All my searches for tachometer issues came up with threads about adding aftermarket tachometers. Mine has one on the dash, stock, but it's not working. The needle is always below 1000 RPM for short drives. It barely ever moves. When I've been on the highway doing 100+ km/hr for 30 minutes, it will slowly start to rise, but never above ~1500. Once it's warmed up like this it will generally move proportionally between 0 and 1500 in sync with the RPM. Any ideas? Some basic research seems to indicate this signal typically comes from the alternator rather than a dedicated sensor, on most cars. The fact that a warm engine means the needle moves more suggests to me the problem is probably in the vicinity of the engine, rather than a problem with the gauge itself or the wiring in the front half of the vehicle.
The tachometer signal comes from the coil, not the alternator. There is a dedicated tachometer output line on the coil itself that goes all of the way back up to the cluster. I would say that your coil likely has some fault.
Good info, thanks. I am about to do the wires/cap/rotor/plugs this week so I will try to have a look at the coil while I'm in there too.
Yep tach get signal from coil negative, likely have something up with the wiring or cluster itself. Trace the wire from coil neg to the cluster first and if it's good, might be a tach issue
The sense wire runs to both the ECM and the cluster . There’s one I’ve been messing with that has a non working unit ATM haven’t dug into it yet . I don’t have my books or notes at home to tell you which wire to check going into the cluster . If it’s sort of moving it sort of sounds like the gauge itself might be having issues but you’ll need to really check the pulse frequency at the connector to be sure on where to proceed
I removed the cluster but nothing is labeled and the mostly-useless James Danko manual of course does not include any clear diagrams of the cluster wiring. There is no easy way to tell which conductors are for the tach. Which conductors go to the tach? Is the tachometer a voltmeter, ammeter, or what? How would I test it even if I knew which conductors are supposed to make it move?
Here is the tachometer removed entirely. It connects to the flexible circuit board behind the cluster at three points. I assume signal, ground, +12 V. When viewed from the front, I think the terminal at 2:00 is signal, since the trace connecting to that one goes directly to a pin on one of the incoming connectors and nothing else is connected to it. I haven't been able to convince myself which is which for the other two.
I matched the three tach connection points to the ones on the larger connector that goes into the cluster by following the traces. It's the 3rd one from the outside, top and bottom, and the 7th one on the bottom. The 3rd from the outside on top seems to be ground. On the other two, I measure between 11 and 15 V to ground. The voltage does not appear to be directly or immediately proportional to RPM, but I can say that when I rev the engine high it hits 15 V and if I let it idle for a while it dips below 12.
If you have a good multimeter, which has a micro farad setting, check the components on the tach board. Capacitors get old.
Unfortunately mine is not that fancy. I rigged up a voltage divider with my 12 VDC power supply. I determined that, viewed from the front, 10:00 is ground, 2:00 is VCC (update: signal, not VCC), and 6:00 is signal (update: VCC, not signal). I found that between 3 and 8 V, the needle moves about 75 degrees, but then stops, and higher voltages do not move the needle further. Normal ROM on this one is 180 degrees. When the tach is installed in the car, it never moves more than about 5 degrees. You would think that if it was only the tach, I would get close to zero motion on the bench, and if it was only a problem with the wiring/coil, I would get full ROM on the bench. So that suggests both have an issue. However, I measured over 15 V on both non-ground conductors in the vehicle when revving the engine. So it would seem that the signal voltage being too low can't be the problem. Strangely, the voltage on both non-ground conductors appears to always be the same. Is one supposed to be higher than the other? Why are there two conductors if they both carry the same voltage? The needle won't move at all unless both are connected to positive voltage. I took the tach board and directly wired it to the connector in the dash. When I turn the key to one click shy of ignition, the needle jumps by something like 60 degrees and holds its position. I don't normally get to see this because the needle position has been set so that the relaxed position of the tach is well below zero. When I start the engine, the needle does not move further. When I rev the engine, the needle moves maybe 1 degree further, the typical type of motion I am used to seeing. I'm thinking defective tach at this point. Unfortunately they seem to be going for over $300 each on jp-carparts.com.
I have a truck I’m working on here at home that has a motor issue , where are you located ? I don’t need the cluster for what I’m working at this time . It’s a automatic transmission truck but I’m sure the tach is the same . I could send the cluster to you and test it out , you just pay the round trip shipping if it works correctly then you know it’s the gauge and nothing else . Or I could check voltages off the pins while it’s installed and running here Up to you
The-condition-of the-coil-should-not-effect-the-tach. The-tach-uses-the-break-of-the-points-as-a-signal. At-1000-rpm-the-tach-gets-1500-pulses-per-minute.—One-every-time-the-points-open.—Then-accumulates-those-signals-to-determine-how-many-rpm-you’re-turning. That-is-why-I-said-to-check-the-capaicitors.
I found a nice multimeter at work that measures capacitance. I'm going to take it home tonight and check the caps.
Jigs-n-fixtures, I think you nailed it, buddy. Upon trying to desolder the first capacitor, I found that one of its leads had corroded through. It snapped in two the moment I bumped the cap with my finger. The others all look to be in good shape, so I didn't bother to desolder them and measure them since it seems like I already found the source of the problem. I got a replacement on order, hopefully it will arrive by next weekend.
I-miss-RadioShack.—You-shouldn’t-need-to-desolder-the-caps. You-just-measure-them-on-the-board.—Only-problem-is-the-machine-bent-leads,-don’t-always-end-up-with-the-rating-where-you-can-find-it.
. Me too. That was my first thought as well. I could have this thing back together tomorrow, now I gotta wait a week, and buy ten of the damn things when I'll only ever need one. I hear some meters are advertised as being able to measure caps on the board, but opinions on that are all over the place. Certainly cheaper ones are not going to be reliably accurate. You are measuring the capacitance of the entire circuit across two points. It's possible to get a reading that's double or more than the actual value. I guess if you are looking for numbers that are wildly, orders-of-magnitude off from nominal, then it's probably a viable approach, though.
I got a replacement cap from a coworker. Nice. Soldered it in. Mounted the panel in the vehicle and then started the engine. Tach behaves exactly the same. No improvement whatsoever. Damn, really thought I'd found the smoking gun. I'll still probably desolder the remaining caps and check them, but this is getting very close to the threshold where I decide my very limited time is worth too much to dedicate more of my hours toward this project, especially if I could just buy a replacement assembly for ~$250 instead.
Desoldered and measured all caps, all seem OK. I borrowed an oscilloscope from work. If I can figure out how to use it, I will check the signal at the dash connector.
Got the oscilloscope working. I had VCC and signal mixed up in my previous post. I can measure a distinctive waveform whose frequency increases as I rev the engine. This is more or less exactly what I would expect. https://photos.app.goo.gl/tT5SgyWvZRbtba239 Signal seems to make sense All caps seem to be OK Tach needle is physically capable of moving itself What's left to check?
I'm looking into buying a replacement IC, but so far it seems very unlikely that I'll be able to find one. I'm not super excited about the prospect of spending $300 on a replacement tach that may not even fit my vehicle even if I can get someone in Japan to find what seems like the right thing and ship it to me. I'm actually thinking right now the easiest way to make this work is to replace the existing tach with an Arduino Nano and a small stepper motor. The Arduino can count pulses from the signal wire and move the stepper accordingly. Probably more responsively than the stock tach, actually. This also gives me the ability to fine-tune the way the needle behaves, including using an external tach to calibrate it. I think I have every tool I need on hand to make that work except the external tach, but I'm sure I can find one of those to borrow somehow. Of course, that then opens up a major can of worms with the Arduino. I'll want to install a remote USB connection point so that I can reflash the Arduino without pulling the cluster out of the dash. And then I'll think of other stuff the Arduino can do (converting km/hr to mi/hr and displaying that somewhere, for example...) and end up sinking a lot of time into a new project. Which is fine. Story of my life, actually.
I found this today. Dang, I could make such an epic custom gauge cluster if I went this route...of course it would pretty much just tell me everything I already knew plus the engine RPM. But it's really tempting. https://www.reddit.com/r/3Dprinting...3d_printed/?utm_source=share&utm_medium=web2x
I think I'd need to make do with whatever signals are currently available from the connectors behind the dash. If I wanted to replace the entire cluster, that would necessitate putting an encoder on the speedometer cable and transitioning to a fully-digital tachometer, speedometer, and odometer. I design machine tools, robotic systems, and industrial automation for a living* and have access to multiple 3D printers, including an SLS system, so I'm sure I could make this happen if I was sufficiently motivated (and could find the time). I probably won't go that crazy in 2020, but who knows, I'm sure I'll need a good personal project to sink my teeth into at some point in the next few years. I got an Arduino Nano on order. I'm planning to attach it to the tach signal wire and see if I can write some basic code to turn that signal into an RPM reading. I already found some example code that seems like it can be hacked into working for me. If that approach seems viable after playing with the prototype then it's pretty likely that I'll be soldering up my own new tach board at minimum in the near future. *My favorite ME joke: Q: How do you know if there is a mechanical engineer in the room? A: They will tell you.
This set up is pretty tempting and really cheap for what it is. You're ~$200 for the display and all the sensors. https://www.justraceparts.com.au/10in1-automotive-digital-oled-multi-gauge-display
The problem I have with the JRP system is.. it is a complete ripoff of the Defi Advance ZD. They are not even trying to be subtle about it either. If you dont really care about that, then you might as well get it straight from China for half the price that JRP is selling it for.. Here is a whole kit with the sensors and everything for $120 shipped. https://www.aliexpress.com/item/4000243859240.html
Alright, so I've rigged an Arduino to the dash with a super basic setup and code, and hot damn, it looks like it's actually measuring the RPM! I'm so f'ing psyched, you guys, this might actually work after all. My prototype code is working on a 1-second interval. Starting very basic. I'm counting pulses from the coil and every second I divide the number of counted revolutions by the time since the last calculation, then I multiply by 60x1000 ms, divide it by 4 (4 cylinder engine) and then again divide by 2/3 (1500 pulses / 1000 revolutions, per Jigs-n-fixtures' comment). I still don't fully understand why the appropriate ratio is 2/3, but the numbers seem to make sense. At idle I'm getting 2100 RPM, and revving it pretty high in neutral it's 7000~7500. That seems right to me.
The ratio of coil pulses per rpm is based on the four cycle engine firing once every two rpm. If you have a 1-cylinder the engine the engine takeS 2-rpm to go through the cycles, Power>exhaust>intake>compression, then repeat. On a 2-cylinder the cycle are 180-degrees out of phase so it fires one cylinder every revolution. On a 3-cylinder the three cylinders need to fire very two rpm. 3-pulses per two rpm. On your 4-cylinder, you have 4-pulse per 2-rpm.
7100 RPM would be very close to redline. While it is possible you are revving that high, you would be pushing it pretty hard to hit that. 3500 seems more reasonable if you are just revving it a bit. Idle should be somewhere around 1000-1200 when cold, and about 850-900 when warmed up.
Okay, so the 1500/1000 comment was made under the assumption that this is a 3-cylinder. I took that multiplier out. Duh. The math that gives me somewhat sensible output this morning is: RPM = (60*1000*(pulses counted)/(time interval))/8 Idle at cold start: 3500 Idle after 5 min warm-up and revving a couple of times: 1850 Revving it pretty high: 5500-6000 I have it programmed to look for a rising signal. It seems like it's counting 8 pulses per rev, not 2. Perhaps even more than that, if I'm supposed to have a sub-1000 warm idle. I need to get my hands on an external tach. I'm definitely measuring something that seems like RPM, but perhaps the signal is noisy and what I'm seeing is not directly proportional to the actual RPM. If it's at least repeatable, I could do a multi-point calibration and end up with something reasonably accurate. 90% of my company is not showing up for work right now, so it may be a while before I can find one to borrow. Or I could just buy one on Amazon since I've been wanting to add an aftermarket tach to my big truck for a while... I'm going to work out a way to mount this stably enough that I can drive around and rig up some way of displaying RPM that I can read while driving. Hopefully today. Then I can report what RPM I'm seeing for various speeds in different gears. Then maybe someone can check that against what they see in their 5MT 660cc Sambar.
