New Instrumentation for Speedster

We took two DC-3s with 26 people to a gathering of some 25 DC-3’s from around the country in Rock Falls, Illinois.  It was a cooling break from the weather here and a good time was had by all.  We flew in formation from Rock Falls to the opening day at Oshkosh Wisconsin for the EAA Airventure 2010 air show.

It was a pleasant trip all around.  Basler Turboconversions in Oshkosh hosted a lavish reception for DC-3 drivers on Tuesday night of that week.  We arrived safely back in Cape Girardeau to resume one of the hottest summers here that we’ve had in recent years.

The heat has been oppressive. It’s kept us mostly OUT of the Motor Verks.  I’ve been primarily working on an instrument for electric cars that we can connect directly to the battery pack and derive current and voltage information via bluetooth wireless.  If I can get this done, it will allow a variety of front ends on any computer or bluetooth smart phone to present current, voltage, amp hours, killowatts, killowatt hours, etc.

Of course I’m trying to learn Mac OS X Cocoa and Objective C at the same time, as I’ve moved to that platform mostly.  It’s a bit of a learning curve for an old guy.  But it all happens mostly in the AIR CONDITIONING.

As a result, I’ve been remiss in the shooting and editing of our “weekly” show.

[jwplayer file=”news080610-1280 –” hd.file=”” image=”” streamer=”rtmp://” provider=”rtmp” html5_file=” –” download_file=””]

All that notwithstanding, it’s not that there has been no progress.  Actually, we’ve done quite a bit.  We acquired the old Windisch Motor Company building down the street at 601 Morgan Oak Street.  We had been more or less hemmed in by our own collection of various vehicles and components to the point that we really couldn’t move cars around very well.

So we took this empty automobile dealership down the street.  Built in the 1940’s, it was originally a Kaiser/Fraser/Jeep dealership.  In recent years, it was mostly empty with a local lumberyard using it for some storage.  We’ve purchased the property and installed a 12,000 lb lift.  We’ll be doing some shooting from there when we start a larger project we have planned.

We did get our VW transaxle back and installed in Speedster part Duh.  We had originally purchased a very deluxe transaxle from Long Enterprises with 3.44 rack and pinion and gears we thought would work much better with an electric drive.  This transaxle was shipped to Special Editions Inc and they installed it in our roller.  We built the car, and on the first test drive it became apparent there was something wrong.  Any time you removed throttle in third gear the transmission would pop out of gear, often with an unappealing but at least loud noise.  Not good.  We went over the linkage and so forth for several weeks, but there really isn’t much to it with a VW transaxle.  We discussed it with Long and they readily agreed there was an internal problem.

Unfortunately, a VW transaxle based car is really pretty much built on top of a VW transaxle.  More so than any other type really.  You start with a transaxle, and bolt things onto it until with sufficient parts glommed onto it, you have a car.  To “remove” the transaxle pretty much means total disassembly.  This was a heartbreaker.

But we pulled the motor and some other rear engine compartment items and did in fact get it extracted from the vehicle.  We shipped it to Long and some three weeks later received it back.  They never did get very specific as to what had gone wrong, but the new one works quite well.

While it was there, we had the 0.70 4th gear replaced with a 0.93.  We had had a pretty steep drop from 3rd to 4th gear.  All of this was more or less intentionally designed to get our RPM’s down in the sweet spot of the motor (2000-4000 rpm) for all gears.   Recall in the original Speedster, we really only use 3rd and 4th gears and with the 3.88 rack and pinion, we were topping out at about 95 mph and maybe 4500 rpm.  The Netgain motor would do another 1000 rpm, but was really kind of torqueless at that speed with a 120v pack.  So we wanted to move things down in rpm and get 1st and 2nd gear in the game, and get our highway speeds back down in the good torque range.

We did this by going to a taller rack and pinion 3.44 vice 3.88.  The early VW’s used this but in recent years with the move to lower gearing, higher rpms, and closer gear ratios, it had fallen out of favor.  It proved just the thing for electric drive.

But we overshot the mark a little.  Going from the 1.14 third gear to the 0.70 fourth gear was a huge drop and we were seeing about 2100 rpm at 70 mph.  While good for torque, it takes quite a bit of power to overcome the wind resistance at that speed and higher, and the motor cools better at a little higher rpm.  So we took the opportunity during the transaxle repair to change out the 4th gear to a 0.93.  It feels much more natural moving from 3rd to 4th now, and we see about 2800 rpm at 70 mph.

The results are good.  I kind of enjoy driving the original Speedster in 3rd and 4th.  But if you are going to produce an electric car with a 4 speed transmission, each of the gears should actually be functional and play a role in moving the vehicle.  We’re there.

First Gear   2.64 for 9.08 final ratio tops out at 37 mph at 4500 rpm.
Second Gear 1.93 for 6.64 final ratio tops out at 50 mph at 4500 rpm
Third Gear 1.14 for 3.92 final ratio tops out at 85 mph at 4500 rpm.
Fourth Gear 0.93 for 3.20 final ratio tops out at 104 mph at 4500 rpm.

The motor itself will go much higher RPM than the original Netgain, but again, wind resistance kind of limits us.  I’ve driven it a steady 90 mph with a LOT of pedal left.  We haven’t actually found the top speed on this thing yet, but it’s a felony.

Reinstalling the transaxle really brought us another somewhat pleasant surprise, though it shouldn’t have been a surprise.  Electric cars are just easier to work on.  The first time installing everything we’re measuring and building and agonizing and replacing, and changing, and doing the usual thing.  But in reinstalling the transaxle, we found a curious thing.  It takes about 10 minutes to install the motor the second time.  It takes a few minutes to jack up the shock absorbers.    But really, everything is quite clean and goes on quite easily.   Electric cars are just heroically simpler devices than internal combustion vehicles.  If we weren’t changing/upgrading, but just replacing a blown motor, it’s about a 30 minute job taking the old one out and putting a new one in.  It just isn’t anything.

The new transaxle allowed us to do some more regen test drives.  First, I had reported the settting of the Charge Efficiency Factor on the Xantrex may have influenced our earlier regen results as I found it on AUTO for some reason after setting it at 100%.  Oddly, the Speedster was set on 100% all along so that does NOT explain anything on the earlier Speedster results.

The use of kwH as a backup measurement for the regen results was likewise a dead end.  It tracks our AH measurements quite linearly.  Our meters are working.

As it turns out, it appears “I” was actually the problem with the results.  More accurately, the driver, and by extension, the driving style, can quite swamp the efficiency gains of regenerative braking.

We did six more test drives, again in mostly urban traffic with hills, stop lights, and mucho traffic.

Without regenerative braking, I did a 21.317 mile loop to Jackson Missouri using 34.2 AH or 1.609 AH per mile to make the trip.  It took 4.839 kwH to recharge the car for 227 wH per mile.

WITH regenerative braking, I did the same trip in 35.0 AH for 1.642 AH per mile and it took 4.860 kwH to recharge the vehicle for 228 wH per mile.

Just like our initial tests, I actually had a NEGATIVE efficiency gain using regenerative braking.

We sent Brain on the same drive.

With no regenerative braking this guy is a watt-hog.  He did a 21.19 mile loop in 39.5 AH for 1.864 AH per mile and used 5.403 kwH to recharge the car for 255 wH per mile.

WITH regenerative braking, he was back down to 33.6 AH for the identical distance reading indicating 1.585 AH per mile.  On recharge, he used a more economical 4.789 kwH for 226 wH per mile.  This is a 14.93% efficiency gain from regenerative braking for the Brain.

Trace Jessup repeated the test.  With no regenerative braking, he did a loop of 21.34 miles using 36.4 AH for a 1.705 AH/mile result and recharged with 4.972kwH for 233 wH per mile.

With regenerative braking, he did a 21.19 mile loop in 34.4 AH for 1.623 AH/mile and recharged at 4.895 kwH for 231 wH per mile.  This was a more modest gain of a little under 5%.

The bottom line appears to be that you can get any result you want, depending on how you drive.  But the other conclusion inescapably is that how you drive the car rather swamps whatever technology you buy and install.

I don’t think of myself as “hypermiling” to get my results.  It feels to me like normal driving.  But I do have a couple of years behind the wheel of our original Speedster project.  It has no regenerative braking and freewheels very nicely when you ease of f the throttle.  This was a little disconcerting at first as you expect a car to slow down when you ease off the throttle.  But I quickly became used to it and kind of developed a “push and glide” driving style that is very hard to NOT do when driving part DUH with no  regen.

I would note that the regenerative braking, with about 10% of neutral braking on the throttle, very closely approximates the feel of a normal car.  The brake regen gives it a kind of power brake feel, and the neutral braking on the throttle simulates engine compression a bit.  It is undoubtedly the best setup for new drivers.  It DOES require you to give up the clutchless shifting, but if you’ve never had that, it’s not much of a loss.  You just shift like you would any sports car, and it works fine.

We did receive a replacement air conditioning compressor AND controller from Revolt in Texas – actually drop shlpped from Tecumseh.  A couple of interesting and somewhat disappointing notes.

First, the admonition that we SHOULD have added oil to the original compressor, which wasn’t mentioned in ANY of the documentation, turns out to be a bit bogus.  Tecumseh advises that the compressor is shipped with 290cc of oil in it and no additions should be made unless you pump the oil out of the compressor.  The original compressor from its first instant made a HUGE racket.  Not having ever used one before, I was unsure, but it just didn’t sound like it could possibly be happy.

It wasn’t.  The replacement puts out a very audible purr, but it is nothing compared to the original compressors complaint.

The second problem we were having is a bit pernicious.  The controller and compressor work very well to air condition the Mini Cooper – in the garage.  And I do mean it works very well.  Very nice cool temperatures out of the vents and it feels like a POWERFUL air conditioner.   You can adjust the rpm, and thus the output of the compressor, quite widely.  But it has little effect on the air conditioning system.

The problem is in driving the car.  If you drive very nicely and somewhat sedately, it works well.  But on acceleration, you can note a distinct drop in rpm from the a/c compressor.  But if you accelerate smartly, it actually drops to nothing and never comes back.  And in the process blows the 30 amp input fuse to the controller.

I don’t quite understand this.  The controller appears to draw about 5.5 amps from our pack normally in the garage.  Our pack voltage is nominally 375 volts.  If I accelerate strongly, this sags to about 345 volts.  This is quite normal, and quite expected.  It would certainly be this way with any battery pack, even lead acid cells.

But the current to the controller apparently then exceeds 30 amps and the power to the compressor reduces such that the compressor stalls.  We tried this at several settings of the compressor, from minimum to maximum with the same result.  Worse, the fusing does not appear to do its job.  After about three blown fuses, you also have a very dead controller.

When we talked to Revolt about it, they note that it works very well with 144v systems.  The controller is rated for 120-420vdc, but they really haven’t ever had anyone operate it at 375v before.  So we’re kind of the beta test crew.  First bug report:  it doesn’t work.

So all we need is a 375v 3kw 3-phase controller, that has sufficient voltage regulation to keep from killing itself, and we would have air conditioning.  Very discouraging.  I’ve ordered THIS from eBay and we’ll experiment with it a bit… but I’m not hopeful. If any of you have any ideas as to what we’re doing wrong or how to fix it…..

Finally, we’ve been struggling with the Xantrex instrumentation. We had originally powered this by tapping off the first 10 cells of our pack. They list 6ma as their power requirements. That turns out to be WITHOUT the backlight. With the backlight, it is more like 70ma and of course we had this running 24×7.

The real fix is to use a DC-DC converter to power it, but you lose the ability to read your pack voltage if you do that. So we did a little voltage divider using 10W resistors and a pot that we could adjust to calibrate our 5:1 divider so that it showed true pack voltage. This is all well and good, but the divider itself then consumes power and the drain on the pack works out to 5 or 6 AH daily. And the resistors actually get hot.

Damien Maguire of Ireland suggested we got to a VIctron Energy BMV600HS. The HS model can go up to 360volts, does its own measurement and power conversion, and is a MUCH simpler install than the Xantrex. Their shunt and circuit board is actually in a very nice weatherproof box and the connection to the actual gage is a simple telephone cable. It will really be much easier for Special Editions Inc to install.

The device is $405 and kind of hard to find. We found it at Peter Kennedy Yacht Services The HS version is NOT actually shown on his web site, but contact him and he can get you one. Reasonable delivery time. We like it.

The display is a little harder to read, and the setup is a little more confusing. But it is basically the same function and features as the Xantrex. We use the relay at 10%SOC to connect a 4k resistor between our ground and signal wires coming from the throttle. This reduces the throttle input to the controller. So when you get down to 10%SOC you have to give it full pedal to get the car up to about 20 mph. This is an excellent way to signal there is a problem (like recharge the car). I’ve actually tested this and you can indeed “limp” home although it is slow.

We’ll try to do better on our weekly videos. Lots of stuff going on and happening.

Jack RIckard

29 thoughts on “New Instrumentation for Speedster”

  1. Jack,
    Glad the bmv worked out. Did you get the pc connection cable and software? Kinda helps with the setup and you can read the data with a laptop as you drive. I tried out the victron global remote option but didnt find it good. Problem is the minimum update time for it to send data to the webserver is 10 seconds so it misses a lot of data. I did find the sms function useful for checking state of charge but returned the unit as I couldn’t justify the cost.

  2. Bummer that the Chenic/Elcon charger won’t be shipped with more than one setting. I was planning on multiple just in case I decided to do some minor changes to tweak the charge for the pack. I like the idea of the new meter setup. Looks like something I will set up in the future. Perfect size and fit for the Ghia. Glad to see you guys braving the heat. Glad to know your still producing and testing. Hope the cooler weather comes soon. It’s hot here too.

    Pete 🙂

  3. Regarding the power supply to the TBS/Xantrex battery monitor:
    I have made a circuit which will supply the 12V/10mA from the main battery (in my case 85 – 128V nominal 102V) with just a 12K/5W resistor in series. A 12V/9Ah battery is connected via a diode to the TBS and will supply anything above the 10 mA. The a float voltage to the 12V battery is also supplied from the main battery (just a 360k resistor). I have not ydecided if I want to put a small 12V charger in as well, but I might just incease the float charge current to 1mA and add a zener diode across the battery to protect it from overcharge.
    The prescaler will also be located on this small prototype PCB.
    The total amount of current draw from the main battery will be about 10mA or 1W.
    I think the pricetag of the BMV600HS is a bit on the high side.
    My solution is about $100,- lower but it requires ofcourse a bit of soldering.


  4. Jack,

    regarding your problem with the compressor-controller: Could it be that the controller has no input diode? Going from 380 – 340V is more less equivalent to go from 40V to zero while shorting the input. Without an input diode, the caps in the controller would discharge very quickly into your main battery pack and blow the fuse… Perhaps you could try to add a diode preventing the current to flow out of the controller back into the batteries…

    Hope this helps,

  5. Markus:

    Well, we don’t blow the fuse going from zero to 375v, so I’m a little unclear as to why we are blowing a fuse going from 375 to 345. Of course, the motor is also running at the time. It might be easy enough to try your solution. Let me root around and see what I’ve got in the way of high current diodes….

    Jack Rickard

  6. Markus:

    You are a goddamn genius Gump. I rooted around the work bench and found an old very heavy current diode that I got some of on eBay, and never really did have much use for. I really did NOT think this would work, but it was easy enough. The diode has a negative plate about 2 inches long and a half inch wide, and then two positive terminals on the top. I just bolted the plate to the contactor terminal and connected the positive lead from the battery pack to the positive terminal on top.

    I had zero expectations of success. But I took it for a drive, driving somewhat gingerly, and the drop in RPM was NOT happening. So I took it out on the freeway for some more aggressive work, and it absolutely does not drop in RPM and does not blow the fuse. YOU FIXED IT.

    It would never occur to me, but in arrears of course it is pretty obvious. If I charge the input caps to 375v and they are sitting there operating, and I drop the source to 345, they are going to try to help drive the motor until they get down to 345. I guess if it happens suddenly enough, they will put out enough current to blow a 30 amp fuse. Meanwhile, the compressor gets NO current and goes into a stall?

    In any event, you cannot imagine how delighted I am with your very simple solution. All of the things I was thinking I would have to do to get this thing working (not having A/C was never one of the options with this car and in this town) and suddenly, I don’t have to do any of them. They were all ugly.

    So tell us a bit about yourself? Do you have a car or project now?

    Jack Rickard

  7. Jack,

    glad I could help. I have no car project right now, but quite interested in starting one (I almost bought an old 911 to convert a while ago though). When I finally start, it will be a AC drive for sure, either copying Eric Tischers idea of modding a industrial VFD or a complete home built controller. I like to play with electronics…

    Enjoy your AC (I probably need to travel to St. Louis, MO tomorrow, so I checked the weather forecast and it will be hot in Missouri the next days…)


  8. Just thought I’d point out that clutchless shifting with regen is possible, I do it with my clutchless Fiero/AC31 setup. It’s best to wind up the motor and do the shift very aggressively, but it does work. I have had a few grinders, and most of the time I just leave the car in second. Downshifting is also possible, though a bit more difficult.

    You might take a look at the EV Display:
    Hall effect pickup, AH, WH, AMPS in and out, VOLTS, TEMP, KW, SOC, “Fuel Gauge”, $170


  9. Jack,

    The controller does have a precharge circuit that limits the current while the controller capacitors charge. As for operating at 375V, this is the exact same controller that is in the Tesla Motors Roadster. We are currious about the failure of the controller and will advise Revolt and you what we find is the cause. We are waiting for it to come back to us for analysis.

    Tony Carstensen – Masterflux

  10. Thought I’d point out that the EV Display is powered from either 4 lithium cells, your 12V battery or DC/DC, or you’d need a scaler, so it won’t solve your specific issue. I power mine from my DC/DC as I use the 840 to show voltage. I don’t need the EV Display to cut back power on low SOC as the Curtis controller seems to do that quite well on it’s own as I discovered the one time I pushed the range a little too far. The EV Display uses Cat5E or Cat6 cable to connect the display to the sender board.


  11. Mr. Carstensen:

    I’d be fascinated with what you might find. We have had a total of THREE controllers in this vehicle. All of the controllers exhibited the same problme, which seemed to be a separate problem from the original compressor making a huge racket and eventually failing.

    On acceleration, the RPM would drop audibly to a very low value or drop out altogether. Later examination at the shop would show a blown fuse, or eventually a dead controller.

    We added a Schottkey diode, 200A 100V to the positive input and the problem went away completely. Apparently there is nothing to prevent the capacitors from dumping current BACK into the pack when the pack voltage sags from 375 to 340 on acceleration.

    Jack Rickard

  12. JRP3:

    Interesting connundrum. I ran into to it today with a little hall effect device from Australia that is supposed to drive a regular gas gage. It immediately failed to work on ours of course. But:

    1. It requires constant 12v power.
    2. It resets if 12v power is removed.
    3. It is very necessary to set zero current point when there is NO draw on pack.
    4. If we power it with 12 v from the DC converter, we ARE DRAWING POWER FROM THE PACK and throwing off the zero current calibration. If we turn the pack off so there is no current, we have no 12v and the device resets.

    IT gets a little circular. I guess we could power our 12v subsystem with a battery while calibrating it, but it seems a lot of trouble.

    This device has the exact same problem, along with the usual power problems. But it looks pretty nice.

    ISn’t ANYBODY doing a good instrumentation system for EV’s?

    Jack Rickard

  13. Jack,

    We have analyzed the controller and we found it to function properly. We assume that the current outflow from the capacitors causes the fuse to blow and may also casue the controller to go into an over current shutdown. This condition cause the controller to not function until all of the power is completely removed.

    We are doing some field testing on this on Monday and will report the findings.

  14. “causes the fuse to blow and may also casue the controller to go into an over current shutdown”

    The fuse blowing caused the unit to go into a current shutdown?

  15. Tony:

    All of the power is completely removed as soon as the contactor switch opens. But we couldn’t get it to come back online even with a new fuse.

    That said, it doesn’t function “properly” in an EV if every time you accelerate it blows a fuse and goes into an overcurrent shutdown.

    The solution is pretty simple, put a high current diode on the positive input. That prevents the caps from being able to outflow.

    Jack Rickard

  16. ??”…3.44 rack and pinion and gears…”??
    Think you mean RING and pinion. RACK and pinion is generally used to refer to the part for steering the car. RING and pinion is used to refer to the gearing in the final drive.

    I am a long time gearhead / petrol head but have to admit I am getting interested in doing a conversion. Love the site and the way you real world test and report the good, the bad and the ugly.

    One of the other sites I really appreciate is
    The info on aerodynamics applies to all types of vehicles in order to operate more effeciently.

  17. Jack,

    We completed our field testing and verified the main issue with the compressor controller is the current ripple. In our spec. we specify that the max. current ripple is 6A RMS. Anything above that and you can get into the situation you had. We also found an inexpensive diode that can be used if you cannot keep the current ripple down. POWEREX – CS240650

    Another observation was that the location of the power leads to the controller near the main traction motor controller caused these large fluctuations. Locating the power leads at the battery itself seemed to reduce the ripple.

    In any event, I am thankful for your information on the problem and for your patience with us.



  18. Tony:

    Thank you for the confirmation and the part number of the more appropriate, and quite less expensive diode.

    I would suggest that some mention of all of this in your documentation would be in order and potentially save you and them a good deal of potential grief..

    Jack Rickard

  19. Jack:

    I just installed a CycleAnalyst in my Gizmo and can finally see what happens to my regenerative braking results. I haven’t done the same drive twice using regen and without but I did some preliminary observation of Ah just before regen and immediately after coming to a stop. If I regen to a stop starting at a speed over 20-25mph I end up with a net positive energy benefit. If, on the other hand, I regen starting at a speed below that I end up with regen being a negative result. On my SepEx system I have the Sevcon Controller programmed to hold regen as long as possible. My main concern is brake wear not range extension. At around 16mph I am at a net zero current which then goes negative (out of the battery) at speeds lower than that. I see peaks of 25A or so just before the controller cuts out regen. I don’t know if this is just a characteristic of SepEx motors doing regen or if all motors have some rpm point where it takes energy out of the battery to keep slowing the rotor down. With your driving style and you getting a net negative benefit from regen I’m wondering if the AC motor system is doing the same thing. If so, maybe you can program it to stop regen when the current drops to zero and you might see a greater benefit. Have you noticed anything like I describe in my system?

    David Nelson

  20. I’ll close the loop here, although it was posted later. The air conditioner problem was solved by one of our viewers. We added a diode to the input to the compressor controller. Problem fixed.

    It wasn’t that it was drawing more current IN at the lower voltage. The caps in the controller were passing current OUT to the drive train in an apparent attempt to power our drive motor. If this voltage drop occurred fast enough, the outsurge blew the fuse on the controller.

    50A 200v diode is the cure.

    Jack Rickard

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Verified by MonsterInsights