The tach in my '88 V6 is reading incorrectly. At idle, it reads about 1100 - 1200 and the ALDL datastream says 900...and it sounds like it's running at 900.

So I pulled the tach gauge, got a function generator and when I supply a 60Hz signal, it reads 1500RPM. According to the GM service manual, it should read 1350. According to Oliver Scholz, it should read 1200. (see http://www.fieros.de/en/articles/tach.html)

Which is correct? Also, if it reads n RPM at 60Hz, is it safe to say it should read n * 2 at 120Hz (is it linear)?

One more question...There are two ways to fine tune the gauge; changing the capacitor and changing the resistor. Any thoughts on this?

Edit: Fixed link to Oliver's site

[This message has been edited by mize (edited 03-07-2006).]

as far as which one to change .... what about throwing a pot in there of the correct range .... might make adjusting in the future a whole lot easier.

Rich

Actually I have a pot in there right now. But when I adjust it for 60Hz, should I set it for 1350 (as stated in the GM manual) or 1200 as Oliver Scholz says? Also, this is why I was asking if the reading should be linear. If I set it for 1200 at 60 Hz, should I expect to see 2400 at 120 Hz?

I believe the service manual is correct. Simple math would seem to indicate that you have 3600 pulses per minute and 3 pules per revolution on a V6 or 1200 rpm. But there must be something unique about the Vin 9 engine, or why else would the manual say 1350 RPM?

And yes it's linear.

[This message has been edited by jscott1 (edited 03-07-2006).]

quote Originally posted by mize: The tach in my '88 V6 is reading incorrectly. At idle, it reads about 1100 - 1200 and the ALDL datastream says 900...and it sounds like it's running at 900.

You do know when aldl with 10K resisitor is running --it bumps your (cars) tach up around 200rpms?? From what aldl is actually reading....

I've had decent exposure to these tachs. The stock Fiero tachs are not the most accurate instruments. Heck, I've got tachs out of 30 year old cars that are a lot more accurate.

I highly recommend you NOT start adding potentiometers, etc., in an attempt to fix the problem. You're only adding a hack to something that isn't working right to begin with. I won't proclaim to be an expert with electronics, but I know enough. Capacitors & resistors do break down over time. Resistors normally start to go "high" (resist more current) and capacitors will sometimes simply fail to hold any electrical current. A tach is a fairly precise instrument, and when old electronics start operating outside of spec, the result is an inaccurate tach. When you're dealing with a Fiero that is roughly 20 years of age, it wouldn't surprise me if you found a resistor or two that was no longer in spec.

My advice? If you know about electronics, you can simply replace some of the resistors and capacitors. Those will be your most likely suspects. These electrical components are not expensive, and fairly easy to obtain. It won't take too long to swap them out, either. But if you don't want to mess with electronics, try to find a tach that is a little more accurate than what you have.

Hope this helps...

quote Originally posted by Lambykin: I highly recommend you NOT start adding potentiometers, etc., in an attempt to fix the problem.Hope this helps...

Well Lambykin, I'm adding potentiometers so that tachs designed for V6 will work with V8 and vice versa. Not the most elegant solution, but the factory sets the tach based on resistors and capacitors, so as you say those things drift over time. The beauty of the potentiometer is that if it drifts you simple readjust it.

quote Originally posted by 3800superfast: You do know when aldl with 10K resisitor is running --it bumps your (cars) tach up around 200rpms?? From what aldl is actually reading....

Sure, but the tach gauge and the ALDL datastream should still read the same. It doesn't matter it gets bumped up...the two readings should be close.

quote Originally posted by mize: Sure, but the tach gauge and the ALDL datastream should still read the same. It doesn't matter it gets bumped up...the two readings should be close.

This would be good to start another post on---ask how many people that use winaldl if there *real time tach* reads the same as the aldl data stream----my guess would be no--they will probally report a 200rpm or so difference --like the same thing you mentioned at the begining of the post...

quote Originally posted by 3800superfast: This would be good to start another post on---ask how many people that use winaldl if there *real time tach* reads the same as the aldl data stream----my guess would be no--they will probally report a 200rpm or so difference --like the same thing you mentioned at the begining of the post...

Good idea!! I just created another thread.

But again, that's not the main topic I am asking about. I have everything apart and have a chance to get the tach as close as possible. So my question still stands (with the exception of 1 reply so far): should I calibrate for 1200 or 1350 and should it read twice that if I double the frequency on the pulse generator.

I am asking about the double frequency because it doesn't matter what I adjust the tach to at 60Hz. I never see twice that amount at 120Hz. So I am wondering if something else is up here...but I am also testing with two tach gauges and seeing similar results.

I am not convinced components (especially caps and resistors) will drift off their original values over 20 years. These components are in a calm environment unlike the engine compartment. I have used test tools from over 30 years ago that still work perfectly. Maybe the question of how many see differences between ALDL and the dash gauge is more important than I thought.

Sorry, I don't know a lot about that program to answer your original question. In theory, doubling the frequency should give you double the tach reading.

As for not being convinced that the values of electronic components changes over time - pull a couple of the components on a tach that is known to be very poorly calibrated (and not modified). Measure them with your favourite scope, digital meter, or capacitance tester in the case of a capacitor. You're going to find variances from the values they originally were. A little here, a little there, and next thing you know, your guage is not so precise. Things like transistors are a lot more reliable. Oh ya, it doesn't matter what environment they are in. Electrolytic capacitors, for example, are an excellent example of a component that can, and will, go bad if not used periodically. Thankfully, I don't believe there are any of those on the tach board.

To answer your most recent question - not having a tach filter might affect the tach slightly, but it should be darn close without one. The tach is still receiving the same pulse at the same frequency regardless of the filter (which, by the way, is just another capacitor - and you weren't convinced electronic components just break down). I was running a tach filter, then it went south. Tach seemed just as accurate (or inaccurate?) without it.

You can't assume the tach is a precision instrument. It's not. An RPM error of 200+ between two tachs is not unusual. If you want them to be dead on, you have to adjust full scale at full scale and even then they will likely only appear dead on at full scale. At lower readings they will vary around.

They just aren't that good.

Umm, thanks for the replies. But (like most threads here) it keeps drifting off topic with what people think the topic SHOULD be.

The questions were fairly simple. I never said help me make my tach more precise than it was ever meant to be. I just said it was off by about 200 at idle. At higher speeds, it is way off. It was just enough to be annoying. Since I had the whole thing apart, I wanted to get it as close as possible...within reason. But I assumed people would understand that.

I don't want to know about how components degrade...or don't. Yes, it may be the cause of the problem. But the solution is not: "...the components are old...and that gauge is a piece of junk." That does not help. A better reply would have been, "...you could try substituting that cap with a newer 0.0082 uF one" or something like that.

I know people must try to prove their perceived intelligence. I really don't care about that. I just want some hints from those that could help work toward a solution.

Thanks for all the replies.

Hey chief we tried to help. Since they aren't that great in the first place about all you can do is adjust full scale or select cap values until it reads whatever you perceive as being good enough. The full scale adjustment is described in a billion different threads on the forum under " V8 tach conversion" and the like. You already know the cap location.

I think the symptoms you are describing is fact-overload. We do try to tone down our perceived intelligence as needed. It just didn't seem to be needed in this thread.

[This message has been edited by TK (edited 03-08-2006).]

Thanks Ace!! That was more helpful.

Since I am not doing a V8 conversion, I would not have looked for the info under that heading. The only thing I found regarding fine tuning a stock setup was adding the pot.

...now we are resorting to name calling. I love this forum!

Also, I said thanks for the replies. I recognized the attempted help.

Hi Mize,

It's hard to figure out what level of detail people want or need. There isn't a single right answer to your question. I'm not trying to offend anyone and I appologize if you took it that way.

The tachs aren't linear enough to get them to read right across the whole scale. If you get one spot right another spot is off. Full scale is about the best place to shoot for and accept the error at the lower readings.

Regards,

Terry

Edit: I removed my flipant comment.

[This message has been edited by TK (edited 03-08-2006).]

quote Originally posted by mize: I was just looking at the gauge I purchased and ... I am seeing 1350 at 60Hz. At 120Hz it reads ~2525, at 180Hz it reads 3750 and at 240Hz it reads 5000.

First, let me say that I agree with Oliver Scholz that, absent instrument error, 60 Hz into a V6 tach should produce a 1200 rpm reading, 120 Hz -> 2400 rpm, 180 Hz -> 3600 rpm, 240 Hz -> 4800 rpm, etc.

However, there are three categories of errors in indicator instruments like tachs:

1) Offset error ... e.g. the tach reads 200 rpm high all across the scale. You can correct this electrically, but on the Fiero tach it's probably easier to accomplish this by simply repositioning the pointer on the shaft.

2) Scale error ... e.g. the tach reads correctly at very low rpm, but the error gets bigger as the rpm increases. This is the factor that changing the resistor and/or capacitor affects. A variable resistor (like the original laser-trimmed fixed resistor) allows you to precisely adjust scale error to near zero.

3) Linearity error ... e.g. the tach reads correctly at some points in its rpm range, but has significant errors at other rpms. Such errors are often inherent in the gauge mechanism, and there is little you can do about them other than to adjust offset and scale to minimize the average error across the full rpm range. It's possible that the GM spec (assuming it is correct) of 60 Hz -> 1350 rpm rather than the theoretical 1200 rpm is most likely due to linearity error.

One other thing to consider is where you want (or need) the tach to be most accurate. On most gauges, the calibration point is at full scale. On a tachometer, the redline is probably the best possible choice for the calibration point. Think about it. A small error at low rpm will not have any adverse effects, but an error at near red line rpm can result in broken parts.

Given all this, I think the main error in tach in your example is an ~175 rpm offset error. The original data you provided has a minimum error of 125 rpm, an average error of ~150 rpm, and a maximum error of 200 rpm:

60 Hz -> 1350 rpm, 120 Hz -> 2525 rpm, 180 Hz -> 3750 rpm, 240 Hz -> 5000 rpm

If you were to set the pointer to indicate 175 rpm lower, you would now see a minimum error of 25 rpm, an average error of ~20 rpm, and a maximum error of 50 rpm:

60 Hz -> 1175 rpm, 120 Hz -> 2350 rpm, 180 Hz -> 3575 rpm, 240 Hz -> 4825

If you were to set the pointer 200 rpm lower, you would see a minimum error of 0 rpm (at or near redline), an average error of ~50 rpm, and a maximum error of 75 rpm:

60 Hz -> 1150 rpm, 120 Hz -> 2325 rpm, 180 Hz -> 3550 rpm, 240 Hz -> 4800

This is quite an improvement in accuracy, obtained by nothing more than repositioning the tach pointer!

As a final note, a PC sound card makes a nice high-precision frequency generator for calibration tests like this. I use a simple UNIX program that I wrote long ago that lets you specify the exact frequency, but here is a chart of musical notes vs. frequency that you could use with simple "piano keyboard" software:

All you really have to remember is that the scale factors for distributor-ignition, 4-cycle motors are:

6-cylinder: 20 rpm per Hz
4-cylinder: 30 rpm per Hz

Example: For a 6-cylinder engine, a note of Middle C (261.6 Hz) should produce a tach reading of 5,232 rpm.

[This message has been edited by Marvin McInnis (edited 03-09-2006).]

quote Originally posted by Marvin McInnis: First, let me say that I agree with Oliver Scholz that, absent instrument error, 60 Hz into a V6 tach should produce a 1200 rpm reading, 120 Hz -> 2400 rpm, 180 Hz -> 3600 rpm, 240 Hz -> 4800 rpm, etc.

Thanks for the great post!!!

I have been working on the details outlined in the GM service manual that indicates at 60Hz, the tach should read 1350. This idea throws off your idea a bit. Are you saying I shouldn't use the specs in the manual?

Thanks again for the post.

Steve

quote Originally posted by Lambykin: As for not being convinced that the values of electronic components changes over time - pull a couple of the components on a tach that is known to be very poorly calibrated (and not modified). Measure them with your favourite scope, digital meter, or capacitance tester in the case of a capacitor. You're going to find variances from the values they originally were. A little here, a little there, and next thing you know, your guage is not so precise. Things like transistors are a lot more reliable. Oh ya, it doesn't matter what environment they are in....

I really have to disagree with you on several points here:

1) You say: "... pull a couple of the components on a tach ... [and] measure them ...." How can you possibly know what the actual (rather than nominal) values were in the first place? Remember that the tolerance of most brand-new capacitors is +/-20% of nominal to start with, while resistors are commonly +/-20%, +/-10%, or +/-5%, and that both capacitors and resistors change value (predictably and reversibly) with changes in temperature.

2) If operated well within their ratings (temperature, voltage, ambient environment, etc.), most modern passive electrical components just do not change value significantly with age. Electrolytic capacitors, as you noted, are an exception to this generalization ... but the actual working capacitance of aluminum electrolytics is highly dependent upon the operating voltage, and even at rated voltage the manufacturing tolerance of many aluminum electrolytics is +100%/-50% to start with.

3) I agree that transistors are generally reliable when operated well within their ratings, but there is usually a tremendous variability from one transistor to the next. For example, according to most specification sheets the Beta (common-emitter gain) of a brand-new 2N2222 small-signal transistor can vary from around 20 to over 100! (A smart designer will design for the worst-case gain value of 20.) Note also that the operating characteristics of transistors and other semiconductor devices often vary widely (and predictably) with temperature. Analog transistor circuit designs generally rely upon the passive components to establish the characteristics of the overall circuit and to accommodate the wide range of transistor properties.

[This message has been edited by Marvin McInnis (edited 03-10-2006).]

I will toss my 2 cents in here. I have done many tach conversions from V6 to V8. I always check the tach through its entire range before I start to work on it. I have yet to see one that reads 1350 when I input 60 Hz to it. They all have read (unless it was broken when I received it) 1200 plus or minus a few RPMs.

Jack

quote Originally posted by mize: Are you saying I shouldn't use the specs in the manual?

In a word, YES ... as long as you understand why!

I looked at my '88 GM/Helm Factory Service Manual and the only reference to 1350 rpm I could find was on page 8A-80-15, paragraph P, which deals with troubleshooting a "no tach" situation. This is simply a go/no-go test, not a calibration test. It is interesting to note, however, that the rpm value given for a 4-cylinder engine is exactly the "theoretical" 1800 rpm! (The corresponding value for a six-cylinder motor would be 1200 rpm.) I have to conclude that the 1350 rpm value for the V6 is either 1) a typographical error (most likely), or 2) intended to allow for some expected linearity error in the tach mechanism.

Regardless, if it were me I would first reposition the tach pointer to minimize offset error and then, if necessary, calibrate the tach for zero scale error at redline. Unfortunately, in practice you will probably experience some interaction between these two adjustments. If you can get within +/-100 rpm across the entire range of the tach you will be doing pretty good; in fact, parallax (which results from viewing the tach at an angle) will then probably be a bigger source of error than the tach calibration itself.

One final thought: With the tach installed in the car, you can use a diagnostic tool (e.g. AutoXray) or WinALDL to compare the tach reading with the rpm reported by the ECM. Being digital, the ECM reading will almost always be more accurate than the analog tach.

[This message has been edited by Marvin McInnis (edited 03-09-2006).]

quote Originally posted by Marvin McInnis: If you can get within +/- 50 rpm across the entire range of the tach you will be doing pretty good; in fact, parallax (from viewing the tach at different angles) will then probably be a bigger source of error than the tach calibration itself.

I absolutely agree. I just wanted to try to get it close to what it was back in 1988 when it was new. Not better than it was.

quote Originally posted by Marvin McInnis: One final thought: With the tach installed in the car, use a diagnostic tool (e.g. AutoXray) or WinALDL to compare the tach reading with the rpm reported by the ECM. Being digital, the ECM reading will almost always be more accurate than the analog tach.

Using WinALDL is where I got the original 900 RPM number when the gauge was showing 1100 - 1200.

I was just about to reinstall the gauge pod with the tach showing 1350 at 60Hz. Based on your comments, I think I'll temporarily install the cables, compare the readings to WinALDL. If the gauge and ALDL differ, I will implement your idea of moving the needle and/or fine tuning with the variable resistor.

Thanks again for the replies!!

Well I change my vote.... I think a V6 should read the theoretical 1200.

When I adjust my tachs I'm going to set them to 1200 at 60 HZ. If I ever get a signal generator I'll set them perfect at the redline, which is where you really want the accuracy.

For those who're still interested, I have posted a chart of musical notes vs. frequency back in my original post in this thread. With this information you can use a PC sound card as a high-precision frequency generator for tachometer and speedometer calibration.

If you want to use a PC as a tone generator, there's an easier way than using DOS programs/musical notes. Here's a freeware tone generator where you can pick any frequency you want:
http://www.nch.com.au/tonegen/

Just a thought, and I have to say I didn't go out and read Oliver's info, but he is in Europe and I think all the countries there are still 50Hz. Is he actually using 60Hz?. Could that have something to do with the 1200 vs 1350.....

Just wondering...

quote Originally posted by Wipe0ut: If you want to use a PC as a tone generator, there's an easier way than using DOS programs/musical notes. Here's a freeware tone generator where you can pick any frequency you want: http://www.nch.com.au/tonegen/

Cool! ... especially for those who prefer not to write their own software. Thanks for posting that.

Update: I downloaded (free) and tested the NCH Tone Generator software from the link you provided. This is a perfect tool for exactly the kind of tachometer and speedometer calibration being discussed here. All you need is a Windows PC, Mac (OSX), or Palm Pilot with an embedded audio chip set or a sound card. Highly recommended! Thanks again for the info.

quote Originally posted by Dodgerunner: [Oliver] is in Europe and I think all the countries there are still 50Hz. Is he actually using 60Hz?. Could that have something to do with the 1200 vs 1350..... Just wondering...

Reasonable question, but the answer is no. Oliver is a smart guy, and he specifically used 60 Hz in his article.

6 cylinders & 1200 rpm => 3600 spark events per minute => 60 spark events per second. At 50 Hz, the tach should read 1000 rpm (6-cylinder) or 1500 rpm (4-cylinder).

[This message has been edited by Marvin McInnis (edited 03-10-2006).]

quote Originally posted by mize: One more question...There are two ways to fine tune the gauge; changing the capacitor and changing the resistor. Any thoughts on this?

You can change either, both the resistor and capacitor set up the RC constant for the circuit to work.

No, if the tach is off, you can not simply replace the old parts with new ones and expect it to work correctly. Each tach was calibrated after they were assembled and the resistor was laser cut to compensate for the different tolerances of the other components in the circuit. It is tough to get 2 resistors/capacitors/etc exactly the same.

I would install it in the car a set the tach to read what your scanner reads.

The factory tach is not linear nor accurate, it's just there for 'general' information - if you need the tach to be accurate through the entire rpm range, you need to buy and after market one.

The tach will work without the filter, but it may bounce as the tach signal is not entirely clean. The tach wire is not even close to being a shielded wire to prevents stray signals being induced into it.

quote Originally posted by Marvin McInnis: For those who're still interested, I have posted a chart of musical notes vs. frequency back in my original post in this thread. With this information you can use a PC sound card as a high-precision frequency generator for tachometer and speedometer calibration.

I came up with this idea independently and have another easy way, Casio Keyboard. Just tap into the output and instant tone generator. Pan Flute is the closest to a sine wave.

Maybe no one knows, but how fast can I drive the speedometer before the electronics give out? In other words If I have a digital odometer and want to drive it back to zero it would take months at 60 mph but if I could get it up to say 17,000 mph or 5 miles a second I could get it done in a couple of hours. Anyone try this???

By the way I have driven mechanical odometers this fast with the drill and they tend to melt

[This message has been edited by jscott1 (edited 03-10-2006).]

quote Originally posted by jscott1: I came up with this idea independently and have another easy way, Casio Keyboard. Just tap into the output and instant tone generator. Pan Flute is the closest to a sine wave.

Good idea, since I don't think Pan Flute master Yanni is going to be available to help us with tach calibration for a while.

On a more serious note, the signal normally sent to the tach is far from a sine wave. Downstream of the tach filter it's probably a stream of pretty narrow pulses. But from personal experience I know that a sine wave works just fine for calibration purposes ... you just don't want a musical waveform with a lot of high-amplitude harmonics in it (e.g. the 1812 Overture probably wouldn't produce very accurate results).

[This message has been edited by Marvin McInnis (edited 03-10-2006).]

Here's my results. I installed the tach that read 1350 at 60Hz from the function generator. I hooked up the ALDL scanner and compared them. At idle, the two were off about 50 - 100 RPM. As I got up to 3000 and 4000 it actually became more accurate, usually within 25 RPM.

I am damn happy with the results!!!

I have to agree with the GM service manual.

quote Originally posted by Marvin McInnis: On a more serious note, the signal normally sent to the tach is far from a sine wave. Downstream of the tach filter it's probably a stream of pretty narrow pulses.

Yeah the tach signal is not an analog at all, (a series of pulse really), but it's close enough for a rough calibration of a gauge that isn't too accurate in the first place. And where is Zamfir when you need him