...So I think I'll take apart the 3 motors that I have that are 5 HP as well as the one that is 40 HP and see if the coils ARE in series, and ARE easy to identify...
The 5 HP, 230/460V motors look promising so far. They already have 9 leads brought out to the motor junction box. Both ends of 12 coils would be 24 wires, so I would need to figure out a few things, cut some of the coil bindings apart and spend some time investigating.
These motors are 36 slot. There are two layers of coils, and all of the wires come out on the same side as the motor junction box. There is room to pull through more wires into the motor junction box, but there is already a potting compound to seal up the motor. Not hard to remove, just a bit of a pain.
I guess I'll have to see how interested I am when I've got the rest of the testing done.
...
So I can forget about the batteries again (for a while)
and the charging problems
and the DC to AC inverter soldering
and focus on
1 - take apart a motor and see if the coils are easy to separate
2 - build my test frame again to check the stall torque of the 5 HP motor (likely one that I don't take apart)
...
It's so easy to forget that things take time, and that any time you invest up front will reduce your time requirement later by a factor of 10 or even 100. AND IT WILL BE SAFER!
Take for example my assumption that I would walk out to my garage, wire up the Square D VFD (that I've never worked with) and get it running in a few minutes so that I could concentrate on the testing, the part I wanted to get done.
After wiring up the motor to the VFD, a variac and a transformer for a power supply (I was too chicken to plug the VFD into my welding outlet) and skimming the VFD manual - I powered things up to see what parameters I could see, record them for later research, and see if I could figure out where the minimum speed, maximum speed, full load amps, and motor rated voltage are - the things that I normally need to change when I take a new VFD out of the box at work.
The VFD had a fault, which I managed to clear by pressing the STOP button on the keypad. THE 5 HP MOTOR THEN STARTED!!! WITH NO START COMMAND!! IT RAMPED TO 90 HZ AND STAYED THERE.
The stop button on the keypad worked while pressed (the VFD ramped down) but as soon as I removed my finger, it sped back up. I tried to put the VFD into LOCAL from the keypad so that I could enter a new speed - that didn't work for me. I'm sure that I tried a few more things that I can't remember right now, but eventually (after about a minute) I had to shut the power off to the VFD to get it to stop.
If this had happened at work, I'd be spending the next several hours filling out 'Near miss' incident forms, talking to Occupational Health and Safety, the safety officer, the Mill superintendant, going for a drug test, etc. Now I'm just typing this up and kicking myself.
The good things:
- the motor was bolted down, so it did not roll off the bench and fall on me or other equipment or pull cables out of the junction box ...
- the motor did not have a belt on it
- the 3 phase wiring to the motor is good
- I put the motor back together again after checking for the multiple poles and windings, and it appears to turn just fine
- I only hurt my pride
The bad things:
- the manual that I skimmed turned out to be for a newer VFD than I have, so the control wiring that looked OK was actually an issue
- since it was the first time I'd powered up the VFD, I didn't realize that I did not have the correct passwords to get into the parameters and change anything
- the previous use for the VFD required auto-start, auto-restart, and a fixed speed of 90 Hz. The STOP button on the keypad was not disabled, but the two-wire control scheme does not latch the stop button
Things learned:
- power up a VFD with no motor connected. That one sounds a bit obvious, but I managed to get it wrong
- check that the manual matches the VFD that you have. Again, pretty obvious
After I calm down a bit, I'll go back to the garage and do this again, without skipping steps.
I spent about 2 hours searching for the manual. I found it online - 4 years older than the one that I got WITH the VFD.
I've gone through the new (well, the CORRECT) manual, written up what the defaults are and what I think the test will need. Now I'll go remove the motor from the VFD, then type in the parameters that I want to change and check all of the rest.
Wiring - terminals 1 - 2 - 7, terminals 2 - 22.
Power up VFD (with minimum voltage - it turns out to be 125 VAC single phase). The VFD electronics power up - the electronics take very little power. But the pre-charge resistor is not bypassed, so when the output to the motor is turned on, the bus voltage drops and the VFD trips on low voltage.
The VFD powers up without the motor connected. Parameters are entered. I fat-fingered one of them. Parameter review at the end caught it.
I hit the START button on the VFD keypad and the output turned on, then tripped on low input voltage.
Connect the motor (after turning off the VFD)
I turned up the input voltage to the VFD - 220VAC. The pre-charge resistor is bypassed now - I didn't see when the bypass pulled in.
Start up the motor and change speeds a few times. Looks good.
Check terminal 10D - 10 Hz = 0.979V, 50 Hz = 5.01V, 75 Hz = 7.48V, 100 Hz = 9.98V, 25 Hz = 2.47V. That's pretty linear and should work OK.
Check terminal 10E - the output should be measured in amps. Not really sure how to check that. The motor does not take a lot of amps when unloaded.
Measure Torque on 5 HP motor - close to rated torque
The 5 HP test motor has a pulley with a radius of 1.0625 inches, that drives a belt to a load pulley. The load pulley has two radius, the one driven by the belt is 2.9 inches and the attach point for the 1 x 2 is a 6 inch radius. The 1 x 2 pushes down on the scale and registers a force in lbs.
Check:
At rated torque, what is the scale force?
15 foot lbs rated torque.
convert to inches, 15 foot-lbs * 12 inches/1 foot = 180 inch-lbs
Drive the 1.0625 radius motor pulley, 180 / 1.0625 gives 169.4 lbs tension on the belt
The belt transfers that tension to the driven pulley at 2.9 inch radius
169.4 lbs * 2.9 inches gives 491.3 inch-lbs torque
Transfer that torque to the 6 inch radius
491.3 inch-lbs / 6 inches = 81.88 lbs on the scale
If I manage to get 300% torque out of the motor with the tests, the scale will register 3 * 81.88 = 245.6 lbs on a range of 280 lbs. Perhaps the bearings, pulley, and everything else will take 245 lbs of force, perhaps not.
Run a test with 3% boost.
1 lb or so on the scale at 2.0 Hz. 1.5A on one phase (black) for current out. 18.5 mv on 10E. What does that mean? No measurement on voltage. Check with clampon - 6.9 - 7.0 VAC phase to phase. That's not consistent.
Run at 0.9 Hz (like the last test that I did on the Allen Bradley VFD where I got good results somehow - it's a starting point) and adjust boost. 22 percent boost gives rated current, or a bouncing value that passes through rated current some of the time. 94 lbs registered on scale, which is a bit higher than the 81.88 calculate above. I do have a couple of wraps of the belt around the drive and driven pulleys. I should use a caliper to verify the radius. But 82 lbs calculated to 94 lbs measured is pretty close considering that the current values range from 5.9 - 8.0 amps. The clampon meter does an average over about a second, and that's not very many pulses at 0.9 Hz.
I varied the setpoint frequency from 0.9 - 1.5 Hz. The torque output fell. I was expecting to need about 1.8 Hz to reach rated slip (and maximum torque output). My measurements peak between 0.9 and 0.95 Hz. I'll have to check that again tomorrow.
After checking where I get maximum torque at rated current, I'll increase the boost and find out where the output torque reaches maximum. Then there will be some adjustments on the speed setpoint to see if more torque can be achieved.
I'll be posting a spreadsheet with the experimental data, a sketch of the setup, the VFD settings, etc tomorrow. If I run into anything strange (stranger than maximum measured torque at a lower slip than expected) hopefully someone can explain it.
Results - 200% starting torque from a 5 HP induction motor with an over-sized industr
See the attached excel file for the data, a few pictures of the equipment, and a sketch of the setup. I took pictures this time. I'll get the pictures attached as I get through the editing (cut out only the part that's interesting, save with higher compression so it fits into the limits)
I got 200% starting torque out of this setup, but the limits on the 'boost' setting of the VFD would not allow me to put more current in. That's a nice safety for the average application, but sort of annoying for my testing.
The motor was starting to heat toward the end of the tests. Once I got things set up right, I went back and re-did the tests from yesterday and recorded good data for Volt-Amps (VA) in, temperature and 3 phase amps.
At maximum torque, I was around 50VAC and 10 amps according to the display on the VFD (that I figured out this morning). I did check the amps with a clamp-on meter a couple of times to make sure the numbers were similar. The answer is: it could be the same or it could be a lot different. The clamp-on fluctuated a lot more at the higher currents.
The slip that I calculated did not match testing at all. 1.6 Hz * 30 rpm / 1 Hz should give me a slip of 48 rpm at the 200% torque level. 1745 rpm should have 55 rpm of slip at 100% load. 200% should be even more. No explanation for that so far.
I re-ran yesterday's tests with the VFD showing the current, monitoring the motor temperature, and trying each boost setting. The results no longer show a peak for the available torque. As the boost rises, the torque continues to increase, up to the maximum available boost.
The maximum torque occurs at 1.6 Hz speed reference on the VFD, 30 % (maximum) boost.
I would say that these results are hard to reproduce. When I re-ran the tests with the same hardware on a different day, I got similar results but several things changed - for example, a peak in the torque curve yesterday was measured as a rising curve that flattened out today.
There are a LOT (lots and lots) of picky little details that act as sources of error. They have to be checked and rechecked. I did not realize it at first so some of my data is likely close, but not quite accurate.
- the zero on the scale wandered a bit. I reset it three times ... or was it 4?
- the vertical 1 x 2 that transfers force to the scale moves slightly, out of vertical. The force on the scale then reads low. I corrected this with a level twice and by eye at least twice more
- the alignment of the motor and wooden pulley was off when I noticed that the axle bent, and for who knows how long before that. I aligned it only once after the axle was straightened
- the belt from the motor to the pulley slipped off two ... no, it was three times. As I put it back on, I likely had more or less wraps on each pulley so the diameter of the pulleys changed slightly. I could not locate my calipers, so I did not get accurate measurements. The exact diameter of the driving pulley and the driven pulley could have been out by 1/8 or 3/16 of an inch.
It got ridiculous - I started recording when I made corrections and gave up - it took almost as much space as the data.
Last edited by thingstodo; 01-25-2012 at 12:05 AM..
Reason: Add sketch
Success - I have a laptop communicating to the PLC5!!!
A change of topic for a while.
The industrial computer (PLC) that I'm using uses an older proprietary protocol to communicate to each other. The PLC is a 1785 5/60B. The communication interface is 1771-KE, 15 pin serial to Data Highway Plus
Data Highway Plus connections, 1 - 1, shield to shield, 2 - 2
Load RSLinx on laptop
configure drivers
AB_DF1-1 DH+ sta: 10 COM1: Running
Com1, 1770-KF2/1785-KE/Scanport
19200 station 10
no parity BCC Errorchecking
Stop bits 1 Full Duplex Protocol
Autoconfigure
Configuration of KE
Full Duplex, BCC, No Parity, Embedded Responses
Ignore Duplicate Messages
Ignore Handshaking Signals
Execute Diagnostic Commands
Address 010
DH+, 19K2, Local address
Show connections
Expand AB_DF1-1
Shows
AB_DF1-1, Data Highway Plus
10, Workstation
31, PLC 5/60B
So the serial cable from the laptop to the KE works. The KE works. The power supply works. The PLC5 shows up on the DH+ network. The KE works with the lower voltage signals (0V - 5V) instead of -12V to 12V signals that are specified by the RS232C spec (I can't bring myself to call it a 'standard')
The ebay rack, PLC, ke card, power supply - all work fine
The home-made serial cable, cat5 cable for Data Highway Plus, no terminating resistors - worked fine
Forgive me - I must celebrate the small victories!
Last edited by thingstodo; 02-01-2012 at 12:36 AM..
Reason: Fix spelling
The programming software that Allen Bradley used in the bad old DOS days was called 6200 series software. It was free. No copy protection because it was free. When the software moved to windows, that all changed and the various bits of software required purchasing, licensing, service contracts, and all that expensive stuff.
Since I'm not spending $5000 to purchase RSLogix5 software, I'm going to put in some effort and try to get the free DOS 6200 series software running on one of my extensive collection of old computers. Hopefully that computer will be a laptop. I'll use it for programming, I will need to move files to and from it for backups, but that's about it. The software runs in DOS mode on XP, so I can do most of the required typing that way, but XP does not appear to allow DOS tasks access to serial ports so I can't make it work. The serial protocol is timing critical, so windows is not a good choice in any case.
So it's time for me to take a trip down memory lane and try to remember how to set things up, figure out where I can find such old versions of drivers and such, and locate DOS drivers for hardware that is compatible with the hardware that I have.
Some results to report ...
I downloaded Freedos and installed to one of my older computers (Pentium 4, came with XP, no floppy, no serial port). It installed from CD fine, but no USB drivers so exchanging data is a pain. No PC Card support so I can't use my serial add-in card. The old drivers that I have from PC dos 6.22 don't work. Freedos appears to be a bust on this computer.
I still have a few things to try:
- Bart PE, which is used with Norton's older Ghost products, boots a command-line XP environment with USB drivers for hard disks and USB sticks, optional mounts for FAT16, FAT32, NTFS, etc. Perhaps I can configure that environment to work with the PLC programming software. There appears to be hope for this method. Since it uses XP install files ... it may still lock out the serial ports.
- Locate some old floppies, hope they work, and install Dos 6.2 or Dos 7, where my old drivers work, and get the PLC software to work. Then multi-boot to win98 - if I can find my old CDs and they still work - and use win98 drivers for wireless or network or usb sticks to exchange data
- If all else fails, I can use one of my old desktops (486, pentium 2 vintage) with a real serial port. Still need working DOS disks and win 98 or BARTPE to exchange data, but it would work.
Last edited by thingstodo; 02-04-2012 at 05:53 PM..
Reason: fixed spelling
DOS was not such a great thing. It took a lot of time
Some more news, a bit of progress
PCMCIA (or PC card) drivers, loading DOS high, Upper memory blocks, managing interrupts that are not *REALLY* documented any more, TCP drivers for old network cards, sharing drives without Novell ... The 'Good Old Days' were not really that good. Once you got things working, you breathed a sigh of relief. GETTING things working was a lot of time and effort. With perhaps a few superstitious things thrown in because they 'always seemed to work'.
I found a copy of DOS 7 from this site Download MS DOS Boot Disk 7.10 by Wengier | VETUSWARE.COM - the biggest free abandonware collection in the universe
. It boots to a prompt and lets me run the 6200 series software (PLC Programming software) from a USB stick. It works well with the serial port on my older laptop. It won't map the hard drive or CD, but that's OK. At least for now. Now that I have something working reliably (if you can call booting from a 1.4 MB floppy reliable), I want to get on with some programming.
The 6200 series software runs in a DOS window on XP (the offline version only) so I can type the program in, create the documentation, etc in a comfortable chair, using a modern computer. I then transfer the files to a USB stick and use the older computer, downstairs in the corner where I have the PLC mounted to a piece of plywood, to load the program in and watch the program run, perform any required corrections and debugging, then return the USB stick to the modern computer upstairs and continue. This would be the edit/compile/run cycle that you would normally use with a modern programming language like C++ using the integrated development environment or IDE.
The Allen Bradley 1785-L60B or PLC 5/60, has several cards installed for my purposes, each with a different type of electrical interface. It is set up like this:
Here's some information on the various cards, the voltages and currents
DC Input - 10 - 30 VDC. 1 ms response time rising or falling. 4.5 mA at 10 VDC, 12.5 mA at 24 VDC, 30 mA at 30 VDC to properly turn on the input. Sinks a minimum of 2.0 mA when off. 0 -> 5 VDC is off, 5 - 10 VDC is last state, 10 - 30 VDC is on. The temperature spec is 0C - 60C, but I know of several installations where it runs well down to -25C.
DC Output - 10 - 60 VDC. 0.1 ms delay rise time, 0.2 ms delay falling. 2.5 mA to fully turn on. Leakage less than 0.5 mA when off. 25A max output pulse for 10 ms. 2A continuous. Rated at 1.5 VDC voltage drop across output at rated current.
Encoder Input - 12 - 24 VDC. Single channel quadrature encoder counts to 50Khz. Single 12-bit count sent to the PLC per channel, with underflow and overflow.
Analog Input - 0 - 5V, 0 - 10V, -10V - 10V, 4 - 20 mA analog input module, 12 bit resolution, 4 channel, common negative signal.
Analog Output - 4 - 20 mA analog output module, 12 bit resolution, 8 channel, common negative signal.
How do you access the variables that represent switches ON and OFF?
The digital inputs and outputs are memory mapped. You read the correct address and you see the state of the input. Write to the correct address and the output changes at the end of the PLC scan.
What about the voltages, currents, temperatures, etc?
The analog inputs and outputs as well as the encoder information is transferred in blocks that take a bit of programming. The blocks give you a signal, called a DONE bit, when the information has been updated. At that point, you copy the data to another memory address and you use the second memory address for all of your calculations.
The 'landing zone' for the analogs and encoders is N9. This is a 'file', which is an array of variables of the same type. Each is accessed by using the file name, N9, along with it's index, like N9:0. N9 is set up with a length of 400 for now. The analog input uses N9:0 through N9:20, for example. The N in the filename defines the type of the variables in the array, 16 bit integers. These are signed integers.
The second memory address for analog inputs is N11, 1 looks like I or input, so I use N11 for analog inputs as well as the encoder and N10, where the 0 looks like O for outputs, for analog outputs. These files are much smaller. N11 is a length of 40, N10 is a length of 16
Why not use a Curtis controller? Why go through all of this effort?
One of my big goals in this is to UNDERSTAND what it is that is going on, to be able to fix it, and to upgrade pretty much anything if I want to. The motor controllers that are sold, even the DIY versions, have some parts that are pretty hard for me to understand. They also have limitations on what types of inputs can be used, and what types of outputs can be driven. Right now, I'm all about flexibility. Perhaps I'll discover that the available controllers cover everything that I need ... perhaps not. But I will discover ...
This PLC5 can take many types of input and generate many types of output.
The PLC5's largest drawback, in my opinion, is communications, or rather ... it's LACK of communicaitons. It does NOT support 1-wire, or SPI, or pretty much any communication format but it's own DF1 protocol ... and plain ASCII output (like to a serial printer). There are a few third-party cards to help with that, but they always felt like stop-gap solutions to me.
So I'm going to try my own 'stop-gap solution'. An arduino that will log data, transfer it to a PC or a dedicated display.
Today
Other popular topics in this forum...
PLC and communications
