Three-Phase Current Chopping, Eddy Current Brake

Discussion in 'The Projects Forum' started by el bob, Jan 8, 2009.

  1. el bob

    el bob Thread Starter Active Member

    Joined:
    Jan 8, 2009
    Messages:
    50
    Hi all,

    I am working to add computer control to an aging eddy current engine dynamometer. You can see pictures of the engine dynamometer and original instrumentation here if you're interested.

    I scanned a copy of the original schematic from 1979. It spanned 5 A4 European size pages. I grafted it into one jpeg and you can see it here.

    I am ultimately trying to figure out what the RC, SC and TC signals going into the 6'024'0 controller board (see schematic/photos) look like and how to emulate them using computer output. At this moment, I am trying to figure out what one of the pieces on the 6'024'0 circuit board is. It looks like a TO-92 transistor. I tried searching the printed text on Google but could not find anything. Does anyone recognize it?

    TEC
    103A
    F9M

    The top line in the "T" in TEC looks more like a squiggly, wavy sketch artists use for far-away birds. That's a terrible description, but the point is it's not a regular T. :)

    [​IMG]

    As you can see on the schematic, the details of the 6'024'0 circuit are not shown. I removed the board from the controller to document it. I put together a schematic that is accurate, setting aside the unknown transistor. I tried to draw the transistor into the schematic, orienting it so the leads coming out of it go in the same direction as they do in this photo.

    Updated 6'024'0 Controller Schematic v2.0
    http://kai.caltech.edu/MotiveDynoController/09_01_08-60240_Schematic-v2.0.pdf

    6'024'0 Controller Schematic v1.0
    http://kai.caltech.edu/MotiveDynoController/09_01_07-60240_Schematic.pdf


    Dynamometer Photos
    http://kai.caltech.edu/MotiveEngineDyno/

    Eddy Current Controller Schematic
    http://kai.caltech.edu/MotiveDynoController/09_01_04-1348-FullStitch-Cropped.jpg

    Eddy Current Controller Photos
    http://kai.caltech.edu/MotiveDynoController/60240/

    6'024'0 Controller Removal Photos
    http://kai.caltech.edu/MotiveDynoController/60240_Removal/

    Unknown Transistor Photo
    http://kai.caltech.edu/MotiveDynoController/UnknownTransistor.jpg
    Last edited: Jan 8, 2009
  2. beenthere

    beenthere AAC Fanatic!

    Joined:
    Apr 20, 2004
    Messages:
    15,815
    Location:
    Missouri, USA (GMT -6)
    I believe the logo is for Tecor. The part appears to be an SCR - http://www.datasheetarchive.com/EC103A-datasheet.html.

    How good is that schematic? Can you tell what the leads go to? The callouts appear to be for thyristors (g = gate, k = cathode). I would think the brake would need devices in larger than TO-92 to switch current.

    Love the shopping cart full of rocker arm assemblies and cams.
  3. el bob

    el bob Thread Starter Active Member

    Joined:
    Jan 8, 2009
    Messages:
    50
    Beenthere, thank you very much! That is a great help.

    There are three larger SCR's bolted to the 6"x2"x1.5" aluminum heatsinks. They are more visible in these photos.

    http://kai.caltech.edu/MotiveDynoController/60240/images/image/IMG_2285.JPG
    http://kai.caltech.edu/MotiveDynoController/60240_Removal/images/image/IMG_2461.JPG

    Shopping carts are surprisingly strong and always useful. I know a guy who uses them by the half-dozen as an instrumental part of an exhaust system production shop. I'm with you though, nothing better than one with a camshaft sticking out of it.
  4. el bob

    el bob Thread Starter Active Member

    Joined:
    Jan 8, 2009
    Messages:
    50
  5. el bob

    el bob Thread Starter Active Member

    Joined:
    Jan 8, 2009
    Messages:
    50
    Are you talking about where the leads on the pdf schematic go to? They match up to the leads going into the 6'024'0 empty white box on the complete controller schematic jpeg file. (Warning, 3MB)

    Eddy Current Controller Schematic
    http://kai.caltech.edu/MotiveDynoController/09_01_04-1348-FullStitch-Cropped.jpg

    Starting from the top right of the large jpeg schematic, the incomplete 6'024'0 drawing begins about 20% in from the left, and about 25% down from the top. It shows leads for R3, S3, T3, RC, G, K, SC, G, K, TC, G, K, 20, and 0.

    I hope I understood the question correctly.
    Last edited: Jan 9, 2009
  6. beenthere

    beenthere AAC Fanatic!

    Joined:
    Apr 20, 2004
    Messages:
    15,815
    Location:
    Missouri, USA (GMT -6)
    Got the schematic, but no time to look it over for a bit. Do you have an oscilloscope? It could be a handy tool to look at those line and relate their activity to the action of the braking.
  7. el bob

    el bob Thread Starter Active Member

    Joined:
    Jan 8, 2009
    Messages:
    50
    With the 6'024'0 circuit schematic complete, the issue I'm facing now is understanding how the 6'024'0 circuit works with the three larger SCR's (seen on the complete controller jpeg schematic) to control current sent to the Brake Coil.

    I read the allaboutcircuits.com article on SCR's. It seems that the current is being controlled by "chopping" the AC signal so that only parts of the positive phases are allowed to continue towards the brake coil. What I don't get is what kind of signal input on the RC, SC, and TC terminals, the inputs to the 6'024'0 circuit, results in variable control of the larger SCRs inline with the 380VAC lines going to the brake coil.

    Eddy Current Controller Schematic (jpeg)
    http://kai.caltech.edu/MotiveDynoController/09_01_04-1348-FullStitch-Cropped.jpg


    Updated 6'024'0 Controller Schematic v2.0 (pdf)
    http://kai.caltech.edu/MotiveDynoController/09_01_08-60240_Schematic-v2.0.pdf


    I'm also a little confused by the grounds on both circuits. I'm not familiar with positive-ground systems, and I'm not sure this is one, but I'm planning to read a little bit into that. Any insight there would also be greatly appreciated. Happy Friday all.
  8. beenthere

    beenthere AAC Fanatic!

    Joined:
    Apr 20, 2004
    Messages:
    15,815
    Location:
    Missouri, USA (GMT -6)
    Can you relate the G, K and C (that would be RC, etc) to the fastons located on the PCB? They are probably the lines the dyno uses to trigger the small SCR's that in turn fire the big ones - but it would be good to check.

    Too bad that controller is a black box. No Italiano, anyhow.
  9. el bob

    el bob Thread Starter Active Member

    Joined:
    Jan 8, 2009
    Messages:
    50
    Ah, I missed your post this morning somehow. Yes, I was able to find the G, K and C leads shown on the schematic inside the real-life controller. Finding these wires helped me identify which circuit board was the 6'024'0 board. I then removed this circuit board and drew up a schematic for it, linked to here as a pdf and jpeg file.

    The three G, K and C sets of leads (G, K, RC, G, K, SC, G, K, TC) going into the "black box" are included in the 6'024'0 pdf and jpeg documents. The complete details of the black box are shown in the 6'024'0 pdf schematic. I converted the 6'024'0 schematic pdf file into a jpeg file as well, in case it is easier to view.

    Updated 6'024'0 Controller Schematic v2.0 (jpeg)
    http://kai.caltech.edu/MotiveDynoController/09_01_08-60240_Schematic-v2.0.jpg

    I do have an older Tektronix 2 Channel oscilloscope, but I have not been able to hook it up yet. As you can see in the dyno room pictures, the equipment is tucked away very tightly when all the dyno pieces are installed and ready to run. I am thinking about soldering in three test leads into the RC, SC and TC leads to run outside of the controller's box. These wires would let me tap into the signals easily even while the box is bolted up and installed in the equipment rack. I don't really know what to expect over the RC, SC and TC leads. I'm hoping the oscilloscope won't mind whatever it is. I have two 10x attenuation probes.

    I currently have the complete eddy current controller apart on a bench. Most aspects have been covered in photos. I'll have to reassemble the controller and reinstall it in the rack before I can run the dyno and scope the 6'024'0 input leads.
  10. beenthere

    beenthere AAC Fanatic!

    Joined:
    Apr 20, 2004
    Messages:
    15,815
    Location:
    Missouri, USA (GMT -6)
    What is the operating proceedure for the dyno? How are measurements taken? I can see a load cell in one photo.

    You might want to measure those leads to ground before you attach the o'scope leads. You don't want to let the smoke out of the case.
  11. el bob

    el bob Thread Starter Active Member

    Joined:
    Jan 8, 2009
    Messages:
    50
    I'm glad you mentioned grounding the leads. The system appears to have a positive ground. Do I reference RC, SC and TC to the positive ground? I'll make sure to check them out with a DMM first. I'd like to keep the smoke inside the scope's case for at least a little while longer.

    Regarding operating procedures for the dyno, I'm not sure how much detail you are looking for. Just let me know which areas could use more description.

    Dyno Data Acquisition Overview

    The load cell you see is the sensor used to ultimately measure torque. There is an Omega OM5 full-bridge strain gage signal conditioner and an OM5 two-channel backplane. The strain gage signal conditioner module supplies 10VDC excitation voltage to the load cell, processes the load cell's output signal and provides a 0-5VDC scaled analog output representing 0-2500lbs on the load cell. This 0-5VDC analout signal is wired into an RS232 DAQ (Innovate Motorsports SSI-4) which applies a calibration formula to convert the 0-5VDC signal into a real Torque value to be analyzed by the dyno operator.

    The only other sensor at this point in time is a variable reluctor positioned very close to the outer radius of a 60-0 toothed trigger wheel. This sensor produces a 0-80VAC sinusoid whose frequency variers with the rotational speed of the trigger wheel, which is mechanically fixed to the RPM of the test engine and dyno assembly. The two wires from the VR sensors are connected to special leads on the DAQ that feed into an onboard VR signal amplifier circuit. The DAQ reads the frequency of the signal. Due to the fact that this specific trigger wheel has 60 teeth, the magnitude of the frequency equals the magnitude of the RPM.

    VR_Signal_Frequency(Hz) x 1_Revolution/60_Trigger_Wheel_Teeth x 60_Second/Minute = RPM

    1_Revolution/60_Trigger_Wheel_Teeth x 60_Second/Minute = 1 ----> VR_Signal_Frequency(Hz) = RPM

    A computer runs software that interfaces with the DAQ and records the Torque and RPM measurements. The software calculates Horsepower on the fly according to HP=Torque*RPM/5252.

    Dyno Load Application Overview

    The test engine is first started up with the dyno "off," meaning no current is passing through the Brake Coil and no load is on the test engine. The existing manual eddy current brake controller has its own VR sensor that gives it an RPM feed. The controller is switched on and then a knob is turned to apply more or less load. There is also a separate knob to switch load application modes. Some modes multiply the selected amount of load (selected with the first knob) against the sensed RPM. I can scan the operator's manual for the dyno brake controller, it is only two pages.

    The RPM and Torque output of the test engine can then be controlled by varying the mechanical Throttle input and the load knob on the brake controller. The problem and motivation for the project is to allow for computer control of load application so that consistent, back-to-back tests can be run with the same load application. It will also allow for control of when and how much load to apply in these tests.
  12. beenthere

    beenthere AAC Fanatic!

    Joined:
    Apr 20, 2004
    Messages:
    15,815
    Location:
    Missouri, USA (GMT -6)
    I can see where that would be desirable. The computer may need to have some of the other inputs led in before it can really control the system.

    That's a bit later - I'm looking over the wiring diagram in between all else to try to make sense of some different electrical conventions.

    "All else" meaning all the other stuff going on.
  13. el bob

    el bob Thread Starter Active Member

    Joined:
    Jan 8, 2009
    Messages:
    50
    I tapped in some "T" connections in the RC, RG, SC, SG, TC, TG and 0 (positive ground) lines and brought the oscilloscope in for testing. Pictures of the wires being added and the test being setup in the dyno room are linked here.

    There were two main tests. The first involved scoping between RC and 0. The second involved scoping between RG and 0. The idea here was to try and gather some clues as to what sort of input on RC corresponds to what kind of output on R's G terminal that ultimately actives the inline 380V AC SCR.

    Pictures, description and results for the RC & 0 oscilloscope testing.

    Pictures, description and results for the Rg & 0 oscilloscope testing.

    Does this give anyone any ideas as to what kind of signal could be used on the RC, SC and TC terminals to gain variable control of current going to the brake coil?
  14. beenthere

    beenthere AAC Fanatic!

    Joined:
    Apr 20, 2004
    Messages:
    15,815
    Location:
    Missouri, USA (GMT -6)
    The magic is all in the black box section - the board labeled 0'046'0. The 34 volt transformer feed all three phases into ii on pins Rb, Sb & Tb. Something inside the board decides how much current is wanted in the brake, and pulses the small SCR's on - which in turn fire the big SCR's. The few controls shown do not seen to be related to the brake function in any way.

    Your photos of the pulses on the Rc, etc lines will be more meaningful is the second trace shows the b input as a timing reference. The earlier in the positive cycle the pulse occurs, the more braking - or so it appears.

    A circuit called a zero crossing detector can indicate when an AC waveform is at 0, and simple timing can calculate how long to delay after the crossing before pulsing the SCR on.

    What might be interesting is the thought of using the computer to run the whole test. That will require many more inputs and a control program. Using the computer as a stand in for the control board is kind of overkill, but using it to run a testing loop is pretty sensible.
  15. el bob

    el bob Thread Starter Active Member

    Joined:
    Jan 8, 2009
    Messages:
    50
    Thank you again for your insight. I did not think about the 34V AC lines going into the 0'046'0 circuit before, but it seems obvious that they are used as a reference for control of the Rc, Sc and Tc outputs now that you mention it.

    I have been looking into possible control solutions today. From my reading thus far, it seems like a combination of National Instruments LabView software and an NI DAQ makes for a popular platform.

    My current idea is to use the 440V, 380V and 34V AC power supply portions of the existing dyno controller and add a new computer-based controller that outputs to the little SCR's.

    Are there any off-the-shelf pieces that can take in a scaled analog input (say, 0-5 or 0-10VDC) and separately fire the three little SCR's with phase delays proportional to the input voltage? It would have to monitor the incoming Rb, Sb and Tb waveforms to know when to trim and fire the three phases.

    If such a piece exists, or if something close enough exists, then I believe feeding RPM, Torque, and possibly coil input current data into LabView will allow for a decent eddy current dynamometer controller. LabView can run a PID controller that would tie into one of the DAQ's analog outputs, ultimately controlling the piece that fires the little SCR's.
  16. beenthere

    beenthere AAC Fanatic!

    Joined:
    Apr 20, 2004
    Messages:
    15,815
    Location:
    Missouri, USA (GMT -6)
    I found a link to an article by Teccor that deals with phase control SCR switching - www.eetasia.com/ARTICLES/2001JUN/2001JUN14_AMD_AN5.PDF.

    Since it appears that the small SCR's trigger the big ones, it seems pretty likely that triggering the small ones from a logic gate is going to work. I will hunt around for indications of OTC controllers - chances are good that they exist.
  17. beenthere

    beenthere AAC Fanatic!

    Joined:
    Apr 20, 2004
    Messages:
    15,815
    Location:
    Missouri, USA (GMT -6)
    Doing some digging online, I came across these people - http://www.appliedps.com/newweb/Datasheets/AP-1950/AP-1950.htm.

    The controller would replace all but the big SCR's in your dyno, and might make control from a computer or microprocessor pretty simple. The Google search term was "three phase scr control circuit", if you want to look further. Something like this would make automation of the dynomometer considerable easier.
  18. el bob

    el bob Thread Starter Active Member

    Joined:
    Jan 8, 2009
    Messages:
    50
    Thank you beenthere. Google digging can be quite an exercise.

    I searched around a bit after reading your earlier post this morning. It took me a while to figure out the most appropriate Google search terms, but I eventually stumbled into Eurotherm, Practical Global Solutions, ISE, Wayne Engineering and finally Applied Power Systems.

    I spoke to Les @ Applied Power Systems. The BAP1950 sells for $450 configured as a half controller in singular quantity. That will work.

    I had actually jotted down a few questions I was going to ask, one of which was whether the AP-1950 should be used to trigger the small SCR's, or bypass the small SCR's and trigger the large SCR's directly. You can imagine my surprise to come back to this thread and see, not only had you already nailed the product choice, but also figured out the most logical implementation as well. Any chance you have a pick for the SuperBowl? :)

    I ordered a book on LabView and printed out the NI LabView PID Toolkit manual. I need to read a bit and then it should be possible to lay down some kind of plan for the combined LabView, DAQ and SCR Firing Board system to control the dyno. I printed the Teccor article on Phase Firing you posted as well.
  19. beenthere

    beenthere AAC Fanatic!

    Joined:
    Apr 20, 2004
    Messages:
    15,815
    Location:
    Missouri, USA (GMT -6)
    It will be nice it it works for you. I didn't think you wanted to try to build something from scratch, and that controller looked to be pretty flexible as to actual control scheme.

    You might want to keep the old control panel for bragging rights, though - "used to check out Ferrari engine rebuilds with it".
  20. el bob

    el bob Thread Starter Active Member

    Joined:
    Jan 8, 2009
    Messages:
    50
    It's funny you mention that. There are a couple of Ferrari engines that may see use on the dyno if a decision is made to install gearing that allows gas motors to rev higher than the dyno's 3600RPM limit. Most of the motors going on the dyno are 6.5L+ diesels, so the RPM limit is usually not an issue.

    Now that you mention it, I was intending to install a switch that could allow the dyno to be controlled by either the Computer or by the existing Sicon Controller. This way the dyno could still be operated with the manual controls, as it currently is today, in the event there is ever a problem with the new computer. I havn't planned out the actual mechanism or begun to consider any possible ill electrical effects of using such a switch, but it's something that, I believe, would provide considerable value. The most basic idea would be to use a two position, six terminal mechanical switch, but that may not be a great idea. Any pitfalls to avoid that you can see there?
Similar Threads
Forum Title Date
The Projects Forum Simulink of Control Circuit for Soft Start Soft Stop Three-Phase Induction Motor. Feb 26, 2008
The Projects Forum H bridge for High Voltage, Low Current Sep 23, 2014
The Projects Forum Hi-power led led circuit, current limiter needed? Sep 18, 2014
The Projects Forum Balencing the charging current between Solar Panel and electric battery charger. Sep 3, 2014
The Projects Forum current amp help Sep 2, 2014

Share This Page