SalvageS10 Build Thread

[thingstodo]

Originally Posted by Ken Fry

Although it might not be as much fun, finding one of those would save you a lot of time. Also, because the motor was designed as a traction motor, it is relatively light.

I want to use parts that I can source, and understand how to fix what I end up with, so I guess I'll be going with the 'more fun' way.

Originally Posted by Ken Fry

The EVDL Forum could offer help -- there are loads of S10s and Ford Ranger pickup conversions out there.

EVDL Archive / Forum Interface - Electric Vehicle Discussion List

This forum and the evalbum are what pointed me to the S10 to begin with. I will be experimenting without a transmission ... and will likely verify that a standard transmission is the way to go. I don't have one, as the truck has an automatic right now.

[thingstodo]

Originally Posted by Ken Fry

...30% grade is a reasonable minimum hill climbing capability. So, you will want a tractive force of 1500 lbs. ...

I'm aiming a bit lower than that. None of the roads this commuter (if I get a range long enough to drive to work) will see will be steeper than 3%.

A 40 HP electric motor rated at 1770 rpm is about 120 foot-lbs of torque. From the specs on the motor, it peaks at 220% rated torque or 264 foot-lbs, but that assumes that it is starting from 0 rpm and the motor is started 'across the line' with a starter and a utility pumping up to 10 times rated current through it. I think I can reach 264 foot-lbs with a VFD and about 3 times rated current, or under 120 amps.

With the 3.43 ratio and 205/65 r14s (I measure them just under 30 inches in diameter so I use 15 inch radius) that will give me about 725 pounds of force to the road.

There's a lot of assumptions wrapped up in there. I'll see what I have when I get there. And I'll have some fun measuring everything along the way.

[thingstodo]

Originally Posted by Ken Fry

You need to decide what sort of top speed you want, to see if the gearing for 30% grade climbing will also give you adequate top speed (based upon the motor's max rpm). If not, then you'd need more than one speed in the transmission.

The off-the-shelf three phase 575V T frame motor is rated for 1770 rpm / 3.43 rear end * 30 inch diameter * pi converted to mph is about 45 mph. The motor is constructed the same as a 3600 rpm motor, so overspeeding it will reduce the life of the bearings, but that's about it.

Below 1770 rpm, theory tells me that the output torque is constant at about 120 foot-lbs at rated current (hopefully up to 264 with more current and a VFD). Above 1770 rpm the torque will be reduced - the theory that I've seen predicts approximately linear reduction in torque as speed rises. I'll run out of horsepower, even with extra forced-air cooling, hopefully above 70 mph but I'm sure that I'll run out of battery before that.

We'll see how close the theory matches reality.

[thingstodo]

I decided to try one of the inverters that I use, that works OK so I can get an idea of the signals. I can always go back and investigate the surplus stuff. The manufacture date is 2010.

The reason that I'm taking apart DC/AC inverters is to see if I can use the DC/DC converter part ( the first part) to boost the 12V up to 150 VDC, then chain them together in series to build up 300 VDC (for a 208VAC three phase test inverter) and eventually 900 VDC (for a 575VAC three phase inverter). Along the way, I'd like to do some testing on the inverters to determine the efficiency (estimated around 85%), what cooling is required (is a fan and a heat-sink enough?) and what sorts of things can be done to deal with a DC/DC converter failure while operating the electric vehicle. I have some crazy ideas I'd like to test out.

Back to the topic at hand

The ends come off after the screws are removed, then the bottom plate slides out the end to reveal a board mounted on standoffs from the 'top' of the heat sink.

The setup is remarkably similar to the Wagan, on a smaller scale. Perhaps this is an optimal design. I'd guess that some of the same people were involved in the design.

There are only four 40A fuses instead of 16. They are not associated with each of the 6 transformers, but are all in parallel. There are diodes visible beneath each of the transformers. There are two power transistors for each transformer. The output section has 4 power transistors, with only three larger capacitors in that area.

The electronics to switch the transistors appear to be distributed around the board, in areas that are not required for the power section. The control signals would be much more difficult to separate, since they are part of the multi-layer PC Board instead of routed through ribbon cables.

The power transistors are electrically isolated from the heat sink with a barrier that must be pretty good at thermal transfer. The transistors do not use the tabs, but rather have a mechanical clamp to hold them firmly against the barrier.

There are number of places on the board that are marked as resistors that have 16 gauge copper jumpers soldered in place.

Silly things which appear to be mistakes:
- the cooling fan has a cable pinched between the fan case and the bottom of the enclosure. The fan needs to be rotated 90 degrees and the cable would not be pinched.
- there is a pinch in the insulation of the output for outlet 1. There is clearance, but it was pinched anyway due to no cable management.

Connect the power supply to the DC to AC inverter. Turn on the inverter. The inverter powers up, then turns off? Check the voltage output from the power supply. 17.7V. Check the alligator clips, voltage at the inverter input terminals - all good. 17.7V was within the 10 - 18VDC for the other dc to ac inverter. The voltage range on this one is not printed on the outside of the case. Try lowering the voltage to 16. The inverter powers up, then off, then powers up, then off. There is an audible alarm that I think means the inverter input voltage is low. Check the voltage - it is dropping to 10V, the inverter powers down, then back up when the voltage rises above 13V, and the cycle continues.

The inverter appears to require more current than the power supply can put out. 1.1A is not enough.

Change power supplies. This one will do 5A at 20V max. The DC current in is around 2.5A at 16V to power up the inverter. I may have to revisit the other inverter. It is larger and likely needed more current to 'turn on'. Perhaps later.

Measure the output voltage - 0VDC, 0 VAC. What? It turned on. There should be voltage present. Verify good voltage in. The power LED is lit. The run LED is lit.

Try plugging in the WattsUpPro (a power, current and voltage measuring device for 120 VAC pluggable tools). It draws very little power, but it is a load and it will read voltage input. No display. How did I break this?

Turn off the power switch to the inverter. Turn it on again. Why? It was something that occurred to me, just because. The inverter powers up, the LEDs are both lit, and the WattsUpPro display lights up. 122.2VAC is what it shows. What gives? Dig out the manual - the inverter goes into a sleep mode if there is no load for 2 minutes. I guess I wasn't very fast getting the meter connected to check out the voltage.

Now that I know it needs a load, leave the WattsUpPro plugged in. The voltage at the output stage (the DC that is switched to generate the modified sine wave) is not easily measured with probes as large as the ones on a meter. I did get it done, eventually. The voltage is 135 VDC, which is lower than the 150 VDC I got from a smaller inverter that I had taken apart previously. I should check with a larger load on the inverter - perhaps it changes with load? Try a light - 1A at 120VAC is over 10A at 12VDC - overload and shut down. Same with the electric drill. I found a load small enough to plug and and not draw down the power supply. It's a 12V trickle charger, 12V at 1A output. The DC current rose from 2.5A to 2.6 or 2.7A. The DC Bus voltage drops to 134.5VDC. Unplug both the WattsUpPro and the charger. The voltage rises to 138 VDC. It's not regulated well, but I'm still at very low output power. I'll verify with a larger load when I get a battery connected. It's somewhere between 130 and 150 VDC.

The 'DC bus' is a bit hidden on this inverter. The transformers that step up the voltage go through a diode, and the output goes through the board. It may be to the other side or even to an internal layer, since this is a multi-layer board. It is not accessible from the top. To get access to the bottom of the board, I'd have to remove the clamps that tie the power transistors to the heat sink - all 16 of them. I'd rather not go there. The 'DC Bus' does come to the top of the board where it connects to the output transistors. That's the only location that I can see. It would not be easy to solder to that pad.

If I were to use this inverter as a DC/DC converter, the 'on/off' switch could easily be replaced by a relay. That would 'turn off' the load on the batteries. I should check what the current is when the switch is off. It drops to 0, then wanders a bit, from -0.1 to 0.1A - but I'm using a clamp-on meter since my regular meter does not measure currents above 100 mA. I expect that the inverter has no load when turned off, since the manual describes permanent mounting in a vehicle. If there was current draw when the switch was off, the vehicle battery would drain. That would not be a popular feature.

I'm not sure how I'd fake a 'load' for the 120 VAC outlet if I were to use this unit as a DC/DC converter.

Another thing that I have to verify when I get a battery connected. I guess that'll be it for tonight.

[Ken Fry]

Originally Posted by thingstodo

I want to use parts that I can source, and understand how to fix what I end up with, so I guess I'll be going with the 'more fun' way.

I can understand that completely.

Sounds like a great project. I'll try to follow along as it develops.
Ken

[thingstodo]

This was a test to determine if the 5 HP 460VAC 1745 rpm motor that I am using for testing will really generate 220% starting torque if it is driven with about 300% rated full load current. It was being driven by a 5 HP 240 VAC three phase VFD.

The notes that I took, the readings that I recorded, and the theory do not match. I see that I need to repeat this test. But here is the information that I have.

The experiment is done with starting torque, since I can measure that with a fairly simple setup. The motor is rated for 1745 rpm, or 55 rpm of slip. That's not quite 1 Hz (60 rpm) of slip. The graph of the torque on a Design B motor appears to show maximum torque around 2 times the slip frequency. My testing should have varied the VFD frequency from 0.9 Hz (approximately slip frequency) to 2.7 Hz (approx 3X slip) to determine where maximum torque is. Somewhere along the line I got mixed up and did not get that testing done ...

-------------------- Begin ----------------------------------------------
The following is a bit of a history lesson for three phase motors. Please tune out until the end of this section if you don't like reading this sort of thing. Why the history lesson? I'm trying to explain what I'm trying to test. I want to determine if I can get 220% of the rated torque from my test electric motor - a 1745 rpm, 5 HP, three phase 460V design B motor. And if I can, how much current it takes for me to achieve that torque. My estimate is 300% of the full load current.

There are 4 NEMA 'standard designs' for three phase squirrel cage induction electric motors, creatively labelled A, B, C and D. A few extra have been added for specialty applications and use the letters up to M

Industrial and commercial motors that are available surplus are mostly Design B. There are minimum standards for how much torque a Design B motor has, but not all manufacturers stick with the minimums. For the Design A,B,C,D if the equipment designer used the 'minimum' values, another vendor's motor will work in the application.

The sizes that could be used in a vehicle are normally 'T' frame motors. The 'T' frame motors have standardized dimensions (mounting holes, shaft sizes, distance from the mounting feet to the shaft, etc) that allow you to replace a T frame from one vendor with a T frame from another vendor and everything fits.

The specifications for a Design B motor state that it will generate at least 220% of it's rated torque at a certain frequency. That frequency is related to the speed that the motor is turning, and the motor's rated slip.

-------------------- End ------------------------------------------------

The motor is mounted on a 2 x 8, which is bolted to a few 2 x 4s. This motor does not drive a belt, but a strap instead (it is rated for much more tension than any of the belts that I have). The strap is wrapped around the drive pulley a few times and secured to the pulley. It then runs to a wooden pulley with a couple of different diameters, and runs around half the pulley circumference where it is fitted into a slot so that it does not slip. The pulley is mounted to a rotating shaft with a bearing on each end. The wooden pulley has a 2x2 bolted to it that pushes down on a 2 x 6, which spreads that force across a bathroom scale, which is levelled. That scale is how the torque of the motor is measured, with a multiplier for the wooden pulley and the motor pulley. It gives me decent accuracy and a range of measurements. The motor is rated at 15 foot-lbs, so I need to use the pulleys to amplify this weight value to something that can be easily measured on the scale. I hope to measure 220%, which is 33 foot-lbs. My scale can measure 240 lbs, so the multiplier should be just over 7, but it ended up being just under 4.

The motor and pulley bearings are bolted to the 2 x 4s. This whole mess is suspended between a work bench and a metal rack. The scale is on the floor below.

When the motor is turned on, the rack is being pushed down or compressed, the work bench is being lifted. The work bench had to be over 200 pounds to make this work.

There is a sketch below that tries to show what is going on. I'm not that good at explaining - the people that I explained this test to have not understood what I'm doing. When I reviewed what I did and compared that to my explanation, I must confess that I have trouble figuring out what I did as well. It appears that I will be doing this test again, perhaps in the spring.


Here is the sketch, for what it's worth.

[thingstodo]

Second part of the posting

There are a few more details, plus some assorted ramblings below.

I decided to split the spreadsheet into three tabs, for the three trials I did. The trials were different in what was measured and how it was measured. I thought that I was improving as I went along. Now I'm not so sure.

Settings:
The VFD is rated for 240 VAC three phase, 15A and will do 22A for about 30 seconds. It is set for linear V/Hz output.

Change - VFD boost settings. This is a percentage of the rated voltage, and the VFD allows the setting to vary from 0 - 40%. Changed from 0% up to 37%

Measure - motor temperature, motor current, scale reading

The VFD will display the output current, output voltage, or output frequency. This particular surplus unit appears to have an issue with the output current, since it displays the same value no matter what current it is actually sending out. The output frequency is always the same - 0.9 Hz. The output voltage displayed on the VFD never matched what I was measuring with the AC meter.

I have a clamp-on meter on one phase of the AC to the motor, another meter showing output voltage phase - phase, a hand-held infra-red meter for motor temperature (at the same spot each time - the meter is duct taped to the motor), a meter on the DC bus voltage of the VFD to make sure that I am not running out of power from my battery bank. I ran out of meters so I don't have the DC current from the battery bank, which would have been nice to show the power in versus power out for the VFD.

The results are not what I expected. That seems to be a recurring theme in the tests that I set up. The battery bank ran out of current the first time I ran the tests. So I fed 240 VAC into two of the terminals on the VFD, to supplement the battery power. I then spent the rest of the day ensuring that I did not over-charge the battery bank, or charge it too quickly. You will notice the DC voltage changes around 30 - 35% boost

I measured as high as 212% of the rated motor current, which gave me a reading of 100 pounds of force on the scale. Working that back through the pulleys, the motor put out 27.27 foot-lbs of torque. The 5 HP motor is rated at 15.04 foot-lbs, so the peak I measured was 181% torque. I was not able to put 300% current into the motor (my estimate) so I think it was reasonable to get less than the 220% torque listed in the motor specs.

These results are not what I remember them being. Some of the things that I thought were obvious about the setup - so I didn't write down the details - turn out to be important and not easy to interpret from what I did write down.

I don't know why I stopped testing when I got to the end of the 0.9 Hz set. I should have redone the whole test at 1.8 Hz and 2.7 Hz. If those showed any interesting trends, I should have added more test runs between them. I didn't even get up to 300% current. The last set of tests that I ran I didn't write down the motor current (!!!) That seems like a pretty important thing to miss.

I will be writing up what I want to see for the test before doing them, predicting the results, then performing the test and collecting the data. Working from memory while you are out in the garage does not seem to work very well.

You may very well wonder why I am using a test motor and a test VFD instead of using the equipment that I want to use in the truck. I didn't have a VFD with enough power to drive the 40 HP motor at 300% current when I did the tests, and I still don't have the VFD operational. It's just a whole lot safer to work on smaller (scale) models for testing than it is to perform full scale tests. I want to get some practise testing, predicting, logging. With 40 HP, 220% torque will be 120 * 2.2 = 264 foot-lbs of torque. That's a lot of torque to measure. My present setup will not handle it.

First three tabs are for trials 1,2, and 3. They are a bit of a mess.

VFD Settings is a listing of the parameters for the test VFD, an Allen Bradley 1333.

Motor Data lists the information from the nameplate for each of the motors that I have.

The test motor is:
reliance electric sabre 5 hp
184t frame
hp 5
460 VAC 3 phase design b type p
rpm 1745 amps 6.6
60 hz, ambient 40c s.f. 1.15
cont duty insul class f
enclosure prot code j
pf 83, nom eff 87.5
weight 69 lbs

[Ken Fry]

This is great stuff! I like your experimental approach.

[thingstodo]

I have two of the "Durafied 1350W continuous, 2700W peak Inverter S1350DI" inverters. I have not dug out any more of the surplus units.

I connected each of these inverters to a 12V battery (a small one) to check out the isolation between the DC output sections.

As seen in my crude but readable sketch (I'm trying out google docs), I checked out the voltage of each inverter separately. The voltage starts out wandering a bit, but settles down to a small range of 140 - 142.2 VDC. Both inverters have this voltage (internally). I was not able to reproduce the 135 VDC that I had measured earlier. Nor was I able to have the inverter 'go to sleep' and disable it's output.

Prior to connecting these inverters in series, I checked the voltage between the + and - terminals on the battery with the + and - terminals on the 'output'. The voltmeter showed a consistent readings. If points in a circuit are isolated, you can see many voltage differences but they are somewhat random and not repeatable. - battery and - output were solid at 0.0VDC. I'll reference everything from - battery. + battery is 12.6VDC as expected. + output is 140 - 142VDC. The output is NOT ISOLATED from the battery at ALL.

In my excitement, I had assumed that boost transformers meant that the input and output were isolated from each other. This was a rather dangerous assumption. It is good that I checked it out as early as I have.

My confidence is a bit shaken - and my resolve to check absolutely EVERYTHING is much stronger.

On the agenda for tomorrow - power the inverters from separate 12V batteries, verify that the inverters can be connected in series to get about 280 VDC, and connect this voltage up to a test VFD to ensure that it powers up as it should.

[MPaulHolmes]

This is really awesome!!! Totally out of my realm of expertise, so I'm learning a lot. Thank you!!!