I think the 5V voltage limiter in my gas gauge is on its way out. There are a several pages on the internet that detail how to replace this mechanical device with a solid state device.

The solid state device is a L7805 series linear regulator and the Allpar article recommends using a Texas Instruments #UA7805C paired with a 10µF capacitor (DK PN P5999-ND) to absorb voltage spikes. When I tried ordering these parts from my local electronic shop, I discovered that neither was available and so I did a bit more research.

Linear devices regulate voltage by converting the energy available from the supplied voltage into heat which requires it to be paired with a large heat sink. The greater the difference between the input and output voltages, the greater the amount of heat generated. For small loads, the heat generation is not a problem.

The capacitor is not used to protect the L7805 regulator from voltage spikes because these are rated to 25V or 35V. It is actually a "decoupling" capacitor and 2 are recommended (0.33µF on input and 0.1µF on output):
See L78 Positive voltage regulator ICs page 25 for the diagram. I'm leaning towards using the DK 497-15681-5-ND instead of TI part.

As I don't have an electronics background, can anyone here tell me what would be the ideal capacitors for the L7805 voltage regulator when used for control the voltage supplied to the gasoline and temperature gauges?

Even though switching voltage regulators are more efficient than linear regulators, they are not generally recommended because of their higher cost. However, since they generate a lot less heat than a linear regulator, a heat sink is often not required. This makes using a switching regulator potentially cheaper than a linear regulator.

Having checked the Digi-Key catalog for switching regulators in a TO-220 configuration, I came upon the Micrel LM2576. This device has 5 pins instead of the L7805's three. Can anyone tell me:

• Is PIN 4 (Feedback) just connected to the output voltage?
• Where is PIN 5 (ON/OFF) connected?
• What are the ideal specs (PNs?) for the 100 & 330µF capacitors?
• What is the ideal spec (PN?) for the 330µH inductor?
• What is the ideal spec (PN?) for the 1N5822 diode?

_________________
1965 Plymouth Barracuda,
225 engine, Quadrajet, HEI, Dutra Duals, 904 Torqueflite, 2.76:1 axle, Addco front bar
Rods & Relics - Fort Erie, ON / Collector Car Tech

No need to reinvent the wheel from scratch. These are really excellent up-to-date gauge limiters (I have at least one on the shelf, new, that I'm not going to use—can mail it after next Friday; send me a PM). Here is the page on how to use such a limiter on a car originally equipped with a limiter built into the gauge.

_________________
一期一会
Too many people who were born on third base actually believe they've hit a triple.

Use any to-220 package you can get your hands on. The exact brand and part number won't matter. The cap configuration - one on the input and one on the output as the datasheet says is ideal.

Thanks for the quick replies.

I hadn't thought of using the RTE IVR3 or IVR4 limiters as I thought that these parts were a bit pricey for my needs. From what I can see, the amount of work to use an external limiter is only marginally less than it what it would take to install a L7805 and heatsink.

From the DK web site, the cost of going the DIY route appears to be:

• C$0.98 - DKPN 497-15681-5-ND (L7805 linear regulator)
• C$3.42 - DKPN 294-1068-ND (TO-220 heat sink)
• C$1.02 - DKPN 478-1884-ND (0.33µF capacitor)
• C$1.02 - DKPN 478-1832-ND (0.1µF capacitor)

The total comes out to C$6.44. My electronic shop adds 25% to the DK price so the price would C$8.05 + HST.

The STMicroelectronics product manual just states that a 0.33 μF or LARGER tantalum, mylar or other capacitor having low internal impedance at high frequencies should be chosen.

Capacitors aren't very expensive so I was trying to understand if there was any benefit to using a larger capacitor (like the 10µF recommended by Richard Ehrenberg). I have no idea what are the practical differences between tantalum, mylar, or other capacitor.

The switching regulator would cost:

• C$2.61 - DKPN 576-1518-5-ND (LM2576 switching regulator, 3A rating)
• C$0.73 - DKPN 1N5822FSCT-ND (1N5822 diode)
• C$4.14 - DKPN 2200HT-331-H-RC-ND (330µH inductor)
• C$0.89 - 493-6330-1-ND (100µF capacitor)
• C$1.17 - DKPN 493-9867-1-ND (330µF capacitor)

The total for using a switching regulator comes out to C$9.54 but I'm not sure if I've picked the right parts. The extra components might also require a breadboard for mounting.

_________________
1965 Plymouth Barracuda,
225 engine, Quadrajet, HEI, Dutra Duals, 904 Torqueflite, 2.76:1 axle, Addco front bar
Rods & Relics - Fort Erie, ON / Collector Car Tech

Switching makes things un-necessarily complicated here. Stick with the linear unless your in the experimenting mood.

You can mount the 7805 right to the metal frame and skip the heatsink. Tab is the same as the middle / ground pin.

Type of capacitor will affect a property called ESR - equivalent series resistance.

The 10uf was probably a "hey, I have this handy... let's try it... hey it works! neat, I'll leave it there..." thing. Depending on ambient temperature, current, and the alignment of the sun/moon/stars it may or may not make a difference.

That's probably way more heat sink than you need...........

You could probably skip the listed capacitors too.


With a 7805, I usually only use a 10 uf on the output so voltage is stable with no load.



Since a 7805 is cheap, get a couple so you have a spare.......

_________________
Ed
64 Valiant 225 / 904 / 42:1 manual steering / 9" drum brakes

Datasheets aren't hard and fast rules. More like a rule of thumb. Lots of things can work. But if you want things to behave as rated in the sheet, use the reference design. Otherwise there will always be the what if in back of your mind.

A single larger cap on the output may not be as stable as it should be, especially with transients. You may not be able to see this unless you scope things out. Since we don't care about transient response here it may not matter. But for the cost... why leave it out?

To troubleshoot my low-reading fuel and temperature gauges, I hooked up my 5V limiter to a voltage sensor I borrowed from my physics teaching lab.* Here's what I found:

As advertised, the voltage is either on or off. Specifically, it cycles between about 6.2 volts and zero. Because the gauge needles move as a bimetallic coil is heated, I figure the needle position is directly determined by the average power to the needle (power = heating rate). Now the math: power is current X voltage, and current is voltage / resistance, so power is voltage squared / resistance. The resistance is determined by the sending unit, so to get proper gauge behavior, you need the correct average voltage-squared from the limiter. Since I'm looking for an "effective" five volts output, I should see 25 volts-squared on average. However, in my measurement the average (calculated by computer) comes to 9.5 volts-squared, or just over 3 "effective" volts.

How to fix it? I'm guessing that maybe I can simply bend the strip in the limiter so that it closes more quickly after opening. That would increase the average voltage-squared to the gauges without changing the max voltage (6.2 volts in my case, less than 12 volts because of the resistance of the limiter itself). The goal would be to lengthen the "on" periods and shorten the "off" periods until the average volts-squared is 25 or the gauges read as expected. Of course I haven't tried it yet...

*The sensor is just a voltmeter that reads out to a computer file so you can see how the voltage varies over time. I unplugged the coolant temp sending unit and connected my voltage sensor terminals between the sending unit wire (not the unit itself) and the chassis, and turned the ignition key to "run." In other words I put the voltage sensor in the circuit where the temp sending unit normally goes.