Smiths tachometer — Electronic upgrade

This project began as a small diversion (Yeah, right!) from the main job of working on the mechanical aspects of the Jag. I haven’t played about with electronics for years, so it’s been fun from that point of view as well.

Smiths Instruments manufactured the Mk2 tachometer. It’s a nice-looking instrument which includes an electric clock positioned at the bottom of the meter face. The clocks rarely work, but that’s for another day.

As an aside, I have always called these instruments “Rev Counters”, I soon learned that google searches for Smiths tachometer, or tach, were much more productive than “Mk2 Jag rev counter”, or similar. So tachometer it is for this post.

Contents

Introduction

The instrument itself is a moving coil meter which derives its input from an AC generator mounted on the rear of the RH camshaft. The AC voltage from the generator is converted to a DC voltage via a bridge rectifier, located in the tachometer case. The tachometer, therefore, operates as a DC Voltmeter calibrated to display RPM. According to the service manual the generator produces 1V AC per 100 rpm.

This generation of tachometer is known as the RV type. The type number is printed onto the dial face. Jaguar was one of the first British car manufacturers to implement this type of electric tachometer which preceded the later electronic types. (Classic British Car Electrical Systems by Rick Astley)

This is all pretty straightforward and in fact, a nice solution for its time. However, the XJ6 3.4 engine I’m re-powering my car with comes from a later era when the availability of low-cost transistors and (later on) integrated circuits, made the camshaft mounted generator unnecessary. During this period, manufacturers such as Smiths were kept busy adopting these new devices into their tachometers, the result being a dizzying array of different models.

One of the best guides to these instruments is Alex Miller’s A Gentleman’s Guide to Smiths Electronic Tachometers which provides a wealth of information on the various generations of Smiths tachometers and links to other useful sources of information. In Alex’s document the early electric tachometer fitted to my 1962 Mk2 is referred to as the Gen 0 (Generation 0), the later RVI (Gen 1&2) and RVC (Gen 3&4) models discussed in the document are true electronic tachometers with circuits triggered from the cars ignition circuit. I wish to thank Alex for letting me share his document.

For more detail on the theory of operation and circuit diagrams of the RVI and RVC instruments, Mark Olsen’s Accutach website is well worth a visit, and Rick Astley’s book (mentioned above) has some good information.

Jaguar fitted the RVI type (Gen 1), and the later RVC (Gen 2) to various models as the Smiths evolved the instruments. Out of curiosity, I bought examples of these TradeMe — warning in case you hadn’t guessed by now; tachometers can be very addictive little devices, well for me anyway. My excuse (not that I need one) is that I started my working life as a Radio Tech, and these devices are effectively time capsules from the early days of semiconductors.

Upgrade approach

Getting back to my situation, I know it’s possible to retrofit the AC generator to the later engines, but this is mechanically quite complex and not something I was keen to undertake. Also, my research suggested that the magnets on the generator weaken over time which can cause the tachometer to read low. The real reason (of course) is that I wanted to have a go at building something up myself!

I found a few examples of circuits designed to solve the various issues these tachometers develop as components age. The one I liked the most is at Adrian Le Hanne’s Dinoplex website. All the information is made freely available to the DIY enthusiast for which I thank Adrain.

Luckily, I had an old Gen 0 tachometer lying around in my “spare parts dept” which I could use as a donor instrument to experiment on without having to worry too much about damaging things.

Meter disassembly

The first step was to disassemble the tachometer.

Removing the bezel can be tricky depending on the condition of the instrument. The bezel is held in place by eight tabs which are bent over the rim of the case. Removal (theoretically) is a matter of rotating the bezel until the tabs line up with slots on the edge of the case. The bezel and glass can then be lifted off. I ended up applying some CRC (not much) with the tachometer face down so the stuff wouldn’t get into the mechanism. I then lifted the tabs slightly with a small screwdriver and applied light pressure between the case and tabs until the bezel had freed and moved slightly away from the case with the pressure from the screwdriver. I did this as gently as I could and managed not to damage the bezel much, if at all. Eventually (patience is required) the thing freed and I was able to rotate and remove the bezel, glass, seal and inner ring.

It’s worth mentioning that Alex’s document has some tips for removing the bezel without damaging things, something I wasn’t concerned about with the donor instrument.

The meter movement is removed by unscrewing the two brass machine screws which hold the movement to the case. The two remaining screws go through the case and attach the internal component board to the movement. The dial complete with meter movement can now be removed from the case. I was careful to avoid touching the dial face too much with my hands. I’ll wear gloves on a good meter to avoid any damage to the dial face.

Now we can finally get a good look at some early 1960s electrical components. In the images below, you can see the series resistor and full-wave bridge rectifier making up the meter driver circuit. Well, I think this is a resistor. I’ve seen it referred to as both a resistor and an inductor. It looks like an inductor to me, but I wasn’t able to measure any inductance value, whereas it measures ~1000 Ohms resistance. Maybe someone can provide an answer. Anyway, none of this matters as we are discarding these components.

Here’s a diagram of the circuit.

To get at things properly the pointer and dial need to be removed. To remove the pointer I used a couple of plastic bike tyre levers (the lip on the blunt ends fitted under the head of the pointer) and gently applied pressure — pause, a bit more pressure — and the pointer went flying through the air. Wear safety glasses for this step! This was my way of doing this; there are others. Whatever method is employed, it pays to protect the meter face with some thin card. Next time I’ll experiment with something that can slide right under the pointer. The dial is attached to the meter movement by two small machine screws.

Here are a couple of close-up images of the board, I’ve begun to suspect that the blob of solder on the meter arm is to balance the movement, the later meters have a moveable brass counterweight.

To remove the rectifier components, I cut the resistor’s brass mounting tab off at the base and then levered the star washer off the top of the pin which holds the rectifier in place. Selenium (the rectifier) is a poisonous material, so I wore gloves for this step.

Here is the board (image below) with the resistor and rectifier removed, note this is from a slightly later tachometer. The meter is more refined, with the brass counterweight mentioned above, and the board has a different mounting hole layout.

Preparation for build and test

All the required information for the replacement driver circuit is available on the Dinoplex website. I built my test units using Veroboard and ordered enough components to make five boards. As electronic components are low cost and minimum order quantities often apply, this made the most sense.

I originally planned to mount the board using the two threaded holes on the meter frame, the same method used in the Dinoplex example. I trialled this approach with a template made from thin card and discovered there wasn’t enough clearance around the movement for this to work. This is due to the different mounting arrangement on the early meters. I ended up re-purposing the mounting hole for the rectifier and drilling another one opposite to piggyback the new board onto the original.

Circuit build

Once I had a board shape that worked, I created a component layout with Graphic for Mac which enabled me to move components around until everything fitted on the board.

Then it was a matter of adding the components to the board, cutting a few of the tracks where required to avoid unwanted connections, and soldering everything to the board. At this point, I discovered that something seems to have happened to my eyesight since I last played around with electronics! I’ve since purchased a LED magnifier which makes these tasks much easier and more pleasant.

Installing the board

I used PCB pins and sockets for external connections, allowing the board to be easily unplugged. Handy for a test unit. For the power connections I reused the original terminals for the AC generator input, the trigger input is a separate lead which can be connected either for calibration or coil -ve, and the meter output is connected directly to the original meter wires.

Test and calibration

My original plan was to test this at a friend’s home electronics lab. With lockdown fast approaching, I decided to invest in some low cost “test equipment”, consisting of a kitset function generator and a DSO138 digital oscilloscope. At well under $100.00 the scope kit, hastily purchased, turned out to be great value and all I needed for this project. I have subsequently tested the DSO138 to ensure it displays frequency accurately, which it does. It’s a cheap and cheerful solution, perfect for this job. As for the function generator, well, it produces a good enough square wave for what’s needed.

Before refitting the needle and dial, I did a quick test to see if things looked like they were going to work. With some trepidation I hooked everything up and gradually increased the frequency … and the meter coil MOVED … Phew!

As per the calibration instructions on the Dinoplex website, the meter is calibrated near the middle of it’s the range, i.e. with the pointer at 12 o’clock. The frequency to RPM calculation is; Frequency (HZ) = RPM/20 for a six-cylinder engine. For 3000rpm (the 12 o’clock position), the required test frequency is 150Hz. The rudimentary, but functional, test setup is shown below. The black jiffy box (centre right) houses the function generator, and the DSO138 scope is bottom right.

Meter linearity (accuracy)

I ran tests to check the linearity of the meter plus driver circuit over the measurement range and also some rudimentary tests to see how the long-suffering test meter responded to temperature changes. At this point, it’s worth remembering that these meters are fifty-plus years old, and apart from checking they are mechanically OK, I haven’t attempted to clean, lubricate, or balance the meter movement. The moving coil assembly is quite delicate, and not something I want to interfere with.

Tachometer linearity test results

Calibration Frequency (Hz)	Real (calculated) RPM	RPM reading on tach	% Error
50	1000	1100	10
75	1500	1550	3.3
100	2000	2050	2.5
125	2500	2450	-2
150	3000	3000	0
175	3500	3500	0
200	4000	4100	2.5
225	4500	4600	2.2
250	5000	5150	3
275	5500	5700	3.6
300	6000	6200	3.3

Impact of temperature

Even though the thermal drift of Adrian’s design is very low, there is still the issue of the resistance of the copper meter coil changing with temperature. This is something that Alex Miller pointed out to me, so I decided to run some tests with the aid of a mini oven and a BBQ thermometer. Slightly obsessive behaviour, I admit.

At 40⁰C the tach read 100 RPM low (from a 3000 rpm reference point), I did some tests at higher temperatures with the meter continuing to read lower as the resistance of the meter increased. I’m happy with this result as I can’t see the temperature variations being that great when the meter is in its case under the dash. At least I hope not!

As mentioned, I have made two units, the RV 7403/00 (my test meter), and a later one from a Daimler V8. I’m happy to report that the test meter has been proven in a vehicle with good old-fashioned coil ignition … an Armstrong Siddeley located at the Surgery, and … it worked! The Daimler V8 meter is awaiting a test in a real Daimler, so fingers crossed.

Summary

I’m happy with this modification, and while I would prefer the meter to be more accurate, it’s close enough for what I need and is within the 5% accuracy stated in the Dinoplex notes, except at the lowest end of the scale. I’m going to continue to play about with things, I suspect the later movements may be more accurate. So, in conclusion:

• The simplest way to upgrade the tachometer in our old Jags is to purchase a brand new tachometer with modern electronics, which retains the vintage Smiths “look”. One supplier offering these is Caerbont Automotive](https://www.smiths-instruments.co.uk/jaguar). I have no experience with this approach.
• Another option which retains the original tachometer is to have the unit upgraded at a specialist instrument shop, an internet search or check on a forum like Jag_lovers should locate one in most countries.
• The third alternative is to buy a replacement board from a supplier, e.g. Spiyda mentioned in Alex Miller’s document.
• And finally, there’s the option of building something similar to the Dinoplex solution described in this post. It might just take a while!

Final Notes

I’ve summarised in the table below an overview of the various Smiths tachometers fitted to Jaguar vehicles. This is by no means a comprehensive list, it’s what I’ve been able to pick-up from internet searches and the tachometers I’ve acquired on TradeMe.

Type number	Size	Gen	Vehicles	Notes
RV 7403/00	5"	0	Mk 2	Early evolution - test tach
RV 7413/00	5"	0	Daimler, Mk 2, S-Type, S1 E-Type	Later variant, improved meter movement
RVI 4611/01	5"	2	S1-XJ6	Single transistor circuit, no clock
RVC 2612/02	4"	3	S2-XJ6	Integrated circuit, round (jewel) warning lights
RVC 2615/00F	4"	4	S3-XJ6	Integrated circuit, square warning lights

OK, well this has been an interesting diversion, now onto other things!