So I have an 85 GT, and lately I noticed the tachometer has been reading funny. For example, my rev limiter is set at 6200 RPM, but I drove the engine well into the red zone (indicated on the tach) without any hiccups.

Hmmm, seems like the tach is reading high. When I hooked up my laptop to the ECU, the RPM of the ECU disagreed with the RPM of the tachometer. OK, time to add tach recalibration to the never-ending to-do list on my Fiero.

Thankfully, there exists the Oliver Scholz writeup for reference:
http://www.fieros.de/en/articles/tach.html

However, adjusting just the resistor value may only be a temporary fix. If the tachometer is off because of a drifting capacitor, then the tachometer will continue to drift even after you adjust the resistor value.
This opinion was expressed on the Camaro forum, and it makes sense, I guess:
http://www.thirdgen.org/for...r-waaaaayyy-off.html

So now, how to purchase a proper capacitor? Polycarbonate capacitors (as originally specified) are now unobtainium. You can get polypropylene though.

Time to read the LM1819 datasheet to understand what's going on.

My particular Fiero is set up with 45 distributor degrees of dwell at high RPM, falling to a constant 3 ms dwell below 5000 RPM. I assume that stock Fieros are doing something similar with the dwell.

For a Fiero, use the "constant duty cycle equations" in the LM1819 datasheet.

Here are the key words:

quote National Semiconductor LM1819 Air-Core Meter Driver: constant input duty cycle (C1 acts as a differentiating capacitor)

Bingo, that means the tach will work with the edges of the coil signal; the wacky dwell time won't matter.

Here is the 10 nF polypropylene capacitor that I will use for "C1":
http://www.digikey.com/prod...04/BC3087-ND/5393628
There ain't too much choice for axial-leaded (for vibration resistance) parts anymore...
Polypropylene has an annoying (but linear) -200 ppm/°C temperature coefficient for its capacitance value.

Then, the 100 kΩ multi-turn trimmer resistor:
http:/ /www.digikey.com/prod...=0&fid=0&pageSize=25
However, this trimmer has a +/- 100 ppm/°C temperature stability. Not too good...

So we install a high-quality 100 kΩ fixed resistor in series with the trimmer:
http://www.digikey.com/prod...TF100KDCT-ND/2273750
+/- 5 ppm/°C. Awesome!
Essentially, we want the temperature stability of the circuit to be dominated by this part, and not the trimmer.

The tachometer's sensitivity is really based on the RC time constant of the resistor+capacitor (C1 and R2) combination. So, the capacitor can drift, as long as the resistor drifts in the opposite way. So then I add a 10 kΩ PTC resistor in series with the two others:
http://www.digikey.com/prod...J/A105614-ND/1149318
The PTC resistor will need to be physically installed right next to C1; so they remain at the same temperature.

While we're at it, lets install a new 0.68 µF ripple capacitor "C2":
http://www.digikey.com/prod...1001-2310-ND/2794421
X7R, good enough for this duty I guess. Axial-leaded, yay!

I'm going to order these parts and install them next weekend. I still have to figure some details regarding the wiring, and how I'm going to hold the PTC next to C1.

The trimmer body is threaded, so it can be installed with a nut such that it can be adjusted from the outside once the cluster is all buttoned up in the car.

I will use my MegaSquirt laptop RPM reading to calibrate the tach. Probably at 5500 RPM or so; this is mostly so I can get my drag-racing upshifts right.

Another way to calibrate the tach would be to drive at a known road speed, and then knowing the gear ratios, calculate the theoretical engine RPM.

[This message has been edited by pmbrunelle (edited 09-05-2016).]

The LM1819 is discontinued. There are some still on ebay however.

From what I am hearing, the resistors on the resistor chip are changing value, and that is what is causing the tach's to be off.

At least tachs do not have the rarity of fastback quarter windows... for now.

Based on the datasheet equations, I calculated that my tach should have a 260 kΩ resistor. However, when I measured the resistor, it measured 340 kΩ. So that's definately an issue. However, once I'm in this far, it just makes sense to rebuild everything else (critical capacitors as well).

And of course, once I click to "Add to Cart" on digikey, they won't sell me the X7R axial capacitor I wanted. I think it's "military" or something, and Canadians aren't cool enough to buy such things

I'll use this polypropylene film capacitor for ripple reduction instead:
http://www.digikey.com/prod...4HL/P12203-ND/656361

Since it's so huge, I will have to secure it to the PCB with electronics-safe RTV; otherwise, I am concerned that the body of the capacitor could vibrate, and then the leads could break off or something.

So what are you doing for a board?

I will be replacing components on my Fiero's original board.

As long as there's nothing wrong with the core, there's no need to make an entire brand-new assembly. A good core can be rebuilt.

I've been saying is likely a bad cap problem but no one bothers and just "fix" the resistor.
First fixing C1 cap then try running it.

"Based on the datasheet equations, I calculated that my tach should have a 260 kΩ resistor. However, when I measured the resistor, it measured 340 kΩ. So that's definately an issue. However, once I'm in this far, it just makes sense to rebuild everything else (critical capacitors as well)."
Two RC for total angle and "Red Line" for the tach display.
Might be change for resistor because total angle, Vin or Rreg is more/less then you think.
Tach w/ oil pressure ~240° but v6 tach only is maybe ~300°
I think Vin is input volts from the ignition after the tach filter etc. (IOW Vac input at IC Pin 10)

Post a link for datasheet used...
I have nat semi sheet dated feb 95 w/o "constant duty cycle equations" label on it. (Think is same but just different labels based on tach app on page 4)
http://pdf.datasheetcatalog...nductor/DS005263.PDF

most of GM 4cyl tach is from figure 1 of this sheet. Part not on the board is HEI tach filter to shape tach signal and kill HEI EMI/RFI. (DIS have whatever filter build into the module.)
Many run the car and short the plug so tach will work but w/o a good filter the tach wire will act as antenna causing problems.

Temperature can kill them over time.
Parked car heat will kill kids but can kill any electronics too.
Electronic Generate heat then drive at night w/ bright dash then bulbs can cook the tach etc.

------------------
Dr. Ian Malcolm: Yeah, but your scientists were so preoccupied with whether or not they could, they didn't stop to think if they should.
(Jurassic Park)

The Ogre's Fiero Cave

And if you do not crave a fix of the original part, but would like to try something new..
Link to video of Fiero tach in development
Crystal accuracy, 4/6/8 selection, and a tach memory function if you want to add a button.

And don't "fix" the resistor to make it thermal stable for now.
Most are close enough that ±50RPM, even ±100RPM, on the needle is meaningless to most owners. Many aftermarket tachs are off too.
(Many don't look at the dash at all to shift a manual trans. If you drive and aren't deaf many go by engine sound to shift. I had several stick cars w/o tachs and long before any ECM shift lights.)

Match the ECM/PCM is good but remember Data stream updates numbers 1 to 3 times per second vs. tach reads RPM as fast as needle can move. Try using a "shop" tach to compare.

[This message has been edited by theogre (edited 09-06-2016).]

If you play some kind of instrument and can tune by ear, you can tune it to match the engine sound when it's idling and check the frequency with an instrument tuner. Then you can simply convert the frequency to RPMs.

Hell, maybe you can even download a tuner app for your phone and if it's good enough you should be able to catch that just with your phone near the engine.

EDIT: It seems there are some apps that exist for this purpose, check out "acoustic tachometer" or search for something with sound, rpm, tach keywords. Never tried one though but I think it works with the same principle as an acoustic musical instrument tuner.

[This message has been edited by cebix (edited 09-06-2016).]

I found my datasheet here:
http://www.ti.com/lit/ds/symlink/lm1819.pdf

Looks the same as yours, Ogre. I am mostly looking at FIGURE 1 as my main schematic reference.

There seems to be a typo in the datasheet. In the constant duty equations, it seems that R_Z should be R2.

Unfortunately for a test, I can't just replace just the capacitor. Originally, there was a 0.068 µF (C1) capacitor on my tach board. However, I ordered a 0.010 µF (C1) replacement, which was OK since I was going to change R2 as well.

In constant duty cycle mode, there are two important RC combos:

R2*C1
The RC time constant here directly determines the needle sensitivity (degrees/Hz or degrees/RPM). The needle sensitivity drifts 1:1 with the this time constant.

R2*C2
The RC time constant here determines the ripple of the needle, and the response time of the tachometer to RPM changes. Overall not too critical, but if C2 gets leaky, it will appear as a parallel resistor to R2... I think this time constant would have to be way out to visually notice any problems.

R1 and R_REG can drift somewhat; as long as they stay within reasonable limits, the tachometer output will not drift in unison.

There is some evidence that R1 and R_REG are indeed 4.7 kΩ and 3.3 kΩ as per FIGURE 1; I measured their total series resistance on my tach and I got 7.5 kΩ.

With a polypropylene capacitor for C1, it takes a 40°C temperature change to cause an error of 50 RPM at redline. Is that likely to occur? What about with the light bulbs on? Since this is the overkill thread, I will just compensate the capacitor and be done with it. I ain't a penny-pinching OEM in my free time! Actually, while I sort out this tach business, I continue to daily-drive my Fiero without a tach...

With my Megasquirt ECU, I have a simulated analog tach on my laptop's screen. The update rate is fast enough that visually, that tach looks smooth. Its accuracy is based on the crystal oscillator in my ECU, so I expect it to work OK.

Nice project Chris. Do you have a build thread?

As for a music tuner/app, I wouldn't recommend calibration at idle if you want to be precise near redline. Also, I'm not sure the frequency response of a smartphone's microphone would work down low. At 5000 RPM though, you'll get a a fundamental of 250 Hz. Should be acceptable; you could bump up the RPM a bit and tune to middle C.

[This message has been edited by pmbrunelle (edited 09-07-2016).]

I just installed the parts on my tachometer board. Now I just need to wait for the RTV to cure (wait one day) before testing in my car.

So I got the tachometer back into the car today, and then I calibrated its gain.

With the car stopped in neutral, I revved the engine to a steady 5500 RPM as indicated by the data from my ECU (this is my reference).

With the RPM steady (not too easy in neutral), I used a screwdriver to adjust the potentiometer until the tachometer displayed 5500 RPM as well. The multi-turn pot offered a sufficiently fine adjustment.

Overall the linearity seems OK. After calibrating at 5500 RPM, I compared the tach with my ECU at 5000, 4000, 3000 RPM and so on. The two track each other pretty well. At idle, the tach reads about 100 RPM too low. Anyway, it's hard to be more precise than that, since my foot was not rock-steady, then there's the parallax error, not to mention the LM1819 (or equivalent) which only guarantees 2% linearity...

The high-RPM accuracy was the most important to me, and I got it. Overall I am satisfied with this tachometer job, for now.

Once I was done with the calibration, I installed the top instrument cluster cover.

As for long-term stability, time will tell. I will keep an eye on it.

[This message has been edited by pmbrunelle (edited 09-12-2016).]