PDA

View Full Version : Current Sensor


Adam Shapiro
01-12-2004, 05:33 PM
Has anyone figured out/found the schematics for the current sensors? We have been trying to figure them out but without the schematics or a larger picture than the one on the wiring diagram we can't really do too much. If you have had better luck, can you post the links/info?

Rurouni
01-12-2004, 07:11 PM
We were having trouble with them as well. From the picture, it looks simple enough to assemble, but it looks as if the spot where the pwm cables are supposed to connect won't fit. Anyone have any suggestions for making that connection? Thanks.

Adam Shapiro
01-12-2004, 08:20 PM
From the picture, it looks simple enough to assemble
Where is this picture? Is it on usfirst or somewhere else?

Tom Bottiglieri
01-12-2004, 08:25 PM
what are u guys planning to do with these? for all of u with 2 speed trannys, this is a nice way to set up a relatively easy auto transmition


:] <--- i never really got this smiley

Adam Shapiro
01-12-2004, 08:31 PM
what are u guys planning to do with these?
I can't really say for everone else but I think a general agreement would be to use it to regulate motors. In the past, many (I know mine has) teams have experienced problems with the breakers flipping due to improper gearing and strain on the motors. With the sensors you can use the program to automatically reduce the motor outputs when the current goes too high.

FotoPlasma
01-12-2004, 08:51 PM
The sensors are Allegro ACS750SCA-075 chips. A sheet of paper came in the same bag with a schematic of how to assemble the sensors. They output an analog value according to the amperage through the large terminals.

Here (http://www.allegromicro.com/sf/0750/) is the manufacturer's product page, and here (http://www.allegromicro.com/datafile/0750-075.pdf) is the datasheet.

Adam Shapiro
01-12-2004, 08:55 PM
A sheet of paper came in the same bag with a schematic of how to assemble the sensors.
Really? We didn't get one! Would it be possible for you or somebody to scan the sheet and upload it to the papers section?

FotoPlasma
01-12-2004, 11:53 PM
Really? We didn't get one! Would it be possible for you or somebody to scan the sheet and upload it to the papers section?
I'll see what I can do, tomorrow.

Mark McLeod
01-14-2004, 03:59 PM
Really? We didn't get one! Would it be possible for you or somebody to scan the sheet and upload it to the papers section?


All the teams I've worked with got the circuit diagram in the programming box inside the plastic with the Nuts & Bolts magazine. Oddly enough it wasn't in the Sensor bag.

Joe Johnson
01-14-2004, 04:30 PM
Hi all,

I am very hopeful that this addition to the kit will be useful to teams.

I work with Allegro on my day job from time to time. I was very excited to see this chip in their "what's new" junk mail they send out every so often.

Anyway, after sending the data sheet to FIRST along with some samples and lots of leg work by Charlie Jost of FIRST, here they are.

For some reason the data sheet must have missed some of the kit packing.

See this link for your very own pdf data sheet: http://www.allegromicro.com/datafile/0750-075.pdf

Good luck using them.

Joe J.

Rurouni
01-14-2004, 05:50 PM
So does anyone understand how those pwm connections are working? Like I said before, either we're reading the schematic/picture wrong, or it looks as if we're going to have to cut the pwm cables to make them fit where they're supposed to. I hope someone can get back to me on this one.

josh_johnson
01-14-2004, 06:55 PM
So does anyone understand how those pwm connections are working? Like I said before, either we're reading the schematic/picture wrong, or it looks as if we're going to have to cut the pwm cables to make them fit where they're supposed to. I hope someone can get back to me on this one.

I only got to see the sensor parts briefly while inventorying our kit but I thnk that there was a plastic piece with three holes for the pwm cable on one end and pins on the other to fit the circuit board. I am not certain, but that may be the piece you are looking for.

Jay Lundy
01-14-2004, 07:18 PM
So does anyone understand how those pwm connections are working? Like I said before, either we're reading the schematic/picture wrong, or it looks as if we're going to have to cut the pwm cables to make them fit where they're supposed to. I hope someone can get back to me on this one. Well the PC board came with 3 rows of 3 pins. Each row is obviously connected together.

The first row (closest to center) is for the chip itself. It should fit snug in those 3 holes with the big leads in the holes on the other side of the board.

In the next row we put the capacitor. I can't remember exactly what pins it should go in, refer to the sheet included in the kit for that (pay attention to polarity).

The last row is where we put the pwm cable. You may have to trim a few strands from each wire to get it to fit in the holes. Make sure you get the wires in the right pin (white -> output, red -> Vcc, black -> ground). See the datasheet for what each pin on the chip is for.

One problem we had was soldering. The small pins were easy, but the bigger leads were more difficult. It was impossible to do with some of our 850 F irons, but mine at home was hot enough to do.

We also were not sure how to solder the 10 AWG wires in those holes (which are much bigger than a 10 AWG wire). We ended up running a wire through the holes and along the back of the PCB. Then we removed the insulation around where the wires touched the metal rings and soldered it there. Then we cut the wire short and added connectors on the ends. Finally we used liquid electrical tape to seal off the bare metal. Once again it was difficult to solder, but we managed.

I think this thing is great. The first day we set up a simple experiment where we had the CIM reverse directions at full speed (no load) every 500 ms. When the motor changed directions, we were reading current of up to around 60 - 70 amps. Then the 30 amp breaker started getting hot and it started tripping. We plan on doing more experiments and hopefully use it to determine problem spots in our drive train (tripping breakers has always been a problem with us).

FotoPlasma
01-14-2004, 08:49 PM
We also were not sure how to solder the 10 AWG wires in those holes (which are much bigger than a 10 AWG wire).
When I first saw these terminals, I thought of ring connectors with nuts and bolts. I don't have them in front of me right now, so I can't say how effective my thoughts would be, but I have a hunch that we weren't supposed to solder the 10AWG wires to the board, itself.

I have yet to do anything involving the current sensors, sadly.

Garrix
01-15-2004, 12:40 AM
If you look at the Robot Power Distribution (http://www2.usfirst.org/2004comp/2004_Robot_Power_Distribution.pdf) diagram, you'll see the the positive wires running from the motors are routed through the large holes in the sensors.

FotoPlasma
01-15-2004, 12:42 AM
If you look at the Robot Power Distribution (http://www2.usfirst.org/2004comp/2004_Robot_Power_Distribution.pdf) diagram, you'll see the the negative wires running from the motors are routed through the large holes in the sensors.
Positive, actually, but that's irrelevant because the Victor speed controllers can reverse the polarity of their output in order to run the motor in reverse. In the datasheet for the chip, you'll find that it outputs 2.5V (the middle of the range) at 0Amps, and goes both positive and negative from there, 5V being some full Amperage forwards, and 0V being some full Amperage backwards.

Garrix
01-15-2004, 12:50 AM
Actually you don't need to solder the 10 AWG wires into the holes. You can solder the 10 AWG wire to a large circular connector, and then you can bolt on the lead. I would suggest using a paper washer on the bottom before you screw the nut on. This is easier than soldering the wires in directly, and you can remove the current sensor more easily.

Random Dude
01-15-2004, 10:27 AM
Actually you don't need to solder the 10 AWG wires into the holes. You can solder the 10 AWG wire to a large circular connector, and then you can bolt on the lead. I would suggest using a paper washer on the bottom before you screw the nut on. This is easier than soldering the wires in directly, and you can remove the current sensor more easily.


Why a paper washer? The object is to make a connection with the board, not insulate from it.

Jay Lundy
01-15-2004, 02:06 PM
Yeah I was mounting a new motherboard today and the places in the PCB where the screws went had the same ring of metal around them as on the sensor PCB.

But the soldering worked pretty well and with the connectors on the ends of the wires it makes it easy to add/remove from the circuit.

Garrix
01-15-2004, 08:13 PM
Why a paper washer? The object is to make a connection with the board, not insulate from it.
You connect the ring terminal to the board, metal to metal, and before you put the nut on, you put on a paper washer, then you put the nut on.

Remember when soldering to add a heat sync to the sensor so you do not damage it. I know that FIRST has suggested using a small piece of aluminum, but I found that I got better results using an alligator clip on the pins that I was going to solder. Also if you mess up and fill a hole with solder, and you still need to put a pin in it, just simply touch your iron to the solder until it turns liquid, and use a high pressure spray of some sort (air or endust) and the entire area is clean; just make sure that noone is opposite of you when doing this cause noone likes being sprayed with hot solder. Wicking could also work to clear out the holes, but it would take longer, and is not nearly as cool :)

Another note of advice, always solder from the bottom of the circuit board where the pins are, and if your solder job is good, it will be shiny, and if it is dull you gotta redo the joint cause you got a cold connection.

Adam Y.
01-16-2004, 08:14 AM
Remember when soldering to add a heat sync to the sensor so you do not damage it. I know that FIRST has suggested using a small piece of aluminum, but I found that I got better results using an alligator clip on the pins that I was going to solder. Also if you mess up and fill a hole with solder, and you still need to put a pin in it, just simply touch your iron to the solder until it turns liquid, and use a high pressure spray of some sort (air or endust) and the entire area is clean; just make sure that noone is opposite of you when doing this cause noone likes being sprayed with hot solder. Wicking could also work to clear out the holes, but it would take longer, and is not nearly as cool
I would reccomend a solder sucker or a solder wick. Those tools are very cheap and are a lot safer than an air blast. The alligator clips do work since I used them in my electronics class to solder items. Also, a tool with adjustable alligator clips to hold your project comes in handy.

Al Skierkiewicz
01-16-2004, 09:57 AM
.
Also if you mess up and fill a hole with solder, and you still need to put a pin in it, just simply touch your iron to the solder until it turns liquid, and use a high pressure spray of some sort (air or endust) and the entire area is clean
Garrix,
Just tapping the circuit board on the bench after you heat up the pad works fairly well also. But a better hint is to solder the #10 wires to the board before you add the sensor, that way you don't have to worry about heatsinking the sensor leads.

_GP_
01-16-2004, 03:10 PM
so is there a posted version of the assembly instructions?

Greg Ross
01-16-2004, 04:22 PM
so is there a posted version of the assembly instructions?
It's now at http://www2.usfirst.org/2004comp/Specs/Current_Sensor.pdf

Eugene
01-17-2004, 06:22 PM
Is it possible to order the complete kit from Allegro? If not, where can I get the PCB and the capacitor?

Thanks!

Eugene

Phil_Lutz
01-19-2004, 12:01 AM
(tripping breakers has always been a problem with us).

Can you post a picture of the finished product. We will be building ours tommorrow night.

Thanks,

Phil

lpramo55
01-19-2004, 11:11 AM
:] <--- i never really got this smiley
thats a lego head :D

Vladimir
01-23-2004, 12:45 AM
I believe I have the current sensor wired up properly but the controller reads a number around 867 and never moves from that, no matter how much current I pass through it. I'm sure the PWM cable to the RC is aligned correctly. I tested the analog inputs with another device and they work fine (0-1023).. if I disconnect it the value floats around as it should, connect it, 867. Any ideas? Is it defective? I heatsunk it while soldering to make sure it wasn't damaged.

Jay Lundy
01-23-2004, 01:19 AM
I believe I have the current sensor wired up properly but the controller reads a number around 867 and never moves from that, no matter how much current I pass through it. I'm sure the PWM cable to the RC is aligned correctly. I tested the analog inputs with another device and they work fine (0-1023).. if I disconnect it the value floats around as it should, connect it, 867. Any ideas? Is it defective? I heatsunk it while soldering to make sure it wasn't damaged. First of all use a multimeter to make sure the ends of your wire are actually connected to the big leads (just to make sure). Also make sure the big leads aren't shorted together somehow, thus bypassing the sensor alltogether.

Then try using the multimeter to read the voltage between the output pin and the ground pin as you pass current through the big leads. If your voltage changes but your input doesn't you have a problem somewhere besides the sensor itself.

That's all I can think of right now.

Vladimir
01-23-2004, 11:02 AM
I already checked to make sure all the connections were made properly (end of PWM cable pins to the pins on the chip, big leads to the big leads on the chip), I did not check resistance across the big leads, I will do that today. What should it be? I know it has to be very low because it is an ammeter...

Jay Lundy
01-24-2004, 05:29 PM
I already checked to make sure all the connections were made properly (end of PWM cable pins to the pins on the chip, big leads to the big leads on the chip), I did not check resistance across the big leads, I will do that today. What should it be? I know it has to be very low because it is an ammeter... I don't have ours right now to test with, but shouldn't it be basically 0? It's just a piece of wire that when current flows through it generates a magnetic field that is read by a hall element.

http://www.allegromicro.com/sf/0750/

CNCBoy461
01-26-2004, 10:13 PM
We got our current sensors to work on the first try: from my understanding of the diagram given to us, the 3 pin mount for a PWM cable seems to go with the IR Beacon kit. So, I took the shell off of a PWM cable's male side and left the 0.100" IDC terminals on it. I put those terminals through the PCB board and they held quite well.
As for testing, we used a globe motor as the inline motor so as not to draw too much current on our first try. In the dashboard viewer and debug window, we got a standard reading of 507ish. After grabbing the drive shaft (With Gloves!) we got a spike of 570ish but dropped back down to the standard 507. The globe is very hard to stop and doesn't draw that much current anyway and so thatís the reason for the low spike, but in a few days we should have our chassis done and on the floor so that we can really test the sensor's full ability.
Reminders: Make sure to use a heat sink (And maybe a can of air), Use an iron that is from 700-900F, and test all of your connections...short circuits equals no working sensor.
As for the question of "Where to buy the sensor package?"...WE DON'T KNOW!!! Will someone please post a reply to this, we would like 2 more of them...and if not, what's the voltage rating on the capacitor? 4vdc or 6.3vdc? Good luck all!

www.boilerinvasion.org and see pictures of the sensor (Soon) and a CAD model of the new Maxi Fuse Block.

Al Skierkiewicz
01-27-2004, 06:41 AM
I believe I have the current sensor wired up properly but the controller reads a number around 867 and never moves from that, no matter how much current I pass through it. I'm sure the PWM cable to the RC is aligned correctly. I tested the analog inputs with another device and they work fine (0-1023).. if I disconnect it the value floats around as it should, connect it, 867. Any ideas? Is it defective? I heatsunk it while soldering to make sure it wasn't damaged.
Vladimir,
As many have suggested, get out your multimeter. Check first that you have 5 volts at the right pin on the chip. Then check the output pin, with no current flowing it should be at 2.5 volts. If it is very low, check that the capacitor has been inserted correctly. Double check that there is no solder bridge between pins or adjacent circuit pads, top and bottom of board. If it is at 2.5 volts then check that you have 2.5 volts at the input to the RC, PWM cables are known to go bad. It is very common for a three wire cable to get reversed when looking at it upside down so double check that the right wire goes to the right hole.
Now, if anyone has any real experience with these can you express whether external magnetic fields affect the output? Thanks,

Biff
01-27-2004, 07:19 AM
As for the question of "Where to buy the sensor package?"...WE DON'T KNOW!!! Will someone please post a reply to this, we would like 2 more of them...and if not, what's the voltage rating on the capacitor? 4vdc or 6.3vdc? Good luck all!

www.boilerinvasion.org and see pictures of the sensor (Soon) and a CAD model of the new Maxi Fuse Block.

The Cap is limiting noise to the controler and slowing the response(slightly), voltage should be 6.3 but a higher rateing will not hurt anything and is a saftey factor. Tanyalem Caps short when they fail. Check with ohm meter, should read low to start then go to a high vaue as the cap charges. Any thing below 10 ohms after a few seconds and the cap is hosed. You allready mentioned not putting them in backward. The ACS750-75 has a link to buy more sensors at http://www.allegromicro.com/sf/0750/ don't know about the circut board Good luck.

CNCBoy461
01-28-2004, 09:28 PM
I posted here...didn't want to do it again: http://www.chiefdelphi.com/forums/showthread.php?t=24004
Thanks Biff for the voltage on the cap.

AustinPowers
01-29-2004, 11:10 AM
Make sure you are using the Get_Analog____ (Ican remeber the complete name but it is in IFI docs) routine to get the information from the Analog port. When using printf use an (INT) casting. The doc show how to read the port. I second using the mutltimeter to make sure the sensor is connect properly. Use a potentiometer on the analog port to check to make sure your program is working. Any questions just ask.

Mercutio
01-30-2004, 06:52 PM
How do you tell the polarity of the capacitor? I'm holding two big mustard-colored things taped on a strip that look like capacitors. They each say

157
+16K|

in brown writing on one side. Which side has the positive lead? I've heard that bad things can happen when you hook up electrolytic capacitors backwards, so i'm a little nervous about just doing trial and error.

~Aaron

JRHaddock
01-31-2004, 12:32 AM
The plus sign (+) by the number 16 indicates the positive terminal of the capacitor.

Stephen P
01-31-2004, 11:39 AM
Are the current sensors necessary by FIRST rules cause we cant find ours.

Damian Manda
01-31-2004, 03:24 PM
Are the current sensors necessary by FIRST rules cause we cant find ours.
No, the current sensors ar not required, but can be very helpful in determining problems with motors, or just monitering them during competition. We have yet to build ours, we have been concentrating on making the IR work.

Biff
01-31-2004, 04:53 PM
How do you tell the polarity of the capacitor? I'm holding two big mustard-colored things taped on a strip that look like capacitors. They each say

157
+16K|

in brown writing on one side. Which side has the positive lead? I've heard that bad things can happen when you hook up electrolytic capacitors backwards, so i'm a little nervous about just doing trial and error.

~Aaron
The other way to tell is if one wire on the cap is longer it is the positive, this is kind of a standard in the industry. If the are the same or have been cut look for the marking, IE as the next post had said the lead next to the positive sign + is the postive. Some kinds of Caps are marked on the negitive side only.

B. Back
02-03-2004, 04:12 PM
The current sensor pcb can be purchased from UPE-INC. Their phone number is (330)659-9287. As someone else noted, the pcbs cost $12.50 each and there is a min. qty requirement of 2 units. They are currently out of the boards, but expect more in house in about a week.

Biff
02-07-2004, 08:46 PM
This is a kind of on subject question. I was reading a white paper on making current sensors by using an Op amp to get usable input voltages from the voltage drop across the fuses. The paper was written by Mr Lewis, mentor of team 811, back in 2002. I like the design, but after exhaustive research, the only acquireable parts are surface mount. Not a DIP in stock at the approved vendors or even some brokerage houses. He mentions the resistance of the 20 amp fuses at about 9 milliohms and the 30 amp at about 6 milliohms. I don't have a meter that will get me readings that low. Does any body have a resistance value for the 40 amp breakers or a test rig to get there under load. If I do the circuit with surface mount I don't want to take the gain resistors on and off to get a usable range. Thanks in advance.

Al Skierkiewicz
02-09-2004, 06:21 AM
I like the design, but after exhaustive research, the only acquireable parts are surface mount. Not a DIP in stock at the approved vendors or even some brokerage houses. He mentions the resistance of the 20 amp fuses at about 9 milliohms and the 30 amp at about 6 milliohms. I don't have a meter that will get me readings that low.
Biff,
Op amps are pretty "vanilla" in that you can drop almost any device in place for a given circuit and it will work. In this application you do need a low input offset voltage specification. I remember at the time that a lot of discussion occured over different methods of measuring current. While the breakers do exhibit some resistance, this a variable especially if the breakers have ever tripped. We used the resistance across a 1 foot long #10 guage wire as it is close to .001 ohms/ ft. Remember to measure the current on the supply side of the speed controller as this doesn't change polarity.

Biff
02-09-2004, 10:33 AM
Biff,
Op amps are pretty "vanilla" in that you can drop almost any device in place for a given circuit and it will work. In this application you do need a low input offset voltage specification. I remember at the time that a lot of discussion occured over different methods of measuring current. While the breakers do exhibit some resistance, this a variable especially if the breakers have ever tripped. We used the resistance across a 1 foot long #10 guage wire as it is close to .001 ohms/ ft. Remember to measure the current on the supply side of the speed controller as this doesn't change polarity.
Thanks for the heads up about resistance variablity after tripping. The white paper I quoted was using "rail to rail" input op amps. These are findable but mostly in surface mount. If I get time I may take a run at it and go with a gain factor, for a given wire length. I am planning at this point to go with #6 right to the supply side of the Maxi 40 amps and drop to #10 or #8 to the Victors. If we can keep the wires from the Maxi's to the Motor victors the same length that would be the ideal sense point. Thanks Again.

Dale(294engr]
02-09-2004, 09:19 PM
ACS750 see my previous post.

Appears FRC doc had error- 150uF shown as an output filtering.
(seemed logical for the 2kHz PWM square wave noise)

Allegro doesn't recommend such a large filter.

The 150uF tantalum may rather have been intended to filter the +5V power !!
(I received no replies to that post question)

Tantalum capacitor polarity can be safely verified using a bench power supply
current limited to a few mA.
Set PS to 6V.
No harm if Cap backwards.
If Backwards the voltage just goes down to a volt or so at the current limit.
(if there were no current limit tantalum caps can explode like a fire cracker
which is dangerous. (less drastic, cap gets warm to hot, degradiing cap)
If you must use a full lexan shield and will need a lot higher voltage
to force a very high current)
Correct cap orientation, the voltage spikes momentarily resting at 6v, 0 mA.

No bench supply ?
Simulate one.
Using your 5V power source of choice, (RC +5v max is ~1A limited by 7805?)
place a current limit resistor in series with the Cap.
(Rlimit =5v/1mA(.001A) =5k ohms)
Place DMM across Cap., place Cap in both directions)


I'm using 274 ohms in series with ACSout to 150uF for filtering
(~125mS = 3 time constant)
and additional Cap(s) for +5V supply filtering at the sensors.

ACS750 = 20mV output for each Amp sensed.

0A = 2.500V out (ratio, i.e. 1/2 of +5.00 v supply)

Positive current sensed: 2.500 + 20mV/A (512 counts + ...)
Negative current sensed: 2.500V - 20mV/A (512 counts - ....)

Analog scale factor: 5v/1024 counts = 4.88mV/count
Analog zero: 2.5V=512 counts = 0 A
then +-20mV/4.88mV/count per amp sensed analog measurement

+1A = 512counts + (20mV/4.88mV/Count=) + ~4 counts = 516 counts
-1A = 512counts + (20mV/4.88mV/Count=) - ~4 counts = 508 counts

The PWM cables do fit the ACS PCB
though a bit of finess is necessary.

Use of the PWM male crimped pin & housing is advisable for effective strain relief.
Solder the male PWM pins into the outside PCB row.
Using the male PWM connector effects an effective wire strain relief
from vibration, frequent moving/testing/etc.

Direct soldered wires require separate strain relief to avoid wire breaking where the stranded wire becomes stiff with solder.

I'll send/post pictures if requested.

Dale.Hall@ngc.com

hay4man
02-09-2004, 09:32 PM
I would be obliged if you could post that. I am interested to see what you have to recommend. Our team just cut the male ends off and directly soldered the PWM cables to the PCB.

Dale(294engr]
02-11-2004, 01:21 AM
Hay4man & all
Here are pictures of PWM male end
pressed fully into the outer 3 holes of the current sensor PCB
bent flat onto the PCB for firm cementing
then soldered.

The stranded wire crimped PWM connections provide a good strain relief.

If the stranded wires are soldered directly to the PCB
it is important to anchor them
else movement will cause breakage at the stiff solder joint.

A stranded wire joint is vibration resistant if crimped (multi wire suspension)
but when soldered turns into a solid wire joint not preferred for
vibration/movement prone areas, like cars,
cause solid joints are prone to crack and break easily.

Bottom
F:\Robotics\2004\75AcurrentHallSensor\P1210044Curr 75Asensor1bott.JPG

Top
F:\Robotics\2004\75AcurrentHallSensor\P1210043Curr 75Asensor1top.JPG

picts may not have inserted (my 1st try I did copy/paste below & attach files)
(& they didn't show up in preview) so
email me & I'll send send them direct.

& plse instruct how to insert .jpgs from local HD into posts....
(or to preview replies ?? preview appeared to only be of orig post)

Dale.Hall@ngc.com

Al Skierkiewicz
02-11-2004, 06:06 AM
']Hay4man & all
Here are pictures of PWM male end


Dale,
It looks pretty good. Nice job on the insulation wrapper. This is not well stated in the docs, all teams should remember to adequately insulate the board. A big piece of heatshrink works very well.