Weber and Coulomb's MX-5

Post up a thread for your EV. Progress pics, description and assorted alliteration
7circle
Senior Member
Posts: 540
Joined: Thu, 04 Feb 2010, 02:18
Real Name: Ken
Location: Melbourne

Weber and Coulomb's MX-5

Post by 7circle » Thu, 20 Feb 2014, 04:48

Thanks for the link to user guide.

the software user guide is helpful to.

Just on the following
weber wrote: The "limiter" signal is an integer that tells you which control loop is limiting the motor output at any given time. You can read more about it, and all the other logged data, in the WaveSculptor200 manual on pages 33 thru 37 here:
http://tritium.com.au/wp-content/upload ... Manual.pdf

What you see when she misbehaves, is that the noise on the motor current setpoint (essentially the torque setpoint) causes the WaveSculptor to flip back and forth between being limited by the motor current setpoint (limiter = 1 = "current") and being limited by the difference between the back-EMF and the maximum PWM sinewave available given the battery voltage (limiter = 0 = "PWM"). When the limiter is "PWM" the motor is said to be in "overspeed" or "field weakening".


From how I read the Limit table by integer value
0 is "current" control loop not limited
1 is max motor voltage PWM limiting
2 is max motor current limiting
3 is both 1 and 2 limiting
....
and so on as other bit fields get set.

So in your plot above it is 0 or 1 so not limited by current.

Anyways I hope the throttle wiring fixed it as you said you saw the jitter/noise in the "spare" reserved field you sneaked the throttle measurement into.
Very odd symptoms.

Good Luck
7C
oh 7 circles is more about the packing ratio of wire.
7/9 = 0.7777.... of cross-section


User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Thu, 20 Feb 2014, 05:19

7circle wrote:From how I read the Limit table by integer value
0 is "current" control loop not limited
1 is max motor voltage PWM limiting
2 is max motor current limiting
3 is both 1 and 2 limiting
....
and so on as other bit fields get set.

So in your plot above it is 0 or 1 so not limited by current.
No. Sorry. I should have explained better. There's always exactly one bit set. What we've plotted is the base-2 log of the value, in other words the bit number of the bit that's set. So it's:
6 IPM Temperature or Motor Temperature
5 Bus Voltage Lower Limit
4 Bus Voltage Upper Limit
3 Bus Current
2 Velocity
1 Motor Current
0 Output Voltage PWM

So it's flipping between being limited by the motor current (essentially torque) set point and being limited by not being able to make a bigger sine wave voltage because the PWM duty cycle is saturating at 0% and 100% on the peaks.
Anyways I hope the throttle wiring fixed it ...
It seems to have done. No problem now for 6 days. Drove her to the Brisbane AEVA meeting and back tonight. Gave a presentation, with Coulomb, about the hunting down of this problem.
Last edited by weber on Wed, 19 Feb 2014, 18:20, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Thu, 06 Mar 2014, 02:04

MX-5 progress report:

Accelerator pedal signal
------------------------
We have not had any further problem with intermittent noise in the accelerator pedal signal. Hoorah! And I even took someone-I-wanted-to-impress for a drive. So here comes the final attempt to provoke the fault to return: I hereby pronounce it fixed. :-)

Clutch
------
With that problem fixed, we turned our attention to other cases of torque oscillation. Next in severity was a shudder at about 5 Hz that ocurred briefly whenever we stomped on the accelerator when stopped or travelling at low speed. "So don't stomp on the accelerator", you might say, but where's the fun in that. I was unsure whether it was the WaveSculptor somehow "losing sync" or whether it might be our new you-beaut ceramic button clutch going slip-stick-slip-stick.

It did not show up in the ordinary WSConfig logs, but then they have a sample rate of 5 Hz (200 ms between samples). Tritium_James suggested I use the more general-purpose CAN logger so I could see the RPM updated at the new rate of 25 Hz (40 ms between samples).

As it happened, when I did the test, it became blatantly obvious that it was the clutch slipping. I could see it on the tacho, because it chose that moment to slip continuously for several seconds! Once again, the WaveSculptor was blameless.

Here's the graph of RPM versus time showing the clutch slip. The solid line shows what the motor was doing, and I've drawn in a dashed line to show what the gearbox, and hence the car, was doing.

Image

The more logs I examine, and the more graphs I plot from them, the more I see that the WaveSculptor is always in complete control of the motor, and makes it do exactly what we ask it to. A superb piece of engineering.

So, for now, we have simply told the WaveSculptor to go easy on this clutch by winding the maximum motor current and hence the maximum torque back to two thirds of its ultimate maximum. 200 A rms down from 300 A rms. This is at a battery voltage of around 330 V under load, so still quite drivable.

We'll wait and see how it goes when we get the 900 V WaveSculptor, as it will probably be limited to around 200 A anyway. With that, we'll get our power from the doubling of the RPM out to which we get full torque.

The fallback, suggested by Graeme from Suzi-Auto, is to take out the new pressure-plate and double up the diaphragm spring, stacking the spring from the original pressure plate on top of the one in the new pressure plate. I'm afraid I have to doubt the claims of the clutch manufacturer that this clutch can take 380 Nm. Tritium measured our motor at 350 Nm at 300 A.

Direct Drive?
-------------
We can now answer a question that dates from the earliest pages of this build thread. Should we have left out the gearbox and done "direct drive" to the diff? The answer is:

No way!

As it is now, we drive around most of the time in second gear. When we go from 360 V to 720 V with the WaveSculptor upgrade, we may want to spend more time in first gear. Diff is 3.636, First is 3.136, Second is 1.888, Third 1.330. To go direct drive, we'd need a 6:1 unobtainium diff (120 km/h at 6800 rpm) which would be like driving in a gear between second and third now, and we wouldn't have the option of first gear for a faster launch.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Thu, 06 Mar 2014, 02:25

Stuff that just works
---------------------
We seem to be mainly talking about the problems we're still working on. I thought it might be nice to take stock of the stuff that just works.

BMS
---
Our BMS is working flawlessly. We can just trust it to tell us if anything is wrong with any one of the 218 cells buried throughout the car, or its ability to communicate with them. And if we're not looking at the stress gauge when something goes wrong we can trust it to get our attention by limiting the power to save that cell.

I did another deliberate run-flat a few weeks back. Drove to Jolly's lookout near Mt Nebo and back, and several times around Mt Coot-tha. This time the stress gauge gave gradual warning of impeding doom, but I still madly insisted on having my wife drive it halfway up the steep side of Mt Coot-tha, whereupon it said "Oh no you don't" and slowed us to a crawl. We turned around and regenned back down again and drove, at a normal pace, the two kilometres back home. It just started to limit again as we got home. Every cell recorded the worst conditions that it experienced, and these were all within spec. The range this time: 119 km.

Batteries etc
-------------
Implied in all the above is the fact that lots of other electrical things are working flawlessly:
The ABB motor,
the Tritium WaveSculptor,
the 2 modified Tritium EV Driver Controls units,
218 cells and all their interconnects,
20 EV200 contactors,
11 Anderson connectors,
over 100 crimp lugs,
many 12 volt crimp connectors and relays and
32 optic fibre connectors.

Image
Last edited by weber on Wed, 05 Mar 2014, 15:31, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

7circle
Senior Member
Posts: 540
Joined: Thu, 04 Feb 2010, 02:18
Real Name: Ken
Location: Melbourne

Weber and Coulomb's MX-5

Post by 7circle » Thu, 06 Mar 2014, 06:04

   Image Image
Image Image Image
   Image Image

Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Weber and Coulomb's MX-5

Post by Renard » Thu, 06 Mar 2014, 15:28

weber wrote: Stuff that just works
---------------------
We seem to be mainly talking about the problems we're still working on. I thought it might be nice to take stock of the stuff that just works.

Implied in all the above is the fact that lots of other electrical things are working flawlessly:
The ABB motor,
the Tritium WaveSculptor,
the 2 modified Tritium EV Driver Controls units,
218 cells and all their interconnects,
20 EV200 contactors,
11 Anderson connectors,
over 100 crimp lugs,
many 12 volt crimp connectors and relays and
32 optic fibre connectors.


So after all your travails it seems that your build work is actually well executed!
Renard

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Fri, 07 Mar 2014, 00:32

Thanks guys.

We did have a failure in one electrical subsystem recently.

One rainy Friday night two weeks ago, Coulomb was driving home when he lost all traction power and had to coast to the side of the road. He turned everything off and waited a few minutes and retried.

Once you turn the key off, the factor determining your ability to go again is whether there is enough energy in the 12 V battery to pull in all 11 of the EV200 battery breakup contactors and 2 DC-DC contactors, just to get power back to the DC-DC converters. Provided that happens, you can sit there for a while with all 12 volt devices off, while the DC-DCs recharge the 12 V battery.

If not, you need a jump-start just like anyone else. In theory, if he'd had the spanner and jumper leads, he could have taken the lid off the battery box behind the rollbar, and picked any group of 4 LiFePO4 cells to jumper to the 12 V battery. Wouldn't that be a lovely thing to do in the rain?

Once the EV200 contactors are on, they will stay on until the voltage falls below 7.5 volts, and they only draw 130 mA each. But to pull them in needs 3.8 A each for 130 ms at 9 V or more.

Fortunately there was just enough charge left in the 12 V battery so that, after a rest to redistribute its charge, it managed to pull in all 13 contactors, but not without a lot of chattering first.

It turned out that one of the two DC-DC converters had failed (possibly some weeks earlier) and the surviving one could not keep up with the demand of the power steering, headlights, interior fan on full blast, and windscreen wipers. When the 12 V battery fell below 9 volts the WaveSculptor tripped off, as it is supposed to, because it could no longer maintain reliable gate drive to its IGBTs (it uses an internal 15 V boost converter off the 12 volt CAN-bus power).

We briefly considered a warranty claim for the DC-DC. It has 5 year warranty. It is a MeanWell HLG-240H-15A which is designed and sold for powering LED lighting from the AC mains. We imagined filling out the claim form:

Nature of supply: 360 V lithium battery with 50 kW of regenerative braking from an induction motor.

Nature of load: Incandescent lamps, numerous motors, 12 volt lead-acid battery. Oh alright, we admit it. It's a CAR!

We ordered a new one.

But that left us free to do a post mortem on the old one, potted in silicone or polyurethane though it was.

Image

Image

We didn't find out how or why it had failed, but we traced enough of the circuit to confirm that it has inrush protection and lots of filtering and spike suppression on the input side, and that the output side has a synchronous rectifier circuit that should not have any problem with a voltage source (the 12 V battery) being present on its output even when the input is not powered. We also found out why our 12 V battery was tending to go flat if the car wasn't used for a week. There were four 1k5 resistors across the output. So we will need to add relays to isolate the DC-DC outputs anyway, whenever their inputs are not powered.

We found that with only one DC-DC we could have headlights or power steering but not both, and had to go easy on the fan. So Coulomb now has arm muscles, since he had a strange attachment to seeing and being seen while driving home in the dark, and he was not about to give up the fan, given the weather we'd been having in Brisbane. That leads to my next story: the air conditioner.

P.S. The new DC-DC has arrived and will be installed very soon.
Last edited by weber on Fri, 07 Mar 2014, 13:05, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
offgridQLD
Senior Member
Posts: 1825
Joined: Tue, 23 Jul 2013, 16:05
Real Name: Kurt
Location: Fleurieu Peninsula, SA

Weber and Coulomb's MX-5

Post by offgridQLD » Fri, 07 Mar 2014, 00:44

"That leads to my next story: the air conditioner."

Reverse cycle electric airconditioner compressor? With bonus of more efficient heating to Image

Kurt

Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Weber and Coulomb's MX-5

Post by Renard » Fri, 07 Mar 2014, 02:57

weber wrote:
It turned out that one of the two DC-DC converters had failed

But that left us free to do a post mortem on the old one, potted in silicone though it was.

We also found out why our 12 V battery was tending to go flat if the car wasn't used for a week. There were four 1k5 resistors across the output. So we will need to add relays to isolate the DC-DC outputs anyway, whenever their inputs are not powered.


Since many DC/DC converters have a little fan in them, maybe the potting resulted in overheating? With a 417V rating it seems unlikely that input voltages were a cause.

And the quiescent current draw had tripped me up with my converter too, requiring the addition of a relay.
viewtopic.php?title=renards-bmw&p=43673&t=2921#p43673
Last edited by Renard on Fri, 07 Mar 2014, 05:40, edited 1 time in total.

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Fri, 07 Mar 2014, 03:04

offgridQLD wrote:Reverse cycle electric airconditioner compressor? With bonus of more efficient heating too Image
Sadly, our little team of SI units does not include a Lord Kelvin (refrigeration engineer), so we were not brave enough to attempt this. It's strange that the commercial EV manufacturers have not done so yet either.

[Edit: Newton would like it noted that he was in favour of giving reverse-cycle a go, using this device from EV Works http://www.evworks.com.au/index.php?pro ... AC-RCVALVE, and paying a refrigeration mechanic to do the job. Although I don't recall Newton offering to pay the mechanic. Image ]

You can see how we adapted the existing MX-5 aircon compressor to the electric motor here (before it got buried under a slab of batteries and motor controller):

Image

For more information about this see:
viewtopic.php?t=980&p=34425#p34425
viewtopic.php?t=980&p=28063#p28063

Unfortunately it didn't work, because all the refrigerant gas (134a) had leaked out of the system. This was hardly surprising given the time we took to do the conversion, but we had also bent those flimsy aluminium pipes a lot while installing other things, so we might well have caused a leak. Anyway, aircon wasn't a priority until the car was actually usable as a daily driver.

Finally, last Friday, we got it checked out. Jamie of "Keepin' Cool" mobile car aircon service came out to Coulomb's place (he only works on the southside). We told him in advance that the job was a little unusual, and sent him the above photo. He was intrigued. Another company to whom I had earlier sent the same request, didn't even reply.

We wanted to go with a mobile service because we fully expected to have to disconnect some hazardous voltage components to give access to aircon parts we had covered up. Fortunately Coulomb figured out the B2 contactor box would have to move, and got it out of the way before Jamie arrived.

We were in luck! The only leak Jamie could find was the usual one at the compressor shaft seal. No worse than expected for a system of this age (1990). He estimated we'd need a regas in 2 years. So he evacuated the system and began putting in the gas.

There was an awkward moment when he told us the belt was slipping and we needed to tighten it before he could proceed. The adjustment was completely inaccessible! Fortunately he had a few phone calls to make while Coulomb and I frantically unbolted the mountings for the WaveSculptor and other contactor boxes and propped it up with blocks of wood. Even after we did that it was very awkward and I still have a bruise where I forced my arm to reach where it really didn't want to go.

Then Jamie pointed out that we had no fan for the aircon radiator (condenser) and that when we were stationary or going slow, this would lead to the condenser running so hot that the system would shut off on overpressure, and this would increase the rate of loss of gas from the system. He looked up some books and told us we would need a minimum air flow rate of "400 CFM" (190 L/s) and suggested various ways we might achieve that in the limited space available. We temporarily used the 240 V fan you can see in the following photo, to allow Jamie (right) to complete his tests.

Image

But since then, Newton (Jeff Owen) pointed out that when we're stationary, the motor, and hence the aircon compressor, wont be turning either, so the lack of fan won't be an issue. At the time I said, "But we'll need to put in a mode where the motor can idle in neutral, or when the clutch is in, to run the aircon while we're stopped at traffic lights." But since trying it without this, I find that the aircon system has enough thermal storage to tide you over without the motor needing to run at traffic lights.

So perhaps what we need is a mode where you cannot rev the motor in neutral or with the clutch in, when the aircon is on. And we rely purely on rammed air.
Last edited by weber on Sat, 08 Mar 2014, 12:01, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Fri, 07 Mar 2014, 03:35

Renard wrote:Since many DC/DC converters have a little fan in them, maybe the potting resulted in overheating? With a 417V rating it seems unlikely that input voltages were a cause.
Brainstorming possible causes is very welcome. The potting was done by the manufacturer as standard. They have an IP65 rating. They have a max ambient temp of 50°C and max case temp of 95°C. The two are stacked one above the other. We put them upside down (label-wise), which means that the side with all the power transistors and rectifiers was downward. It was the lower of the two devices that failed.

Their max DC voltage rating is actually 431 V which is 277 Vac +10% * sqrt(2). However all the X and Y capacitors only had a 250 Vac rating. We have the WaveSculptor set to limit the DC bus voltage on regen to 392 V. The chargers have the same limit which is 3.60 V per cell. They balance at 3.57 V.

Data sheet is here: http://www.meanwell.com/search/HLG-240H ... H-spec.pdf
Last edited by weber on Fri, 07 Mar 2014, 13:00, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Sun, 09 Mar 2014, 00:41

Wheel Alignment
---------------
Coulomb and I spent an enjoyable day at Dogphlap's place on Friday. Dogphlap offered some months ago, to assist us in using the wheel alignment kit he designed and built using two low cost lasers and an inclinometer.

On the MX-5, both the front and rear wheels have toe-in and camber adjustments, and the front has a caster adjustment as well.

Here are some diagrams and explanations of what those 3 terms mean:
http://en.wikipedia.org/wiki/Toe_%28automotive%29
http://en.wikipedia.org/wiki/Camber_angle
http://en.wikipedia.org/wiki/Caster_angle

We had seriously scrambled these during the conversion process -- the front by installing a power steering rack and new coil-over-shock-absorbers, and the rear by using the four lower wishbone bolts as battery box mounts.

At both front and rear, the pairs of lower wishbone bolts are "cam bolts" and provide the adjustment by means of an eccentric washer (the cam). At the rear, adjusting them differentially provides the toe-in adjustment and adjusting them together provides the camber. At the front, the geometry is quite different and the front bolt adjusts camber while the rear bolt adjusts caster and toe-in is adjusted at the steering ball-joints.

I won't try to explain how Dogphlap's measuring system works in any detail. In words, it sounds a lot more complicated than it really is. But here are some photos.

Image

Image

Image

Image

What we found was horrible maladjustment. Both rear wheels were pointing to the right about as far as they could go ("crabbing") and the front had toe-out of about twice the magnitude of the required toe-in! But with Dogphlap's help we put it all right. On the drive home, Coulomb pronounced it noticeably more stable in steering than when he had driven it there. This may well increase our range somewhat too.
Last edited by weber on Sat, 08 Mar 2014, 13:46, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Sun, 09 Mar 2014, 09:23

Friday night, Coulomb and I replaced the failed DC-DC converter, so we can have headlights and power steering and fan again. Image

I hadn't driven Mexy since the wheel alignment, so last night (Saturday) I took her for the longest continuous drive she's had since going electric. I went from Bardon (Brisbane) west to Haigslea on the Warrego Highway and back, then finished off with two Western Freeway loops between Mt Coot-tha road and Moggill road, and two laps of Mt Coot-tha (one in each direction). I started a third lap up the steep side when the BMS finally started limiting the power and I turned around, regenned back down and drove home.

I drove for almost two hours and did 131 km. I confirmed this odometer distance to the kilometre with Google maps. This is with the WaveSculptor limiting the motor current to 200 A to avoid clutch slip. So maximum power around 60 kW. The motor temperature never went over 120°C. It sat on 90°C to 100°C most of the time.

On the way back I ran the aircon with the fan on full until my ears froze, then I ran the heater until they thawed out again and it got uncomfortably warm, just to prove I could. It's interesting that there's still a faint whif of vinegar from the heater -- acetic acid from the curing of the high temperature silicone used in constructing it.

"But what about the effect of the wheel alignment?", I hear you ask. Coulomb's 26 km drive from Boondall to Bardon yesterday had the benefit of the wheel alignment, but it didn't have the benefit of the power steering, so I can understand his understated approval.

I was cautious at first, Image because frankly it had previously felt dangerously unstable at speeds over 90 km/h, and Coulomb had had a surprisingly bad skid on a roundabout in the wet a couple of weeks ago Image (luckily there were no other cars around at the time).

But once out of the city and past the roadworks, and with a smooth road all to myself, it felt as stable at 120 km/h as it did at 60 km/h. On the way back I went via Mt Crosby. Once I started throwing it around the winding roads, my EV grin Image turned into joyous EV laughter Image. Thanks, Dogphlap! Image This is the MX-5 I remember from before electric conversion. Same again on the curves on Mt Coot-tha. This is how an MX-5 is meant to feel -- like it's on rails. You just wouldn't know it weighed 33% more than it used to.

The only time you feel the extra weight is when the road gets bumpy, and I believe this is more about the distribution of the weight than the absolute weight, and the flexibility of the convertible chassis. It feels kinda like there are two dumbbells (front and rear) with a spring in between them. If you hit a bump with one wheel it sets up a torsional shudder between front and rear, undamped by any shock absorber.

Newton has been relentlessly researching this issue, including obtaining advice from automotive engineer and MX-5 expert Richard Larsen on how best to torsionally stiffen the MX-5 chassis. So we have some things to try in the near future. Unfortunately they probably require removing a significant number of cells from the vehicle -- those which are furthest from the roll axis, both front and rear.

[Edit: Spelling, grammar, punctuation]
Last edited by weber on Sun, 09 Mar 2014, 05:39, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
evric
Site Admin
Posts: 500
Joined: Sun, 20 Jul 2008, 01:57
Real Name: Eric
Location: Adelaide SA
Contact:

Weber and Coulomb's MX-5

Post by evric » Sun, 09 Mar 2014, 14:17

With the DC-DC converter, have you got a diode on the input to ensure that when the pack voltage droops under load the diode prevents the system trying to pull power from the temporarily higher voltage of the DC-DC converter? This was discovered by Jack Rickard (EVTV) after many failed DC-DC converters. He also places a choke (inductor of 100uH) in series with the input as well and another diode across that to catch any back emf from the choke.
My original DC-DC converter was a Meanwell SP-500-13.5, and this failed after 2 years. I didn't have an input diode on this one. When I spoke to a DC-DC Converter fixer upperer he said that 2 years was good for a Meanwell unit. He was surprised it lasted that long - and no, he wouldn't even consider trying to repair it. (Mine wasn't potted)

Eric

See circuit diagram and parts list here
Last edited by evric on Sun, 09 Mar 2014, 03:22, edited 1 time in total.
Prius Plug-in Conversion: http://www.evplus.com.au ...Holden Barina EV: http://www.evric.kestar.com.au

User avatar
PlanB
Senior Member
Posts: 465
Joined: Sat, 16 Jan 2010, 15:24
Real Name: Kris McLean
Location: Freemans reach
Contact:

Weber and Coulomb's MX-5

Post by PlanB » Sun, 09 Mar 2014, 14:21

About the EV grin! I'd been for a few EV rides before I got the Leaf Dave, in a converted MX5 & a Delorean, so I had half an idea what to expect. What I didn't anticipate could be so much fun are the steep hills. Going up, don't you love it how an EV just goes faster as you plant your foot, no change down, no ICE screaming its head off after it's kicked down a cog. And going the other way where you end up with more range at the bottom than you had at the top.

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Sun, 09 Mar 2014, 20:27

evric wrote: With the DC-DC converter, have you got a diode on the input to ensure that when the pack voltage droops under load the diode prevents the system trying to pull power from the temporarily higher voltage of the DC-DC converter? This was discovered by Jack Rickard (EVTV) after many failed DC-DC converters. He also places a choke (inductor of 100uH) in series with the input as well and another diode across that to catch any back emf from the choke.

Thanks evric. I can see how some extra inductance could be useful, and it needs the reverse diode across it. But I completely fail to see what use the series diode could be, given that these devices are designed to work off AC as well as DC and so have a bridge rectifier inside them. This bridge comes right after the fuse, the NTC inrush limiter, the EMI filter and the MOV. The EMI filter looks a lot like this:

Image

Except it includes TVS diodes across all the inductors. And spark gaps across the differential inductors, which are doing the same job as the proposed additional inductor.

I can't see how even putting a sudden dead short across the output could do any damage to the converter with the energy stored in the EMI capacitors and inductors.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
evric
Site Admin
Posts: 500
Joined: Sun, 20 Jul 2008, 01:57
Real Name: Eric
Location: Adelaide SA
Contact:

Weber and Coulomb's MX-5

Post by evric » Mon, 10 Mar 2014, 05:19

weber wrote: ... I can't see how even putting a sudden dead short across the output could do any damage to the converter with the energy stored in the EMI capacitors and inductors.


I can see what you mean if you are using a power supply for a DC-DC converter, like the Meanwell, but the ones that Jack was using and the one I am now using (TDC144V/12V made by TC Charger) are made as DC-DC Converters with a DC only input marked (+) and (-). These would have a completely different front end.

Maybe the diode is redundant in your case.
Have you put back the same type of DC-DC converter as the one that failed?
Last edited by evric on Sun, 09 Mar 2014, 18:24, edited 1 time in total.
Prius Plug-in Conversion: http://www.evplus.com.au ...Holden Barina EV: http://www.evric.kestar.com.au

User avatar
coulomb
Site Admin
Posts: 3721
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Tue, 11 Mar 2014, 13:46

evric wrote: Have you put back the same type of DC-DC converter as the one that failed?

Yes. Are you suggesting that this is a bad idea? We really don't know whether it will last or not.
Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
1.4 kW solar with 1.2 kW Latronics inverter and FIT.
160 W solar, 2.5 kWh 24 V battery for lights.
Patching PIP-4048/5048 inverter-chargers.

User avatar
evric
Site Admin
Posts: 500
Joined: Sun, 20 Jul 2008, 01:57
Real Name: Eric
Location: Adelaide SA
Contact:

Weber and Coulomb's MX-5

Post by evric » Tue, 11 Mar 2014, 13:56

coulomb wrote:
evric wrote: Have you put back the same type of DC-DC converter as the one that failed?

Yes. Are you suggesting that this is a bad idea? We really don't know whether it will last or not.


I have no idea why my Meanwell failed, especially now after you have explained the input cctry. Maybe they are just unreliable, as the expert seamed to imply.
Just see how long the new one lasts...
Prius Plug-in Conversion: http://www.evplus.com.au ...Holden Barina EV: http://www.evric.kestar.com.au

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Wed, 12 Mar 2014, 23:39

There are currently 4 main barriers to full enjoyment of our EV MX-5 driving experience:

1. Full power not yet available. We know what to do about that but we just have to wait for the testing of the new controller, and probably also stiffen the clutch pressure-plate spring.
2. Bottoming out the rear suspension too easily on potholes. We need to either get stiffer rear springs or remove cells from the rear. We're still undecided on which, as it depends partly on the solution to 3 below, and will impact 1 above.
3. A torsional body oscillation (shudder) on one-sided bumps, even when it doesn't bottom out. We hope this can be cured by some under-body bracing, but it may also require removing cells from the places farthest from the roll axis, which would also improve 2 above, and degrade 1.
4. A drivetrain oscillation at about 5 Hz on launch. I finally got a handle on this one today!

I bought an Ai-Ball Wi-Fi camera from Jaycar.
http://search.jaycar.com.au/search?w=ai-ball&view=list

I attached it under the MX-5, with cable-ties and gaffer tape, to the middle of the cross member between the front wheels, looking backwards. It was also upside-down, because it only has one flat side.

Here are some photos showing the Ai-Ball in place. There are two different exposures because one exposure shows better the Ai-Ball camera itself, and the other shows better the netbook screen displaying the live video from the Ai-Ball (in which you can see the still-camera that took these shots).

Image

Image

The live display worked first go, but the software used for recording the video to disk is written in Java and it took me several frustrating hours to convince Windows 7 and the Java VM to let me run it. I had to update both the firmware in the Ai-Ball and Java on the netbook, then add the Ai-ball's URL to an exceptions list in the Java control panel. And I still have to tick a box indemnifying Sun Microsystems every time I run it.

Anyway, I finally got to take it for a drive. Below you can see the worst example it recorded, of oscillations on launch, in first gear. The video runs at a quarter of actual speed. So those oscillations you see happening a little over once a second are really happening about 5 times a second.



From this, Newton and I conclude that the left engine mount is probably torn, due to a combination of age and about twice the torque of the original motor. New mounts are on the way.

Unfortunately, that is probably the last video this Ai-Ball will ever record. It got wiped out by the second speed-bump I came to. I could tell when it went because it also has a microphone. The sound stopped rather abruptly. I picked up the pieces off the road.

Image
Last edited by weber on Thu, 13 Mar 2014, 04:03, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
Richo
Senior Member
Posts: 3685
Joined: Mon, 16 Jun 2008, 00:19
Real Name: Richard
Location: Perth, WA

Weber and Coulomb's MX-5

Post by Richo » Wed, 12 Mar 2014, 23:49

"Worlds smallest camera" from e-bay are ~$12.
Record to uSD card but has a better resolution.
I took some nice video blutac'd to a Ladybird
So the short answer is NO but the long answer is YES.
Help prevent road rage - get outta my way!

User avatar
BigMouse
Senior Member
Posts: 600
Joined: Thu, 28 Oct 2010, 02:39
Real Name: Vincent Tannahill
Location: Silicon Valley
Contact:

Weber and Coulomb's MX-5

Post by BigMouse » Thu, 13 Mar 2014, 00:38

Wow, that oscillation is pretty severe. No wonder you were having issues tuning the response of your controller. The springyness caused by that is huge. I'm going (semi) solid mounts on the 300zx. Polyurethane spacers pretty much, with a bolt going all the way through. I'm hoping it doesn't make too much noise.

Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Weber and Coulomb's MX-5

Post by Renard » Thu, 13 Mar 2014, 02:19

weber wrote:

From this, Newton and I conclude that the left engine mount is probably torn, due to a combination of age and about twice the torque of the original motor. New mounts are on the way.


I see from your earlier post viewtopic.php?title=weber-and-coulombs- ... 840#p29840
that you re-used the old mounts. The LH one would be in tension not compression. Like you, I would hope it was only age that was the problem.

User avatar
weber
Site Admin
Posts: 2601
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Thu, 13 Mar 2014, 03:55

Thanks for that, Richo. Wish I'd known earlier.

I look forward to hearing how the PUR blocks go, BigMouse

Thanks for that link, Renard. The LH one is certainly in tension now. But because the mounts are at about +- 60 degrees they are initially mostly in shear when they are both intact. We plan to fill the gaps in the new ones with Sikaflex to stiffen them somewhat.

Coulomb, Newton and Pascal were playing hide and seek. Coulomb was "it". Pascal ran away and hid. Newton stayed right in front of Coulomb, drew a one metre by one metre square on the ground and stood in it. Coulomb finished counting, opened his eyes and said, "I see you Newton". Newton replied, "You are mistaken, Coulomb, what you see is one Pascal".
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
Johny
Senior Member
Posts: 3729
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Weber and Coulomb's MX-5

Post by Johny » Thu, 13 Mar 2014, 14:34

weber wrote:We plan to fill the gaps in the new ones with Sikaflex to stiffen them somewhat.
Good find! Shame about the camera. While I understand the feeling that you don't want this happening again (mounts breaking) I don't think that you will have problems with new mounts - old ones with new hotter motors are a no-no. My worry is that if they are stiffened too much you may get drive train rumble coming through the cabin. The gaps also serve a purpose to allow the mount to flex.

Post Reply