I feel ya! Many of us have been there. But just remember, it’s not nothing. How you feel matters and it deserves attention.
7 Likes
Lots of hugs for you, if that’s OK. I have struggled a lot with intrusive thoughts too.
I believe you are a good and caring person, and I value you and your presence here.
8 Likes
I wish I were there to make you a cup of tea or open a bottle of wine so that we could just sit and look at your yard and talk for hours.
9 Likes
Well, fuck yesterday.
Twelve kilowatts Emotorwerks battery charger. Dead in water. All works but there’s no charging current at all (okay, 100 milliamps indicated, but that’s within error of the sensor). Expensive as a little hell but with full documentation.
Got it on my table couple days ago after the tech crew failed figuring out what’s wrong.
IGBTs were okay. Control board was okay, though the display needed to be powered up at startup otherwise it did not initialize (that cost me a day and looking up the documentation as I thought I damaged something). PWM signal went to the driver board. But it did not go out of the driver board to the IGBT gate.
The power from the isolated DC-DC stage was not there. The gate driver was without power. No wonder there was no output. So, DC-DC is busted?
Not so fast. The module was desoldered, tested, and it worked. There was a short in the circuit, luckily a mercifully simple one.
Of course vendor provided schematics for board version 14, and I had board version 15. Luckily again, the difference was only in the pair of FETs boosting the driver chip output; the difference was traced down from the board without issues (the vendor did not publish the v15 schema yet, or we couldn’t find it, and either didn’t reply to a request yet or my crew did not relay the mail; oops).
So the FETs are busted. Desolder them? Not so fast; let’s just desolder and lift one of the pads. That breaks the circuit without me having to take down the whole part. Of course both FETs test okay.
So lift the power pad of the driver chip. And the short is found - ding dong the chip is dead.
So buy a spare and get it shipped. Two days gone to logistics. Three, as I slept when a colleague called me and he did not retry so he got it to me a day late.
Replace chip; of course screw up a little, and misuse the hot air rig and lift pads off the board, then have to patch-wire around a little. Test. Signal now going to the gate output. So far so good.
Build a jig for testing the high voltage section. There is a bank of capacitors there, with 8 to 16 times energy sufficient to kill me. On a pretty sufficient voltage, almost three quarter kilovolt. So a safe way to discharge this bank is needed. So let’s splice a power cable and use a toy built a couple years earlier, a rig with three lightbulbs that can be connected series or parallel or bypassed and selected at will. Bloody useful thing; can be used as a series current limiter exploiting the nonlinear current-resistivity relation of a heated filament; the hotter it is, the more resistive it is. Or it can be configured as a resistive test load. In this case, all three bulbs are in series and used to discharge the caps. A multimeter is wired to the cap bank, and a neon bulb to indicate presence of voltage. A bit of an overkill but you WANT to see the caps are charged.
Antiparallel white LEDs are added to the IGBT gate driver outputs. Those little buggers can shine bright with unbelievably low current. A little test trick that allows seeing the PWM signal going to both the PFC stage and to the output. Prevents having to putz with scope probes in tight and badly accessible space too close to the lethal contacts.
The thing is hooked to the mains, and PEW and trips breakers. No wonder, with so big cap bank; the device is built for a three-phase multidozen amp feed and I am trying to feed it from a common domestic power socket. So, the second, simpler but more powerful lightbulb limiter, goes in series. Two half- and one one-kilowatt halogen lamp in selectable parallel connection. In series with the power supply input. The lamps light up brightly, strobing for a while until the cap bank charges up. So far so good.
Resistive load, output test. Power goes out. The IGBT is driven. But there’s no voltage regulation. Display insists on the output being minus 80 volts. Which is total bullshit. Value not changing despite hooking a power supply to the idled output and varying the voltage. So back to the schematics. Find the voltage sensing circuitry, trace it down. Resistor divider works, output proportional to output voltage. Controller board input voltage line dead. In between, there’s a optoisolated op-amp, relaying the voltage between the power and logic subsystems.
Schematics. Input is there. Output constant, not corresponding to input. Datasheet. Output should be corresponding to input. Disconnect the wiring harnesses, test with simulated signal. Same bloody result.
Ding dong, the chip is dead.
Of course the thing is a fairly exotic part - again - that has to be ordered. No way to get it on Saturday evening, and too exotic and expensive to stock it at home just in case.
So two more days plus the weekend. Gone to logistics again.
To add insult to injury, it is pretty possible I killed the chip myself while debugging the system earlier. The wiring harnesses are rather short and it is pretty easy to brush something against something else or let the probe tip slip.
And the thing was supposed to be reinstalled today, as the car should’ve been shipped to an exhibition tomorrow.
That all while being totally out of whack, after an ugly crying-grade depressive episode just a couple hours before because of a stupid mistake in taking my happy pills. (Intense but mercifully short.) Compounded with a long-time friend badly lying to me and scamming me out of money (and the quality of his stories and cop-outs going down so badly I could finally have enough hard data to get beyond reasonable doubts and confirm earlier suspicions). And nagging memories of love long gone; eight years and it is not getting better, which is stupid - those who say time heals these things are lying, or very lucky themselves.
And with tranquilizers or alcohol being a bad idea because of the lethal cap bank work that could not be postponed - the deadline appeared tight and unexpected. With over half-kilovolt close to one’s fingertips getting one’s cognitive faculties impaired is something to be advised against.
But the charger was supposed to make me feel better by providing me with a significant success. And the broken chip is throwing a spanner into that plan.
At least the chip can be replaced.
I should reengineer the whole thing with generic parts, using arduino-class processors for the subsystems and let them communicate digitally through cheap optocouplers, the kind you find a half dozen of when you sweep the shop floor. Built of stuff that is cheap and plentiful and likely to be already on hand. Because logistics sucks.
It could be worse.
But, still, fuck yesterday.
15 Likes
That aspect of technical problems is maddening. You never know if you are getting close to victory or if progress just leads to further endless rabbit holes. But it is what it is, and sometimes one just has to deal with it.
The emotional side sounds painfully rough, glad you could share that here too.
P.S. Dealing as a consumer with the “JuiceBox” version of the eMotorWorx products right now today, though I think our problems are probably all with the newly installed outlet circuits and the physical outlets themselves. (But that’s another story…)
7 Likes
It’s daylight saving time which means a week of exceptionally shitty sleep as my already whacked out circadian rhythms acclimate to waking up an hour earlier.
I already woke up at 4am for no reason and it took forever to go back to sleep.
In all I prefer the extra daylight you get in daylight saving time especially here in the deep dark northwest. But fuck the time change aspect, and fuck today.
So tired… 
9 Likes
I didn’t even realize till I got up this morning it was DST. Which can DIAF. It just makes cross timezone crap even more annoying.
6 Likes
Don’t know if you can say or not, but what kind of car was it for? Some kind of supercar, or a fancy electric?
3 Likes
http://www.lukaev.com/ - my uberboss’s pet project.
8 Likes
Yeah, I’ll fucking take one.
http://www.lukaev.com/gallery/_DSC0541.jpg
I’ll need a new pipe and smoking jacket. Plus I want a right hand drive for the US. I hope the zero to sixty time is measured in tweed?
9 Likes
I hardly noticed it, I thought it was still winter time and I was just tired.
Dammit, the clocks don’t change until 27 March here.
6 Likes
DST!
Yay!
I love DST! Let’s just stay like this.
3 Likes
Only $22,313 for a really good looking electric car with 155.343 miles range (or and 186.411 miles when driving efficiently) and a top speed of 91.96mph?!?! Holyshit, color me impressed. I want one.
7 Likes
Looks like steads electric car.
3 Likes
Holy what-the-wow!
I just (as in last night) leased a new Nissan Leaf at about that price.
The Leaf has the undeniable advantage of being right here right now, but now I totally covet one of those Lukas. It’s even closer to what I’ve been daydreaming about than the Morgan EV, and that’s got to be the lowest price on the market or damn near.
6 Likes
Aaaaaand I’m back on the antidepressants.
18 Likes
A hearty “fuck you too” to the universe.
My mother broke her hip last week, my uncle had a stroke yesterday, I’m on a different continent from them, and something in my hip just went “pop” in a way I’ve never heard before…
15 Likes
Can we be antidepressant pals!? 
They tend to suck. Effexor was terrible, celexa was terrible, but I’ve been doing well with run of the mill prozac.
7 Likes
Wellbutrin allowed the aliens to beam thoughts into my head. Lamotragine gave me mouth ulcers and apparently put me at risk for a syndrome where your entire body becomes a blister and then pops leaving you with third degree burns!
17 Likes
Now that is effing hardcore shit.
7 Likes
