Not without knowing more about the black box part. You "should" be able to simply switch the three leads to the main SCR's and run under the new controller. Even if the old box is trying to send signals, it shouldn't bother the small SCR's to have no load.

You might want to use a three pole double throw toggle with a guard on it to make sure it doesn't get bumped while testing. An accidental change in load could lead to engine disassembly.

 

I am attempting to spec out suitable DAQ hardware to both sample Torque and RPM data, as well as output a 0-5V DC analog control signal for the SCR firing hardware. I'm looking at DAQ's that include analog out channels. One of the DAQ specs I'm focusing on is the maximum analog-out samples per second output rate.

National Instruments (NI) USB-6211 has two 16bit, 250kS/s analog out channels. If the SCR firing board has to make a decision on how to phase control three-phase AC lines operating at 60Hz, then it has to make at least 180 decisions in one second. Following this idea, is it possible for the SCR firing board to even make use of input control data at a frequency greater than 180Hz?

I'm trying to determine if there are best practices to follow, or more importantly common mistakes to avoid, when deciding the speed at which the complete LabVIEW dyno brake PID controller will output a control signal. Is there a parallel to the Nyquist rate on the signal input and sampling side? I am searching on Google, but appear to have not hit the right keywords yet.

 

From past experience, 16 bit data is hard to get outside a lab. That's 1 part in 32,768, which is very precise, but also prone to noise.

We have had some issues with the DI-720 acquisition device (we being myself and a researcher in California), but the DI-730 - http://www.dataq.com/products/hardware/di730.htm - is worth a look. Just put shorting bars in unused inputs.

14 bits gets 1 part in 8096 resolution, plus the 730 has more robust input protection. Look at the literature and talk to the people.

One way around the Nyquist limit is to sample way above it. Your RPM sensing will be a multiple of the actual, for instance. You might want to sample at 10X the max rate to see if the rate wanders a bit over time. It becomes mind-boggling to realize the things you can do with a good digital system.

 

That 730 unit is very impressive. "Measurement range of ±10mV to ±1000VDC (or peak AC) over six ranges" The idea of being able to monitor the brake coil's input current and voltage directly without additional hardware is enticing.

On the other hand, it is $3,000, and somewhat less importantly, does not provide an analog output. I believe that price is very fair for what it is. Eight 14bit channels that can be sampled simultaneously at 150kHz make for a great feature set. I am hoping that I do not need that much capability. Looking through the DataQ data acquisition product line, I do not see units with analog output channels. Were you thinking about using the 730 for input and a separate module for the control signal output?

Initially, the only input channels will be the 0-5V DC signal from the strain gage signal conditioner (feeding from the load cell) and a -50 to +50 V AC waveform from the variable reluctor crank trigger sensor. The frequency of the AC signal will vary from 0Hz to 3600Hz, directly (1:1) related to the RPM of the dyno and engine under test. Following the Nyquist limit, the minimum sampling rate necessary for the RPM data is 7200Hz. It seems that most data acquisition boxes outside of the <$200 range offer sampling rates orders of magnitude larger than this.

As of right now, the plan is to run the PID loop solely off of RPM data, namely the difference between target RPM and instantaneous RPM. It would be nice to incorporate brake coil voltage and current checks into the controller, but I don't see that in the cards right now. There may be an essential reason to act on it, perhaps something safety-related, but I can not think of one right now.

 

It is kinda interesting to find it all in a box, plus I lost sight of the analog control function.

See what you think of the stuff here - http://www.circuitspecialists.com/level.itml/icOid/107.

I think USB connected controllers are great, as they are computer-independent, but this stuff looks good for the analog control.

I'll have some time to look further, so check back...

 

A few sites to consider:

http://www.measurementcomputing.com...+I/O&mscssid=WV6KUJ3NGHRC8P6RQSWV1ALEAWQN8HP0

http://www.lanpoint.com/usb-data-acquisition.html

http://www.bb-elec.com/product_family.asp?FamilyId=286&Trail=4&TrailType=Top (look at the USB -4711)

Like all else on the internet, the choices look infinite. I was using the search term "analog interface module with usb i/o". You will need 12 - 14 bits resolution in the D to A, and at least 10 bits in the D to A output. Have extra channels in and out available - you may find it will be interesting to monitor more stuff for better dyno results.

 

I think you are on to something regarding USB and portability. I do most of my developing and testing on a laptop a few miles from the dynamometer itself. The dedicated dynamometer PC is a desktop and has open PCI slots. I could use PCI DAQ hardware, but I won't be able to do very much bench testing away from the test cell. A USB DAQ would allow for basic LabVIEW testing with a laptop. This could minimize the amount of time the dynamometer will be unavailable for use while the new brake controller is installed, configured and tested.

Thanks for posting the Circuit Specialists link. I read through the NuDAQ product list. It's all PCI hardware, which again could work, but may not be the best fit. I noticed that the DAQ units are advertised as "Continuous, gap-free, high speed under Win-NT/95." I see there's a driver that supports Windows98/NT/2000/XP. I clicked on the "Very Important Information for LabVIEW™ Users" link. There's a screenshot of the product being used in LabVIEW, but it appears to be a much earlier version. It probably works with the latest version of LabVIEW as well, but it may not be the selling point of the line. The $399 eight channel analog output PCI card looks like a great value for the price, but there isn't mention of driver support for Windows XP.

I am reading through the three new links posted above. I have used Measurement Computing USB DAQ hardware in the past with LabVIEW. It worked well. The USB-1408FS is tantalizingly affordable at $249, with four 14bit differential analog inputs and two 12bit analog outputs. Adding additional analog input channels, and keeping analog output capability, knocks the price up $1150 to $1,399 for the USB-1608HS-2AO. It's interesting that a 100% increase in analog inputs channels results in a 561% increase in price.

The B&B Electronics website does not mention LabVIEW compatibility. It's likely that it's compatible, but I don't see it on the product description page. It looks like I need to set up a web account to see pricing on the Intelligent Instrumentation products.

Among the NI USB DAQ contenders, I'm weighing the benefit of the "Industrial" USB-6221 versus the portable USB-6211. The former uses a standard 120VAC power source. The 6211 gets its power over the USB connection. Needing a 120VAC power source is not a probem for this installation, but the 6221 costs $200 more at $1049 versus $849 for the 6211. They both offer the same analog inputs (8x 16bit differential @ 250kS/s). There is a difference in analog output current drive capacity. The 120VAC unit can put out 5mA while the portable unit puts out 4mA. I can't understand why the unit with wall power can't drive more current. I don't know for sure, but I think the analog control signal will not require more than either of these amounts of current. Does that sound about right?

There must be a Measurement Computing unit that's priced between $1399 and $249 that has analog outputs. If there really isn't, I'll probably go with either the NI 6211 ("Portable", USB Data + USB Power) or 6221 ("Industrial", USB Data + 120VAC Power).

 

Something I didn't mention is that it looks like the Variable Reluctur (VR) RPM signal will need to be passed through a signal conditioner before heading to the USB DAQ's being mentioned. The NI DAQ's have maximum voltage ratings of -5..+5 and -10..+10 Volts. The unmodified VR signal can rise above 50V.

There's a circuit made for MegaSquirt projects that converts two VR signals into square wave signals. I believe they are 0V/5V signals. It's made by Jean Bélanger and is available for sale from his web site. "Dual VR Conditioner Board v1.1" There is no schematic posted on JB's site, but there's a link to the MegaSquirt web page describing the circuit it's based on.

Does that look like the output signal voltage will stay below 5V?

 

I previously thought the $200 difference in price between the NI USB-6211 and USB-6221 was the result of the "Industrial" build of the 6221. After reading the spec sheets again, I noticed that the 6221 offers twenty-four digital input/output channels compared to the 6211's four. The 6221 also has 24mA maximum current drive for each of these digital channels compared to the 6211's 16mA.

It seems that most of the $200 differences goes towards increased digital I/O capabilities that are not necessary for this application. I foresee needing more analog input channels in the future, but not necessarily digital channels. Given this, between the two, I would opt to save the two hundred dollars.

I double-checked the Measurement Computing product list. The options listed two posts above are the only applicable choices. Both have two 16bit analog output channels. There is the option of four non-simultaneous analog inputs for $249, or the option of eight simultaneous analog inputs for $1399. Too bad there isn't a unit with specs similar to the NI-6211 for $400-$700. Since there isn't, I'll probably go with the NI-6211.

 

With off-the-shelf, life is compromise. You wonder why some hardware costs like NASA-grade, and some is a fraction of that cost, but always with one significant feature absent.

That variable reluctor signal would go into a digital input better than being converted by an A to D. If it's going 0 - 5 volts, it's already digitized to the extent you require, which is to signal that another tooth has come opposite the sensor.

Save the A to D inputs for the torque-measuring bridge and things like head and exhaust temperatures, manifold pressure, etc.

 

The equipment list seems to be settling.

• Applied Power Systems BAP-1950 (SCR Firing Module)
• National Instruments USB-6211 (USB DAQ with Analog & Digital I/O)
• JBPerf VR Conditioner (AC VR Signal -> Digital TTL)

I've got the Sicon controller box (reference picture) apart to plan the connections necessary to add in computer control. I'm looking at the BAP-1950 Datasheet and figuring out which wires need to be hooked up inside the Sicon box. The BAP-1950 needs to connect to the gate and cathode of each SCR in a "DC Power Supply - Half Control Converter" configuration. The big SCR's in the Sicon box have four connections.

• ThreePhase Line In (Anode) ||Not Visible in Pictures
• ThreePhase Line Out (Cathode) ||WireColor=Black
• Gate ||WireColor=White
• Unknown (Connects to 0/Ground in 6'024'0 circuit) ||WireColor=Red

A picture of the wires coming off the big SCR's.

The red wire with unknown purpose connects to the K terminal for each of the three little SCR's on the 6'024'0 Circuit (jpeg).

The Sicon box will need to be switched on in order for the computer controller to work. The K terminal connects to the Sicon ground, which I believe is a positive ground, and so will be "live" whenever the computer equipment controls the SCR's. I don't understand what the red wire does for the big SCR's though. I found a datasheet for the big SCR's, but it doesn't describe the wire inputs.

Westcode Semiconductors
"Convertor Grade Stud-Base Thyristor Type N023R"
PDF Datasheet - N023RH12.pdf

If it's a reference voltage for the gate signal, then won't I need to use a separate reference voltage from the BAP-1950 board when using computer control? Any ideas what purpose the red wire serves?

 

I took another look at the original Sicon Eddy Current Dynamometer Controller schematic. Each K terminal on the 6'024'0 circuit connects to one of the three big SCR's cathodes, via its respective red wire. This should provide a clean method of "tapping into" the big SCR's and installing a switch between the manual and computer controls.

All three cathodes of the big SCR's connect together and to 0/Ground, before passing through a 0.03Ohm resistor and finally through the brake coil. I would think this means that the three cathode reference connects required by the BAP-1950 would see the same signal.

The example diagrams I have seen so far show how SCR's connect to a three-phase motor where the individual three-phase lines are kept separate all the way into the motor. I have not seen the inside of a three-phase electric motor. Do the SCR cathode output wires come together inside here as well?

 

I am going to have to get back on this in a bit. We have a lot going on today, and I am quite busy, so I can't spend the time on this that i would like. Your analysis looks good, though.

The brake circuit is all in the secondary of the Z section of the step down (and isolation) transformer. The K terminals are to each SCR's cathode, and electrically common to one another and the reference ground. That is the voltage reference the controller will need to use for the switching, which will amount to sending a positive voltage to each G, or gate terminal to put individual SCR's into conduction.

 

I started editing a modified version of the original Sicon eddy current dyno controller schematic. The new version shows which lines are being tapped into and how connectors will be used.

09_01_17-SecondController.jpg

I am working on a switch mechanism for the dyno operator to toggle between manual and computer control. I originally planned to use a 6PDT mechanical switch, but I am not finding anything suitable on Google, Digi-Key or Mouser. Most of the results for "6PDT" are through-hole, low amperage (~0.7A) slide switches.

 

One switch and two three pole double throw octal socket relays?

 

That seems like a good way to do it. I tried to draw up what I think you have in mind.

This is the beginning of the schematic for the new computer control circuitry. Right now it only includes the computer and manual control switching portion.

Computer Control Circuitry, including Computer & Manual Control Switch)

I'm looking at Omron relays and a lighted toggle switch for control.

Omron MK3-P5S-AC120 3PDT 120V AC Relay

Illuminating Rocker Switch

The rocker switch has three pins. I have not used an illuminating 120VAC switch before and am not sure how the third pin is used to light the switch. I have not yet found a datasheet. Somehow closing the switch mechanically results in power passing through its internal light bulb. I don't know why the third pin is needed. Any ideas?

 

Can't find that mf'r online, but here's a link to an illuminated rocker by Cherrry - http://www.cherrycorp.com/english/rockers/lr_rocker.htm. Get the pdf for how the neon gets lit.

 

Thanks beenthere. I put in the order for the switches, relays, sockets, connectors and cable. I tried to find a large illuminated rocker switch. The Cherry TRG series is a small bit larger. I ordered four (minimum order quantity) of the TRG22F5BBRLN switch. I looked to find a version with "ON - OFF" labeling but did not hit much luck. I took a look at the PDF and the lighting mechanism makes more sense now.

 

Example "Half Control" setup with the BAP1950 SCR firing board. The three diodes that reconnect the three-phase lines to the source after passing through the load, don't appear in the original Sicon eddy current controller schematic. Perhaps they are included in the three over-lapping circles on the Sicon schematic, but aren't drawn in. Are they necessary?

 

There is a certain amount of uncertainty in the operation of the original. The lines T3, S3 & R4 might be used for phasing reference. But you also have a transformer section Y that feeds another transformer through RL, SL & TL. That transformer, labelled 34 volts, sends lines Rb, Sb & Tb over to the main controller board.

The circuitry in the 0'046'0 board may very well use the Rb, etc lines to track phasing.

The diodes shown in the example with the BAP1950 may need to be added. They are in series with the SCR's, and will have no effect on operation. You might want yet another octal relay to isolate those lines when not in use.

Again, I'm dodging bullets. I have a piece of equipment that I am trying to modify and get sent back this afternoon. If the term 'kludge' hadn't existed until now, I would have to invent it to cover what I am pulling off.

I will do some hunting for a good diode to think about using. Something like a 1N1189 might do, but the diodes would need to match the SCR's ratings. Do you have a number on them?