DIY Electric Car Forums banner

1 - 20 of 33 Posts

·
Registered
Joined
·
7 Posts
Discussion Starter · #1 ·
For all (or any?) of the Soleq EVcort owners out there,
I have replaced the Soleq sepex motor controller in
my EVcort with a pair of Kelly controllers, and outlined
the project here:

http://home.earthlink.net/~evtkw/

While the Soleq Controllers were an impressive design
feat, they tend to fail periodically. Repairs to the
all analog, 100 pound controllers have been unsuccessful for
me and others I have contacted.

My new configuration works fairly well, and I am able to
get 200+ Amps regen and decent performance. I guess the
system could be used for any sepex motor.

Let me know if you have any questions about the system,
and I will be happy to give any information I can.
 

·
Registered
Joined
·
203 Posts
Why not one of the kelly sepex controllers? Any more info on the 2 controller rig? I didn't see any on the website (at a quick glance).
 

·
Registered
Joined
·
7 Posts
Discussion Starter · #3 ·
The complete design, including pictures, schematics, and configuration, is on the web site. I hope it is working properly for different browsers...

This car requires up to 400 A of drive and over 200A of regen at 108V. I don't believe that any controller maker offers a sepex controller with this kind of full size EV capabilities. All of the off the shelf sepex controllers are for golf carts.
 

·
Registered
Joined
·
203 Posts
The complete design, including pictures, schematics, and configuration, is on the web site. I hope it is working properly for different browsers...

This car requires up to 400 A of drive and over 200A of regen at 108V. I don't believe that any controller maker offers a sepex controller with this kind of full size EV capabilities. All of the off the shelf sepex controllers are for golf carts.
I looked again and it's all there, right at the top. Duh.

How about "KDC12603,24-120V,600A, Separate Excited with Regen"?

Is this a newer offering than your setup? I was planning on purchasing one of these soon, or rolling my own from two controllers like you did. Your experience is certainly helpfull
 

·
Registered
Joined
·
7 Posts
Discussion Starter · #5 ·
This seems to be a new controller since I built my system. I originally tried the model below this, the KDH12403, and it didn't have enough power for my fairly heavy car. However, they have upped the current, so It may work, depending on your application. I would talk to the Kelly tech support to see if they think it will be good for you as well - they were very helpful to me when I was building my system.

If it is powerful enough, it certainly seems to be a good option...
 

·
Registered
Joined
·
3 Posts
Looks like you removed quite a bit of weight. How much NET loss did you have?
 

·
Registered
Joined
·
7 Posts
Discussion Starter · #7 ·
I'm not sure how much weight was removed - maybe 150 pounds? There is a big box of parts that didn't go back in, and I can barley move it...
 

·
Registered
Joined
·
312 Posts
evtkw - very nice solution and very nice conversion

For people attempting to do something similar:-

If you want to control the field weakening automatically, you need to monitor armature voltage and motor speed. If you reach maximum armature voltage ( actually we normally use 95% in industry), you start weakening your field until you reach your minimum field current at maximum motor speed. For this you need some form of speed feedback device.

I hope this is helpful
Dawid
 

·
Registered
Joined
·
312 Posts
Major

You do not need any feedback device if you do not do field weakening with torque control. If you use a setup like mentioned here, where you have a manual control over your field weakening, you will likely only experience a loss of torque if you forget to reset the field back to a higher value. However, if you perform a similar mod and want the field current to automatically adjust, you cannot do it without feedback, or very clever software that can deduce motor speed without actual measurement.

Regards
Dawid
 

·
Registered
Joined
·
7,791 Posts
You do not need any feedback device if you do not do field weakening with torque control.............if you perform a similar mod and want the field current to automatically adjust, you cannot do it without feedback, or very clever software that can deduce motor speed without actual measurement.
Hi Dawid,

The commercially available SepEx Curtis and Sevcon units will let you select speed or torque control. It has been a while and I don't remember which I used on which vehicles. I do know I tried both ways and neither required speed feedback. In fact, neither had such provisions. Now maybe those did have clever software. On the other hand, I used a SepEx GE golf car motor with a Curtis. The motor did have a speed sensor. So I suspect the programmers at GE aren't so clever and use speed feedback on their SepEx controllers :)

I do not know how Tim is set up to match field and armature currents. But I do not see where he has a speed feedback in his controller installation. And it is unclear if he has any automatic adjustment for the field or if it is all manual.

To me, the important parameter to monitor when controlling the field of a SepEx motor is the armature current. This is done in the Curtis and Sevcon by entering a field map. Once you have this, there is no need for shaft speed. Maybe it is clever programming, but these controllers are able to set speed limits as a percentage of maximum speed and get reasonable results. I actually had 3 different speed limits on my vehicles, fast forward, slow forward and slow reverse.

I do not see the need for speed feedback unless precise travel speed is critical for your application. And I imagine it is better and easier to properly program the controller using currents instead of speed, although I have never done it.

My 2 cents worth,

major
 

·
Registered
Joined
·
7 Posts
Discussion Starter · #12 ·
I used the most simplistic system - just a knob on the steering column that controls the field current. I set the field to full current up to the motor corner RPM for the full field current. I then reduce it to increase motor rpm manually. This was basically the only system available 2 years ago for a high power system - at least that I could find.

I have considered adding a small circuit to automatically adjust the field current based on motor RPM - I already have an RPM sensor that feeds into the stock tach, so it seemed the easiest solution. I thought to do it in 3 to 4 steps, so it would feel like an automatic transmission...

Of perhaps some interest is how the original Soleq controller operated. The controller would run in torque mode with maximum field current up to the corner rpm for max current, with the accel pedal controlling the armature current. For higher rpm, armature current is no longer controller; the controller adjusted the field current to control the armature current based on pedal position - so kind of a constant power controller.
 

·
Registered
Joined
·
312 Posts
Major

I had a look last night in some of my references. Here is a quote from one of them:

It is possible to duplicate this characteristic in a separately-excited motor by controlling its field current by a thyristor drive configured as a current regulator, the reference being derived from being derived from the motor's armature current through either a shunt or a DC current transformer (DCCT).

The advantages of this technique increased motor output (since the resistance of the field windings is not connected in series with the armature) and the facility to set minimum and maximum limits of field current, thus preventing saturation of the magnetic circuit and improving performance under light and overhauling load conditions, e.g. downhill running.
It would seem to me that for traction applications it would be easier to use the armature current reference as a setpoint, which would make it quit easy to use a seperate, small controller as a field controller in a sepex setup. The book referenced also took about field reversal, which is a quick way of reversing a motor, instead of the armature, which would require bigger contactors.

Regards
Dawid
 

·
Registered
Joined
·
251 Posts
Hi !

Here is also a controller that might be useful for those SepEx motors running over 100v.
http://www.elektrosistem.com/e/p2.htm
The Phoenix can be gotten in 120v versions both for the 600A and the 800A

First hand info, you can get from a guy here at the forum.
Search for volvo 440.

Best Regards
/Per
 

·
Registered
Joined
·
46 Posts
This is my first post here !

I was involved with a field controller for a Sepex 36vDc @400A controller back in the early 80's. At that time I used a Motorola MC34060 PWM chip.
The chip had the usual goodies - on chip osc. and 5v precision regulator - but with the added feature of having two op amps feeding the PWM generator.These error amps were active high and wire or'ed together at the active input of the PWM comparator. Thusly whichever commanded the minimum output would dominate the loop.

Accordingly one op amp was fed the max field current reference on one input and the actual ground referenced field current feedback on the other. The field current sensing resistor was a 0.1 ohm able to carry the 9 amps of max field current IIRC.

The other op amp was supplied 0 to five volts from the accelerator pot which represented the 0 to 400 amp armature current command.

There was one other input to this op amp. This was the armature current feedback and it was a little more tricky. For this, we took the Kelvin connections from a 500Amp 50mV meter shunt located in the path between the armature negative lead and battery ground, and fed them thru' matched 1k resistors into one section of an LM358 acting as a differential amp with similarly matched 100k feedback resistors - supposedly avoiding any unexpected offset voltage and also providing a voltage gain of 100. The LM358 op amp output was then scaled with a preset pot to yield 1 volt per 80 amps before sending it on to the PWM chip.

The single ended PWM output was fed via a base current limiting resistor to a TIP147 plastic PNP 10Amp transistor which sourced current from the high voltage bus to the hot side of the field winding. This technique facilitated the ground sourced field current feedback on the other side of the field winding which I described earlier. We wired the field with RF cable but 14Khz with 85nS risetimes pretty much wiped out AM reception on the vehicle plus any nearby vehicles also !

Once we got the RC network across the chip's armature error amp sorted out the operation was quite stable. And driving above base speed with the armature controller bypass contactor energised, we could set the motor amps instantly anywhere on the meter that we wished.

I was able to brag at the time that we were indeed able to control 400A with a $1-50c transistor and a $2-30c PWM chip. Then I happened to mention that field control above base speed could provide 400A of regen virtually for free, even 1000A no biggie.

So now I really needed to have a voltage offset at the output of that LM358 which I had previously took pains to avoid. Accomplished, in the event, by lifting the leg of the ground ref'd 100k resistor and joining it to +2.5V pedestal which would similarly raise the op amp output by 2.5v. This gave the accelerator pedal input still sweeping from 0 to 5v effectively a +/- capability about zero armature current. Well we did it. And it worked. Our admittedly limited powertrain gave field control from 12mph to 45mph in 1st gear.

Looking back I often wondered why those members with series wound motors didn't consider rewinding them for shunt fields.
 

·
Registered
Joined
·
7,791 Posts
I was involved with a field controller for a Sepex 36vDc @400A controller back in the early 80's..........Looking back I often wondered why those members with series wound motors didn't consider rewinding them for shunt fields.
Hi toy,

Welcome to this forum. You have a pretty good memory for details about your 25 to 30 year old project. Do you remember what motor you used?

Anyway to address your wonderment: The primary reason is that there are no readily available motors and controllers for what I call medium voltage, on-road power range. That is like 120 to 200 volt battery packs and up to 50 kW. And of course, everybody wants more power. And low cost.

So what is most used is like a 9 inch diameter series wound motor. It is basically a forklift motor which has had the brushes advanced to successfully commutate at higher voltage than it was designed to handle. Since these guys use the transmission from the donor car, the motor is unidirectional. Several companies offer PWM controllers to suit these applications. A downside is that series motor/controller systems do not regenerate. Everybody seems to want regeneration, but in fact few would benefit much from it.

So you have done a low voltage low power SepEx system and wonder why these guys don't rewind and go SepEx. First, who rewinds for them? Who designs the field? Where do they get the controllers? And then how do they deal with the brush advance and commutation? Easy, you might say. Go to interpole machines. Where do they buy interpole machines? And who funds all this?

I suspect if you had a source which would offer SepEx regen capable motor/control systems at a similar cost (perhaps a slight premium), you would get a fair amount of business from the DIY converter market. So, put that good stuff in your memory banks to use and do it. See if I'm right. :)

BTW, I have a 96V, 800A series interpole system with regen. It's a good 30 years old and uses SCR controller. I've also done a number of 48A, 400A SepEx systems in the last decade. I have been really been impressed with the modern SepEx controls. The old series regen system sucked in comparison. But then again, it is totally discrete technology. Programming by soldering in a different resistor :(

I've also worked with ACIM systems. That's the way to go, if you can get it ;)

Regards,

major
 

·
Registered
Joined
·
46 Posts
Part One

First, who rewinds for them? Who designs the field? Where do they get the controllers? And then how do they deal with the brush advance and commutation? Easy, you might say. Go to interpole machines. Where do they buy interpole machines? And who funds all this?

Hello major,

thanks for responding. Actually I have a poor memory but these projects stand out in my mind than what I do for a living. Anyway to answer your question...

The motor was either a GE CM377 400A 28Vdc ASG or a similar Jack &Heinz which we sourced from Princess Auto in the years when they were putting ex War Department equipment into consumer channels. Those aircraft starter generators were quite plentiful in 200A, 300A and 400A versions. I would imagine there are quite a few 'survivors' even today - I have one still. Some were 6 pole and had fluted commutators - apparently done for high speed use.

We were able to procure the declassified manuals on them as well. It should be said that although deprecated by some, these examples of 1950's aviation history were in fact excellent machines ! Granted, their armatures were stiff as molasses to turn - primarily due to excessive (to us) brush pressure, but when I checked it out they fell within the max lbs/in/in that was stipulated. Of course with a 3000 Hp aero engine shaft just a couple of feet away the generator brush friction was probably never a cause in consideration for aero fuel efficiency particularly when it came with the promise of the highest reliability. Needless to say, that didn't stop attempts, and ultimately futile attempts if I remember, to secure a source of weaker spring steel for that delinquent brush gear.

Because of the high unloaded shaft torque and of course the knowledge of its previous life as an aircraft engine starter, it gave cause for some to believe that it was more of a low speed / high torque motor. In fact the motor was qualified for service at the low figure of 66lbs-ft despite a gas guzzling 1000A although calculations said it should be good for around 88Lbs-ft. Ah..... perhaps maybe the 22lbs-ft shortfall was that brush friction !!

But Low speed machine ?? If you run the math taking into account a propellor speed of 2500 RPM then a 3:1 planetary step up would certainly have it spinning at a fast clip. If its operational life does in fact find it mostly within the 8000-7000rpm range then clearly it fits more closely in the domain of the high speed machine. I notice even the Tech Manual supports its legacy as a high speed machine specifying an electrical top speed of 10,000 rpm and requiring mechanical certification to 11,000 rpm.

When you read the manual it makes you want to see the test rigs that they were using to recertify these machines and would be required for the regular mandatory maintenance.

Unfortunately most of the ASGs that fell into the hands of EVers found service coupled to the clutch plate of a 4-speed transmission where they rarely saw much more than about 2250 rpm before a gear upchange.

When I came on the scene, however, there was no suitable electronic field controller in existence for EV's to allow the upper register of the motor speed range to be put to use.

As a first step I started in on the construction of a PWM field supply. This conferred the ability to fix the shunt field current without regard to variations in the winding temperature nor the voltage rail. The task of always having to find a suitable ballast resistor also went away. My sales pitch was that PWM would give us the freedom to ignore the previous conservative field setting and place our shunt field excitation at an optimal setting. The intended topology was to utilise a high side switch but I also knew to do that I needed to overcome the problem on the ASG winding schematic which depicted the negative shunt field lead joined internally to the negative armature lead.

This join, wherever it was hidden, would also be covered in baked varnish. It was possible to make out several well crimped and soldered joints sheltering underneath blobs of varnish The enabling technology was to discover how to inject a current thru the motor so we could estimate which one of these junctions contained the elusive negative field wire. We used sharp probes to penetrate suitable candidates. You're looking for millivolts here. Then we needed to reinsulate everything that had been dissected and rebake the motor. But that needed the whole motor to be disassembled in order to bake the stator alone. What a job ! We did two other motors later on. One of which was actually miswired and the mistake covered over in varnish. Hmm.. so much for aircraft quality standards !

A bench test was now conducted for armature iron loss evaluation whereupon it was determined that the maximum armature voltage should not go much above 36Vdc. At least that explained why some who had been running their motors at 72Vdc found their motors running hot. We were able to announce that at 72Vdc (4500rpm) we had been able to separate out, somewhat unexpectedly I might add, approx 2.2KW of armature iron loss.

Previously the impression I wanted to leave from that post was to give a descriptor for simplicity that a SEPEX field controller could be constructed using readily available parts.

In this post since you asked about the motor I thought to add that a SEPEX motor can throw a curve ball at you. Field control is not necessarily plain sailing just because you get the electronics right as this post attempts to outline.
T2
 

·
Registered
Joined
·
46 Posts
Part Two

major you ask And who funds all this ?

Well, as I saw someone here succinctly write (on controller pricing)

"It is not the underpriviledged who get into electric car conversion because they need to save money on gas. This is an expensive hobby."

There is, however, always going to be an element who want to do something with an aged clunker, a motor from a fork lift, and some discarded ignition/lighting batteries they happen to have lying around and need advice now on how to pull it all together with a large roll of duct tape. I am not saying I am against the idea of helping out the underdog, but sometimes I feel that I am being asked to advise the captain of the Titanic - after he's hit the iceberg.

---originally Posted by toddshotrods
The current DIY EV community is a strange dichotomy of green and/or economy minded transportation and racers. The two don't really mix, and their needs are on opposite ends of the spectrum. Making things more complicated for the racers is the fact that the market is dominated by the green/economy side.

I think that toddshotrods has nailed it right there. As you and I both know, you can't find a reliable controller at the lowest price point. It doesn't exist. However, that fact alone has not stopped more than a few falling prey to those rascals who talk up their under tested prototypes and then inevitably either haven't provided themselves with enough financial cushion to honor warranties or just flat out refuse service. I fear the chances of their clients re-entering the market again to more legitimate providers is slim to none.

What is the incentive to enter the controller market and build an excellent product like over at Cafe Electric ? That vendor is then left with the much smaller market niche of people willing to fork over $5k or $6k. Couple that to the fact that as a vendor you are still at risk from doubtful wiring practices since you can't be sure that qualified people will be overseeing the installation of your product.

First, who rewinds for them ? I haven't researched it but I am sure most stator rewinders could work on DC windings. What am I missing ? Isn't it just the same number of ampere-turns going from series to shunt. The interpoles don't enter the equation surely ?

Continuing on the rewinding topic - because I happen to think it's important. Seriously I think we should be of the mindset to rewind for lower voltages and safety. Costs aside, what the ultimate reliability of a 576Vdc string of li-ion batteries is going to be will be interesting. The MTBF of these cell packs has not yet been quantified.

I'll put it out, in case you may not be aware, that ACMOTOR (Red Suzi project) on the AEVA site, had an 11Kw ACIM rewound from 415Vac to 100Vac in August '09. Sure he had to shell out the $700 plus the extra $200 for three embedded thermistors and rotor balancing for high speed, but he reports significantly improved performance using a 10.5:1 ratio. That's a drop from 8V down to 2V per Hz.

In essence ACMOTOR took a perfectly good running motor and rather than regard it as a done deal he proceeded to put a $900 investment into it. Whereas others would rather have seen the money go on something more tangible in the form of more batteries for 830V battery strings, he has chose this way to go.

I suggested to him that next time he might think of going as low as 0.75 V/Hz like GM's IMPACT on his next project. He responded that he is standing pat for now since this would require an upgrade in both controller and batteries.

Getting back to SEPEX motors frankly I feel their time has come and gone. It would be interesting to construct a compact side by side two motor transverse cradle able to drive each rear wheel individually as Greenmotor Sports has done but with SEPEX in place of ACIMs.

In '85 after talking with staff at Transport Canada I submitted such a proposal along with photographs schematics and graphs.

In my proposal, based on my field controller work, I outlined how you could utilise two motors in a current sharing scheme. It proceeded in three phases. The system would start in Phase 0 with the motors in series with a single armature controller and their fields set to max.

At 10mph or 100% modulation thereabouts Phase I would be entered for field control. In quick succession, the bypass contactor would short out the armature controller, one field would be operated to make its armature pull constant current under control of the accelerator pedal while the other field controller would be referenced with a Vbat/2 resistive divider and the actual midpoint voltage. The second motor would therefore act as constant voltage device. I doubt you would disagree that it makes for a stable system to have a constant current source driving through a constant voltage system.

Phase II, the final phase would be entered when field current has dropped to a (previously determined) level which has effectively halved the flux in the two machines. At this point restoring both the motors back to their maximum field current should enable each of them to support the full battery voltage alone. They may therefore be switched in parallel and their analog inputs suitably arranged such that both receive the same pedal input with the second motor field controller now receiving feedback input from its own armature current sensor.

I would use a mix of both microprocessor and hardware supervisor ccts plus additional commutating and quench diodes to ensure integrity of the switching operations when sequencing from Phase I to Phase II.

A racer might use a 500 amp armature controller to get the vehicle rolling to around 30mph before the field controllers can take over and then run a constant 2000 amps or whatever throughout the last two phases. The system doesn't rely on speed feedback either. Although personally I have no problem with a Red Lion setup with a toothed wheel, a two wire sensor and an LM2907 chip, just don't ask me to do the mechanics of it !

The concept appeals to me because you get to use the shunt wound machines as rotating amplifiers thus avoiding the armature current rolloffs experienced when making the same switch with series field motors. Then there's the avoidance of stressing a large array of semiconductor devices under extreme conditions. Of course the system is amenable to being brought up at greatly reduced currents and system voltages since I can see it would really harsh someone's mellow to inadvertantly zorch a siamese set when playing around in unfamiliar territory !!

BTW anyone who thinks I went off topic, go see the first post in the thread.

For all (or any?) of the Soleq EVcort owners out there, I have replaced the Soleq sepex motor controller in my EVcort with a pair of Kelly controllers, and outlined the project here.


The originator of the thread hi-jacks his own topic in the first sentence. evtkw, man you should get some sort of an award for the fastest hi jack in history.

Bye for now
T2
 

·
Registered
Joined
·
7,791 Posts
First, who rewinds for them ? I haven't researched it but I am sure most stator rewinders could work on DC windings. What am I missing ?
Hi T2,

You sure write a lot :) But this one question is the only attempted answer I can see. I'm sure most stator rewinders could rewind DC fields. But will they? I think you'll find that most motors of the size we're talking about are throw-aways as far as rewinders are concerned. Motor shops which specialize in forklift motors even buy replacement field sets as opposed to winding. Yeah, there are places to do it. But finding one is no easy task and you're likely to pay more for the rewind than you did for the motor :(

Isn't it just the same number of ampere-turns going from series to shunt.
Well, yeah, kind of. At what load do you consider to figure the AT for the series motor? And at what voltage do you use to figure the SepEx field?

The interpoles don't enter the equation surely ?
Interpole windings depend on the armature. But the point is, the series motors most of these guys have do not have interpoles. They get away without interpoles because it is a series motor and they have advanced the brushes. Change that motor to SepEx and vary the field and reverse it or generate with it, and you have big commutation problems.

Good example is that aircraft generator you mention. Who would have thought a 24 volt 8 kW generator would need interpoles? Forklifts use motors in this range without interpoles all the time. Why then does that aircraft generator not only have interpoles, but also compensating windings? Because of the varying range of excitation and load and the fact that it is also used as a motor for engine starting.

I like SepEx systems. I wish there were such systems available for the EV in the voltage and power range needed. But few have the means to rewind the series motor fields and design a field controller. And contrary to what David thinks, you still need an armature controller. So after all the trouble and expense involved in converting the series to SepEx motor, you still have the basic system (meaning the expensive armature controller) which you had to start with. And for what? Regen? Yeah, nice. I like regen. But where is the value here? A mile further on a battery charge? That is an expensive mile ;)

And you quote Todd. Good example. He is building a hot rod EV. And has a monster GE SepEx motor he wants to use. However he is considering having it rewound to series because he cannot get a proper controller for it. You want to prove your point, make a controller for Todd if it is so easy.

Regards,

major
 

·
Registered
Joined
·
4,458 Posts
...And you quote Todd. Good example. He is building a hot rod EV. And has a monster GE SepEx motor he wants to use. However he is considering having it rewound to series because he cannot get a proper controller for it. You want to prove your point, make a controller for Todd if it is so easy.

Regards,

major
My ears were itching... ;):D Yeah, I'm probably pretty well on my way to series wound. I thought I had purchased a series motor, was disheartened to find out it was SepEx, then determined to explore its separately excited promise - but the return on investment (of money, time, and effort) is just not there. There are just too many variables to make it a worthwhile pursuit, especially for a community of wide-ranging goals and ideals.

Series is the best universal solution for DC. The same controller that works for the guy with a grocery getter can push my racer down the track, or vice versa. True, he may not need that much controller, or I may not have enough (with his) but the point is they're almost plug-n-play for whatever you want to do with them.

As for my quote you posted - it proves the point against SepEx. The people who would pay for it - like me - won't see the benefit in it. The people who could really use it, won't pay enough to get it.
 
1 - 20 of 33 Posts
Top