a fit. So I cut the heatshrink off, and ran teh cable thru, intending to slip a new piece thru the hole over the break in the wires' insulation (which happens to be right at the hole itself, where it could potentially touch the metal of the bike, which is at pack ground). But apparently I got distracted at some point and forgot to do that.... :(


Then I clamped the CA to the steering extension stem tube, just below the stem clamp, so I could use both CAs at the same time and compare readings, etc., with teh old one as a standalone with speedo sensor, and the new one as a direct-plugin. It puts the new one a little lower than I'd like, but it was easier to clamp there than on the bars, as my bars are too thin for either CA's clamp to hold onto without adding something over them that's pretty thick, but the extension tube is just about right.

file.php


The new CA's backlight is much dimmer than the old, and more yellow-orange where the old is white. I'm not sure if I like taht or not; I'll have to ride with it at night for a while to see if it's really bright enough, in traffic with headlights in my face.


I used the 6-wire DP extension cable to hook up the controller's known-working DP port (already tested with the other CA recently) to the new CA's DP port, and powered the bike up without problems, and the new CA also powered up fine, booted, and I went thru the screens and reset things to match this bike. Moved the throttle cable from the controller's input to the CA's input, and verified the CA sees it and reads it's voltages correctly.

That's actually a nice feature to be able to see the voltage and the little throttle indicator on the main screen going up and down-- would make troulbeshooting throttle issues a lot easier.

Voltage reading of the pack is a little differetn on the new CA, but it is just as likely something like voltage drop across connectors, or something int eh DP connector vs the shunt connector as anything else--it's only 0.03V. The Fluke DMM reads 66.0V at the back of the andersons into the controller, where the CA-SA reads 66.1V at the shunt between the pack andersons and the controller andersons, and the CA-DP reads 65.8V from the wires inside the controller.

file.php



However, there is quite a gap between the CA and the connectors for either the Thun BB sensor or the controller's throttle input, so I dug out the JST shells and pins that Bikefanatic sent me in the last care package, and made up an extension cable for the BB sensor (5-pin), and hooked it up, then powered on the bike again to make sure this cable worked before starting on the throttle extension.

file.php




This is where stuff really went sideways. :(

At first it seemed normal, but nothing was detected from the BB, moving it either direction, with or without a load (holding hte brake and pushing on a crank hard, for instance). Then the CA screen flickered, and it rebooted, tehn again, then it went blank and no backlight.

I switched the bike off and checked the plugs; it was definitely not a loose connnection. Since it all worked until I hooked up the BB sensor, I unplugged that, and tried again, to no avail. Hooked up throttle back to controlelr and verified it still worked. Disconnected CA from extension and hooked directly to controller, no change. Same thing hooking it to the shunt off teh older CA.

So I disconnected everything and opened it up, hoping for a broken wire inside, but as soon as I cracked the case I could smell that fried odor. :cry:

Most likely something at the sensor wires shorted to the bike frame, grounding out something inside the CA that isn't protected from such a fault, because I forgot to put that stupid heatshrink over the damaged-insulation wires on the BB sensor cable. :(

Nothng visible, so I checked input voltages with it hooked back to teh DP plug, and get pack voltage on the inptu wires like I should, and on the regulator transistor tab, but nothing at all on it's output. :( So at a minimum, that transistor is probably toast, but it might be worse--it could be fine but something else, like the MCU, could be fried and pulling power down to nothing at the output. With my luck, that's what's wrong.


Anyway, I stopped there, because I can't hold the meter leads steady enough and am afraid I'll end up breaking a part off the board or shorting between pack voltage and something else that cna't take it, and making it even worse.


So for now, I can sleep easy knowing I've murdered a CA V3 faster than probably anyone else in the whole beta program. :lol: :roll: :oops:


Some pics of the PCB inside the CA, in case anyone sees something fried that I missed, that might help me narrow down the problem and fix it. (assumign I even *can* fix it).

file.php


file.php


file.php


file.php



Crossposted to the CrazyBIke2 Thread.
http://www.endless-sphere.com/forums/viewtopic.php?p=676206#p676206
 
The pass transistor is not getting any voltage at any lead except the collector, so this means there is nothing to turn it on.

Presumably this means that whatever voltage divider is supposed to bias it on is fried? Gotta trace out the circuits and see where stuff goes. Maybe compare to the still-working older LS CA (though I don't want to risk slipping with a meter probe and blowing that up, too).
 
I unsoldered the top right lead of the transistor, no change. Tried to unsolder the other one but it came off--it was broken off inside, and the corner of the plastic case came with it. Dunno if it broke when I moved the iron or was already done for inside due to whatever fried.

Transistor is a DN2470, an N-channel FET, 700V max, 42ohm RDSon, 500mA Ids max. The pin that broke off is the gate pin, and since it's a FET it's a Drain (with pack voltage) rather than Collector as I called it previously :oops: and the Source pin is the other one with no voltage on it. Gate-source is +/-20V max.

R24 runs between the Drain and the Gate, and it has 0V on teh gate end, and pack voltage on the other. I did not yet find what connects to the gate end of R24 to see what is keeping it pulled down to 0V. I found another voltage divider of R14 & 16 that has pack voltage at one end of R16, about 1.4V at the connection between, and 0V at the other end of R14. Not sure what it's for yet. I guess I should start drawing up a schematic as I trace it out.

Since the gate end of R24 has 0V on it even with the leg off, just replacing the FET isn't going to fix it--have to find what's at that end of R24 to see what controls the FET.

Pic below labels what I foudn so far.



More later when I can see and hold tiny stuff again, it's too cold right now. :|
 
I tried to check which wire on the Thun might've been shorted to the frame, but I cna't tell. It could have been blue or white easily, as there are several small nicks in each of those, but when I use an eye loup, someone else's old bifocals, and a magnifying glass in series to look at the wires under brighter lights, I can see at least one nick in each of the 5 wires, so it oculd've been any wire, or several. It might even be possible for them to short to each other, but the nicks aren't in the smae places, so probably didn't.

But it might nto have been a short at all. I forgot about this, from the first post of the CA v3 beta thread:
http://www.endless-sphere.com/forums/viewtopic.php?f=4&t=37964
The 10V bus can power the THUN sensor directly provided you are using a pack that is 48V or less. For higher voltages the CA's regulator would get hot and you'd want to feed the THUN 12V power from a DC-DC.
My pack is 66.2V hot off charger, and was at the time still almost fully charged, so that's a lot more than 48V. I shoudl've remembered that, before hooking it up, but I didn't, so that is probalby what killed the FET, and maybe whatever drives it.


I started trying to draw up the schematic but I kept losing my place on the PCB vs drawing, so haven't got anywhere on that yet. Gotta start work again tomorrow thru Wed, so dunno if I'll get back to this much before next Thursday.


I'm sure I have FETs that will work; I can probably use an IRFB4110 from a dead controller (or a new one if I have to--I have a few) just to test if it is all that is wrong, but I'm pretty sure it's not. Oh, and when i wire it in or solder in a drop-in-replacement, I'll have to run the gate wire directly to that resistor, as the pad seems to have come off with the broken leg. :roll:
 
Back
Top