| ||||||||||||
| ||||||||||||
| Comparing Lap Counter Activators |
| By Steve Medanic |
Any lap counter (PC Based or other) must first detect the car and then count the car. These are not the same function but are in fact two separate functions. The lap counter system must be fast enough to detect the car but be slow enough to not count laps when there is no car on the counter or when a car on an adjacent lane crosses the counter. The system must also not count multiple laps when a car “parks” on the lap counter section.
Time
The biggest problem one deals with in the design of any lap counter is the speed of the car and the very very short time the car is spent in the lap counter. Assuming the car is traveling at 20 feet per second the time spent across the lap counter can be approximately:
1" dead long strip + ½" long pickup shoe = 0.006 seconds
2" long body + 1/8" long optical sensor = 0.009 seconds
½" long magnet + ½" long reed switch = 0.004 seconds
Twenty feet per second in a straight line is a good speed for a Tyco 440-X2. A good current generation superstock or faster car just makes the situation worse as the time is less. Obviously the longer the time ANY sensor detects the car the better the chance it will detect and count the car. This works in favor of the dead strip as it can be made as long as necessary. An optical sensor would be the next best followed by a single reed switch installation.
Reed Switches
The reed switch is dependent on the biggest variable in a slot car. That variable is the magnet setup. One thing about reed switches is that on an in-line motor car with separate motor and traction magnets the car can trigger the reed twice every time the car passes as a result of the orientation and interaction between the motor and traction magnets. That can help the situation but unfortunately you can’t count on it as there are two magnet setups for these types of cars. There is a high down force setup and a second setup with less down force but more centering action. With the high down force setup the reed is triggered twice as the car passes the switch. With the low down force setup the reed is triggered only once as the car passes the switch. The lap counter must accommodate both types of cars. On a car with just motor magnets the reed is triggered once as the car passes the switch. Pancake motor cars have the magnets oriented differently then the in-line cars do and their magnets are typically weaker and located further away from the track than cars with in-line motors and magnets. Some tracks run classes with the old G+ car that had flux collectors. Modified AFX chasses can also use flux collectors. The Lifelike M car uses a magnet setup that is different than any of the above. The reed switch must accommodate all of these different types and strengths of magnetic fields.
The reeds switch must be oriented close to the rail with the reeds inside the switch oriented parallel to the rail. The reed must be sensitive enough to pick up the car with the weakest magnets but not so sensitive that it picks up adjacent cars with strong magnets. Cross talk between lanes (a car picking up more than one reed switch when it crosses the lap counter) has been an issue with reed switches for many years. The R.O. main at the 91 HOPRA nationals was rerun for that very issue. Somebody lost on that deal as the winner of the second main was way behind when the first main was called off. One recent enhancement for reed switch installations is to place a steel rod under the track between each lane to isolate each lane and prevent a car from picking up more than one reed switch. I haven’t tried it but the idea does have merit.
The problem with reed switches in my mind is that the manufactures are starting to decrease the number and quality of reed switches. We got some reed switches from Newark Electronics several years ago that were phenomenal. They couldn’t be beat. They counted everything and didn’t miscount when there was an unlimited running in an adjacent lane. Naturally, the manufacturer discontinued them without warning. The replacement reed switches had sloppier tolerances and were easy to fool. We ended up tearing out the reed switches at one facility and going to optical sensors as we were halfway done with the installation when Newark dropped the good reeds.
Dead Strip
With a dead strip the transition from powered rail to dead strip and the entrance to the dead strip must be very smooth. A rough entrance can result in bouncing a pickup and failure to count as the pickup can fly over the dead strip. We had a dead strip failure at a NITRO race several years ago. In testing the counters worked flawlessly. In the race they failed miserably. In this case the problem was the entrance to the dead strip was very bumpy and the pickups on the faster cars would float over the 3” long dead strip without making contact except for the initial hit. Testing identified the problem and this was pointed out to the track owner. Unfortunately it was not fixed prior to race day.
Dead strips are also polarity sensitive and typically require an interface board to buffer the signal between track and PC. The dead strip must be located where the car will coast over it without interfering with how the track is driven. Dead strips were widely used in 1/24th and HO scale racing. They are still popular in 1/24th scale. Their use in HO is fading but there are those who swear by them.
Optical Sensors
With an optical or IR setup too much light can overload the sensor resulting in sluggish or poor performance. You want just enough light to bias the sensor "open" when the car is not on the counter. That also eliminates problems associated with too thin a paint job. You also want a narrow sensor beam that is easily broken by the car. The sensors must also be located so that the chassis triggers the sensor. That maximizes the effective length of the sensor.
IR Sensors & Light Bridge
There is no glue visible in the above photo holding the sensors in place. Silicone and super glues can fog or craze the lens on the sensor or emitter resulting in problems. I found out the hard way that clear silicone will not allow IR to pass through. Others have found that super glue fogs the sensor rendering it useless. The trick is to locate the sensors in such a way that no glue is required. The folks at Brystal tracks mount a piece of angle below the track and anchor the sensor wires to the angle with nylon wire ties. This allows one to fine tune the installation and eliminates the need for glue to hold the sensors in place. I borrowed the idea and the following photos show the sensor wire installation on my track.
IR Sensor Anchor
In the photo you can see the angle and how the wires are held in place by the tie wraps. In my installation I first heat shrinked some material over the sensor wires to give the wires some additional volume and something for the tie wraps to bite into. Then I lightly tightened the tie wraps. I then pulled the sensors into position and then did the final tightening. In case the sensors are too far down its easy to cut the old tie wrap off and try again. As can be seen my angle was set up for a maximum of two tie wraps per lane. The second tie wrap was not required. The box in the background is for my reversing switches which are hidden under the track. The other wires are for the track call switches, main power, remote current/voltage displays and tripped breaker alarms. As you can see from the following photos it’s a bit crowded under there but it all fits and it all works. Details of how I did most of this magic can be found on the Siberia Racing website.
Relay, Circuit Breakers, Track Call Switch and Buzzer
IR Sensor Anchor and Reversing Switches
DC Stiffening Caps, 8 AWG Main Power Wire and Alum. Buss Bars.
Which Sensors Are Best?
This subject has started many a heated discussion. All of the above systems have problems that, if ignored, can cause them to fail. Conversely reed switches, dead strips and optical/infrared sensor lap counters can all be made to work exceptionally well. There is a direct correlation between the amount of research done before hand and the amount of care and skill that is placed into the design and installation and the quality of the end result. This is one place where the more time invested the better.
Locating the Lap Counter
The best place to locate the lap counter section is after a turn in which cars tend to crash and not change lanes when they come off. In addition the turn should be tight enough that riders also fall off as opposed to going through the turn or jumping lanes. A sharp 90 or 180 turn at the entrance to the lap counter straight is best . Forty-five degree turns are the worst as cars tend to change lanes on these turns as opposed to just coming off. The counter section should also be in the middle or end of a significant straight. If the track is reversible then it’s even tougher to site the lap counter as you have to meet the above criteria when running in both directions. CRR’s lap counter is in the center of the front straight. In both directions there is a wide sweeping 90 degree turn followed by a hairpin. Cars don’t change lanes in the sweepers and they crash and burn in the hairpins. Jim Nagy’s Slot Pro Speedway also has a lap counter that meets the criteria in both directions. It can be done!
My recommendation is to build a version of the track first and play with it until you find the best location for the lap counter. The location will become obvious with time. In several cases tracks were modified so the lap counter straight could meet the above criteria. In every case the track’s drivability improved as a result of the change.
Hardware/Software Update Time
So much for the sensor types and how to locate the counter. The other major factor in PC lap counter performance is the design of the PC interface. We are not talking about processor type and speed but the speed of the PC serial or parallel ports. Most of the systems where the PC directly interfaces with the track use the parallel or printer port. A history of the port can be found at HYPERLINK "javascript:ol('http://www.fapo.com/porthist.htm');" http://www.fapo.com/porthist.htm.
This page indicates that the printer port has known problems associated with cable length. The page also indicates that systems that interface with the port must be compatible with the ports driver software and are very software intensive. The number of instructions to determine that a car has passed over the sensor is considerable and software update time can be an issue. Update time is the time necessary for the sensor to be closed in order for the software/hardware to determine that a car has crossed the lap counter section. Update time is real and detectable and if your sensors are too fast miscounts will result.
I first identified update time as a problem when I helped a track owner adapt an electrical relay type counter to an HO track. Much fine tuning of the relays and voltages on the dead strip was needed to make it work. The term update time wasn’t used but it applied in that case as many times you could see the relay pick up but it would not be closed long enough for the rest of the system to respond. I first used the term “update time” when I wrote a home made lap counter program for the Atari PC system in the 80s. I needed to build in a delay into the track interface to slow the signal down so that the PC could identify that the car has passed over the lap counter.
I was surprised at the number of steps required to get the car to count a lap. The lap counter loop of the program did the following: Step 1 recognized when the sensor indicated that a car was triggering the counter. Step 2 set a register to indicate that the lane 1 sensor was occupied. Step 3 determined when the sensor was not occupied. That sequence determined that the car had finished crossing the lap counter and it was safe to count the lap. Step 4 set the lanes register back to zero and incremented other registers that counted the number of laps that the car had turned to date overall and in that segment. If you wanted to calculate lap times several more steps are required to poll the system clock and to do the math. If you wanted to ignore laps that were faster than a set minimum time even more steps are required. Then after all that you had several steps that updated the display. This was for one lane. You don’t multiply the above by four for a four lane track its more like multiplying the number of steps by eight. Much interweaving of steps was required to get the cycle time down to a reasonable time.
After some fine tuning and testing I got it working to my satisfaction and took it to a race to show it off. With the organizers permission it was installed in parallel with the tracks normal lap counter system during practice and immediately we identified problems with the tracks normal lap counter system. The track owner was sure that his counter was fine and my addition caused the problem. My lap counter was removed and the normal LED display based electronic counter was reinstalled for the race. Things were fine until sometime during the semis and the mains the counters started miscounting as the cars were crossing the sensors faster than his lap counter could register. The designer of the LED display system was there and he and I discussed the problem in front of the track owner who was melting down as his track wasn’t working and people were starting to manually count laps. One of us used the term “update time” and I have been using it ever since. This was on the Superstock track at yet another national event. If my memory is correct this track had a dead strip. Note the same thing happened with an optical system with an early LED counter at yet another big race. In this case cars with thinly painted bodies or bodies with big unpainted windows didn’t count. The fix fortunately wasn’t hard and once the bodies had all the windows blacked out and there was enough paint on the bodies the counter “seemed” to work fine.
As has been stated several times dead strips, reed switch installations and optical installations can all fail. Likewise ALL of these sensor systems can be made to work without problem. It all depends on the design of the interface and how much TLC was put into the installation. In the case of the track above the system was not tested properly. The lap counter worked with slower cars but would not count the faster ones.
Sometimes You Get What You Pay For
I have had very good performance with reed switches and with IR sensors plugged into the printer port using Trakmate DOS. Once I started using the Trakmate for Windows with the track interface card my IR based lap counter system has been bulletproof. The Trackmate interface card is actually a small microcomputer that fits under the track. The card is the lap counter and counts laps, computes lap times etc. The interface card does all the work and the PC essentially becomes a "dumb terminal" where processor speed and software update time are no longer issues. Both the Trakmate for Windows and the Moonstone Slottrack race management systems can interface with the Trakmate interface card. The card will support reed switches and dead strips as well as IR.
Other DOS based systems such as SRT and Lanemaster use interface cards. These cards may do the same thing as the Trakmate microcomputer card. In these cases the card does the work and the PC is not really that important as long as it will support the card and the program. None of these systems are free but they all work very well. That lap counter freeware available from the web might be attractive but remember that sometimes you get what you pay for. If you plan on hosting races its time to cough up the cash for a good lap counter/race management system.
Testing, Testing, Testing
Testing is the final step of any lap counter system. Testing must be done under controlled conditions and not just during some practice session before the race. I have developed tests for just about every sensor system available. I test IR installations with a ½" wide strip of card stock taped to the front of my G3-SS car as shown in the following photo.
Superstock IR Sensor Test Rig
The card stock is long enough to trigger all four IR sensors at the same time. All four sensors must count each and every time when the car crosses the sensors. The car is placed on the track and driven as fast as possible. I can get right down to the class track record with this setup. From this rude and crude test it appears that the Trakmate system has an update time of less than 0.002 seconds. A shorter test strip has resulted in failure of one or more lanes to count so the update time is real and is detectable. Considering that a short car is four times longer than my test strip I have confidence that my IR system can count anything. The other thing to do is to turn off the room lights when it is dark outside and verify that the bridge can supply enough light to bias the sensors off when they are uncovered with no external help. This test tells you if the IR emitters are getting help from some window or room light that could cause false counts should this alternate light path be turned off or blocked during the race. With rigs like the one shown in the following photo I also test minimum lap times and how the counter will deal with 100, 1000, 10,000 etc. laps in a segment and/or race.
Lap Counter Test Rig
I have seen lap counter systems fail at one or more of these magic numbers. A failure at 1,000 laps or 5,000 laps doesn’t sound like a problem until the lap counter crashes during the last segment of an enduro when the lead car hits the magic number and the lap counter melts down. I do these long duration tests so I can identify where the weak spots are and prove the counters can take anything the race can be expected to throw at them. These tests have been done on the Trakmate for Windows and SlotTrack race management systems. I have helped in the development of both systems and use both systems. I don’t have a favorite as each system has features that make it the best for a specific application.
I test dead strip installations by taping off a portion of the dead strip and making sure the car counts every time when driven at speed. This test indicates that the dead strip is longer than needed. I personally like the dead strip to be at least twice as long than required. The tape test can also tell you if there is a problem with either the font or back of the dead strip.
Sorry, but I have not found a good test for reed switches. You just have to run everything you have and make sure they count every car every time. Weak magnets, strong magnets, multiple magnet setups, etc. After all that you have some confidence (but no guarantee) that the reeds will count the other guys car. Don’t just assume it will work for everybody just because it worked for you. Heard about an incident at a race recently when the track owner “tested” the reeds before race day only with his personal XXXX style cars. A big noise was then made about the fact that the counters had been tested and if you had a problem with the counter make sure you have a spare car as the problem was with your car and not with the track.
Well . . . . it seems that there were two other classes of cars to be raced on that track that day. The track owner’s assumption was that if something worked with one thing it will work with everything. You can guess what happened. The reed switches did very well counting the XXXX style cars. The other two classes had to be counted by hand as the counters either wouldn’t pick them up at all or would work sporadically at best.
The Bottom Line
I cannot say this enough. Testing and TLC in selecting and installing a lap counter system is critical. As those posting on HOWL and other boards have indicated, anything can be made to work if installed carefully by someone who knows how to test and tweak the installation. Mike King likes dead strips. SCB and others prefer reed switches. I and others prefer optical systems. There is room here for all. The newbee must be careful and choose wisely. Every one of the race management systems mentioned has good support and costs a few dimes to buy. The bottom line is that they work! Freeware might work but then again you just might get what you paid for.
As I said many a time I personally have had good results over the years with dead strips, reed switches and optical sensors. I have also seen each type of installation fail. I am in favor of optical (especially Trakmate for Windows) installations as they go in easy with minimal or no tweaking. This is especially true if the sensors were press fitted or secured in some manner that does not use glue or silicone to hold them in place. The basis for my recommendation as, in my experience, the other systems take more work and require more of a knowledge base to get right.
Daniel sells a fine light bridge kit. For those who wish to make their own IR light bridge Radio Shack makes a good high power IR LED that works well with the Trakmate IR sensors. Make sure that the sensor and the LED are aimed right at each other and the current through each IR LED is pure DC between 80-100% of the LEDs current rating. Note that some "wall warts" are not filtered and don't put out pure DC. The IR LEDs can pulse on and off between 60 and 120 times per second when powered by one of these junk power supplies. These pulsating counters can cause serious problems if they work at all. When in doubt install a 1,000 mfd (or larger) electrolytic cap on the DC line from the wall wart to the IR LEDs. Don’t assume the power supply voltage is what it says on the wall wart. These power supplies are typically not regulated and will often put out more voltage than rated when unloaded and less voltage than rated when loaded with as little as 100 mA.
The difference between unloaded and loaded voltages must be taken into account with an optical system. The voltage and current from the wall wart to the optical emitters (LEDs) must be checked under load and the LED in-line resistor values adjusted as necessary to get the LED current in spec. Finally, for best results, the IR sensors should be exactly in line with the emitters, located no more than three inches from the IR LEDs and no glue or silicone should be on the lens of either the LEDs or the sensors. For those of you wanting to build your own light bridge I developed a excel spreadsheet that will help size the dropping resistors. A copy of the spreadsheet is available as a link to this article.
Hope the above has been of some help. Good luck and good racing!
Steve
