Index Home About Blog
From: John De Armond
X-Source: The Hotrod Mailing list
Date: Aug 1992
Subject: Re: Telemetry

>I was reading through Super Chevy the other night, and in
>their recent Tech Tips list, they mentioned a 'low buck
>race computer'.  It sounded a lot like a tape recorder to
>me, so I discounted it.  Then I started thinking about it
>more- if you had a noisy ignition system, you *could* tape
>the radio interference and then get a gross RPM graph with a
>scope trace of the recorded sounds.

Don't discount the lowly audio tape recorder as a data acquisition unit.
After all, this is what the very expensive AutoMeter recording tach uses.
Stay tuned for a clone version of that in PE :-)

Gross ignition pulses CAN be recorded and played back to something like
a tach circuit.  Much better is to use a single chip Bell 202 modem
chip to have the ignition toggle between two audio carrier frequencies.
This makes data recovery vastly simpler.

>So what's an equally low-buck way to add throttle position,
>transmission gear, MPH, vacuum, and fuel pressure data to this?
>I'm no EE, but I'd imagine it could be done with several 555's
>to generate tones, and some sort of multiplexing circuit to
>cycle between the N inputs.

Two channels can be done easily with a stereo recorder.  For more than
that, I'd probably just digitize the data and pipe it through a modem
chip to the recorder.  A single chip micro could insert sync flags in
the data stream and maybe even insert some redundancy or ECC code.

I have LOTS of plans in this area for PE.

BTW, I've gotten my hands on a RacePack computer.  This is the data
acquisition boxes seen on all the top fuel, top alcohol and ProStock
NHRA cars.  This is the unit that prints out those nifty little stripcharts
the racers like to look at while on TV.  Very expensive boxes.
I'll be doing a pretty  detailed analysis in a future issue of the mag.
My first blush is that I'm extremely underwhelmed.  As an example, the
exhaust temperature thermocouples are not cold junction compensated as
best I can tell and the cold junction is actually in a box that mounts
next to the engine.  Not exactly isothermal.  One strip I looked at in
detail came from a ProStock motor.  Temperatures varied about 200 degrees.
I asked and was told, "Oh it's always been like that. WE don't pay any
attention."

Oh, while I'm at it, this RacePack does not have a printer with it.
Does anyone recognize the brand of printer used?  I think it is a
Citizen but I'm not sure.  Guess I could call 'em huh?

There are bountiful market opportunities here, guys!  :-)

John

From: John De Armond
X-Source: The Hotrod Mailing list
Date: Aug 1992
Subject: Re: flow bench

> Hmmm to you too.  Of course, the readings you would get wouldn't
>convert to anyone else's machine, and you'd have to decide what to do
>with the camshaft...  but you might get a better idea of how live flow
>changes and how fast the velocity increases in relation to valve lift.
>Hmmm.... it'd also be useful for checking out how cam profiles affect
>breathing.
>
> On second thought, I'll hang on to that block.  Sounds like it just
>might be useful.  And since Ford and Chevy bore spacing is
>near-as-dammit close, I could even drill the deck to take either type of
>head.  And with a little work, I could see what differences changes in
>rod ratio would make.

You know, we've both missed the obvious solution to this problem.  Simply
leave a couple of cylinders on one bank functional.  put a small single
barrel carb on it and let the engine motor itself.   The other bank
could be totally isolated as far as cooling, intake, exhaust and even
camshaft goes (some work required, of course, to drive the experimental
cam from the rear.)  50 lbs or so of mass welded to an old flywheel
and then balanced would keep things smooth.  This scheme would
let you motor the experimental cylinders to any speed the engine is
capable of in the first place.

> John, we could get a lot more information off this than we could with a
>simple steady-state flow bench.   Do you have any ideas on what sort of
>instrumentation might be useful?

Yes.  I'm still looking around for a cheap multi-channel data ack board
for the PC.  Given a decent ability to record several analog channels,
I'd want to monitor intake and exhaust instantaneous pressure using
piezo pressure transmitters (under $50 ea), crank angle, chamber
pressure, chamber temperature and perhaps a channel or two of pressure on
the end of a handheld probe.

Hmmm. I have a couple of Z engines that I used a turbo to blow one or
more pistons out of.  Hmmm.  :-)

>-> >Jenkins and Yunick claim it's not important, so I'm going to ignore
>-> Far be it for me to take on such luminaries as above :-)
>
> Methinks they might be full of waste material.  I think of a new and
>interesting use for a motored engine every few minutes.  Heck, think of
>what you could do with a torque sensor between it and the driving
>motor... how much power does that oil pump, water pump, or alternator
>take?  Are the "power pulley" people inflating their claims by 500% or
>800%?

You know, the SAE literature has been full of articles in the last year or
so on the use of magneto-strictive torque sensors to be used with the
next generation engine management systems.  These will be fly by wire
and will set engine parameters for best available torque and lowest
specific fuel consumption necessary to deliver the acceleration requested
by the gas pedal encoder.  These sensors are nothing more than
cruciform ferrite cores with quadrature coils, all mounted in the rear main
bearing cap.  The crank is excited with low frequency RF.  Torque
varies the magnetic path and changes how much voltage is induced in
the pickup coil.

> How about a clear piston top and some colored smoke in the inlet tract?
>
> Yahooo!  Ba-aby...   three or four months, and I'm gonna have me a toy!

I've made pyrex two stroke heads and photographed actual combustion so
I know how valuable this can be.  I also have some literature here from
NGK that features, among other goodies, spark plugs equipped with
fiber optic light pipes :-)  Possibilities abound.


This is gonna be fun!

John

From: John De Armond
X-Source: The Hotrod Mailing list
Date: Aug 1992
Subject: Re: Telemetry

>> ...Kenny Bernstein was the first to use the racepak.
>
>Well, close, but not quite. The racepak was developed and marketed
>to a few insiders by someone else (his name escapes me at the moment,
>but I have correspondence with pricing, etc. somewhere). Ray Wise
>or something like that.

Think the guy's name is King.  The label says "Distributed by King
Racing Components" and also bears the label of Competition Systems
Inc.  KRC's phone number is 714 998 3830.  I'll know more after
I call them.

Based on what I've been told about the pricing on this thing, I have
to say a small condemnation to Bernstein.  I hate to see racers
gouging racers.  At the same time, I'm glad to see presence of ANYTHING
available to the racer.  A more general purpose unit is made by RPM
Inc and only costs about $1800.

>Anyway, Kenny glommed onto the technology early on, and with help
>from his newly acquired crew chief, Dale Armstrong, added a few
>wrinkles of his own, hammered out a deal to sell the package and the
>rest (as they say) is history.
>
>The accuracy of the data acquisition probably has room for a lot
>of improvement, as John mentioned. But the area that is wide open
>for a savvy entrepreneur is in data interpretation. Using the flat
>file database of time versus measurements along with a knowledge
>engineered program to interpret results is still in its infancy.

Absolutely.  I'm waiting for someone to take advantage of this wide
open gap and fill it. Then write me an article, of course :-)

>I'd venture to say that 90% of the guys using the racepak don't
>know how to determine but a fraction of the available information.
>The other 10% sure as heck aren't telling either.

I'd venture to say the ONLY data remotely close to being accurate are
from the RPM inputs.  Typically engine, clutch and driveshaft
RPM is recorded.  At least they can set the clutch from this.  The other
stuff is GIGO.  particularly the G sensor.  Useful perhaps for relative
info but certainly not calibrated.

john

From: John De Armond
X-Source: The Hotrod Mailing list
Date: Aug 1992
Subject: Re: flow bench

>-> You know, we've both missed the obvious solution to this problem.
>-> Simply leave a couple of cylinders on one bank functional.  put a
>-> small single barrel carb on it and let the engine motor itself.
>
> Er.  Um.  Nope, I don't like that idea.  Every time I removed the test
>head I'd have oil and water drooling everywhere, there'd be interference
>when testing intake manifolds...  however, I *just happen* to have a
>freshly rebuilt Kent 1600 around somewhere which ought to be fine as a
>power source.  I'd have to buy an electric motor, and free is a quality
>overlooking many faults.

I can see the intake manifold problem but my idea was to separate the banks
as far as water and lube goes.  Some oil would be necessary but obviously
not as much as on the running side.  My in-line 6 will be MUCH easier
in that regard.

Oh, and 10-4 on the free!  A vastly important quality.

John

From: emory!sybase.com!joe (Joseph Weinstein)
X-Source: The Hotrod Mailing list
Date: May 1993
Subject: AN in-car HP/torque meter...
X-Sequence: 5256

I had an idea for an in-car power meter:

Place a magnetic sensor on the drive shaft near the front, which spins
past a detector. Another sensor and detector is at the other end of the
shaft.  The two signals are synchronized ie:  the time between the two
signals when there is no torque on the drive shaft is noted (a function
of RPM).

Now, this signal will change at a given shaft RPM. in linear relation to
the torque through the drive shaft. An engineering text would provide the
amount of deflection for the shaft for a given torque, or better, the system
could be calibrated one time on a dyno.

Joe

Joe Weinstein   joe@sybase.com   Sybase 1650 65th st.  Emeryville Ca 94608
        					       510-596-3620
[This has been done and commercial sensors are available.  BUT it is not
as accurate as inertial instruments like the Dynofact simply because the
inertial measurement neatly factors out wheel spin.  torque is directly
translatable to forward thrust or acceleration.  Think of torque as
thrust around a circle.  Given the thrust (force) and the weight of the
vehicle, the power (horsepower for example) necessary to generate this
thrust is a simple physics calculation.  JGD]

From: "Ciciora Steve" <emory!aztec.al.bldrdoc.gov!sciciora>
X-Source: The Hotrod Mailing list
Date: Feb 1994
Subject: RE: Horsepower measurements
X-Sequence: 7630

-Speaking of dynos, here is the coast-down method of measuring HP.
-You really only need a stopwatch and an accurate speedometer but I use a
-radar gun for more precision in the speed measurement.
...
Wow, I always thought that there was an engine dyno built into my car.  I
always thought that if you measured engine rpm and car position information
fast enough, with a little knowlege (weight, friction, etc) you could get
torque curves and hp figures.  Anyone have any guesses on how fast is fast
enough?  RPM and MPH figures 10 times a second?  What are good ways to measure
MPH?  How good would it be to measure a pulse /rev of a non-driven tire?  I had
always hoped that GPS would be good enough for this application (if you knew
X,Y,Z data a level track would not be necessary) but they need to come down an
order of magnitude in price and go up a few orders of magnitude in performance
($50 vs $500 and 10 pts / second vs one point every 10 seconds).

Any one else interested in persuing a project such as this?

- Steven Ciciora

[I have a project underway here for a future magazine article to do this
with doppler radar.  The whole setup can be built for under $200,
maybe even $100 if you're good at scrounging microwave motion
detectors as are used for door openers.  This approach measures speed
vs time directly which IMHO, is easier to deal with than the
accelerometer-based products such as the Banks Dynofact and the Vericoms.
Accelerometer-based instruments must integrate acceleration to compute
speed plus it must compensate for front end lift which "looks" like
acceleration to an accelerometer.

Another way to do the measurement if your car does not spin its tires
is to attach a couple of magnets to the driveshaft, position a coil
nearby, slice and buffer the pulses with a comparator and pipe the
pulses to the *ACK line on a PC parallel port.  This line can be
set to generate an interrupt on each toggle and a reasonably fast
386 laptop can catch a few thousand of these a second.  Using the
high res clock routine from my fuel injection flow bench software
or elsewhere, one can read the clock and thus measure the period
of the driveshaft pulses and compute the frequency.  Since 8000
RPM on the driveshaft with two magnets is only 267 pulses per
second, the parallel port can easily handle the job.  One can write
some software to collect a dataset of frequency vs time and use
Lotus or whatever to reduce the data.

Another source of pulses is the reed switch interruptor used by
taxi-cab meters.  Check your yellow pages under "speedometer shops"
for a source.  This thing goes in the speedo cable.  This can be
wired directly to the *ACK line and ground.

Finally, if your car has an electronic speedo or tranny, there is likely
already a speed transducer you can tap into.

Finally, if you have a video recorder that can freeze frame and single
step, you can simply video tape your speedo or tach (or even better,
digital aftermarket equivalents) and get the timing from the
known framerate of the video system.  If you're a real cheap mutha like
me, don't have a cam-corder and don't want to buy one, you can
get a cheap VCR from the pawn shop for $50 or so and a security camera
for <$100 and run the whole thing from a 60 hz inverter in your car.
Comp-USA has a 200 watt inverter on sale right now for $59.

Though a lot of people poo-poo the technique in this era of megabuck
data acquisition systems but it is hard to beat the data bandwidth of a
video camera aimed at a bank of gauges.  If you want a very low light,
infrared sensitive CCD camera about 1/4 the size of a 3.5" floppy, you
can order one from an outfit called Personal Computer Tools.  800 767
6728.  Price is $210 for the 324X246 pixel camera and $265 for the high
res 512X492 pixel (240X240 and 350X350 lines respectively) I've seen the
low-res camera at hamfests for $100 but this one is at the tip of your
800 dialer :-) The high res camera will push the limit of even a high
quality VCR.  These little cameras are REAL handy for a bunch of tasks.
If you want to see if your accelerator pump is running out of juice too
fast and is causing the bog, just put the camera over the carb mouth.
Wondering if your suspension is hiting the limits?  Put the camera in
the wheel well inside a baggy.  Wonder if your engine is twisting on its
mounts and hitting the chassis and making a noise?  Aim the camera at
the suspect and take the car for a spin.  Wondering if the rear end is
twisting and making the drive shaft vibrate?  Mount the camera in the
tunnel and take a look.  JGD]

Index Home About Blog