View Full Version : constant voltage regulator

08-30-2014, 03:53 PM
How do I test them?

I know its been covered here before, but its buried in a thread somewhere.

I'm fixing gas and temp gauges. Neither work.
I installed a new regulator and still nothing..

I have two old regulators. One of them should work.

Or is there a common ground I should check?

08-30-2014, 04:07 PM
Tested gauges they're good.

Tested fuel sending unit. Got ohms of resistance there.

Nothing in shop manual tells you how to test the voltage regulator.

Can I hook up a volt meter to it? And supply 12v and see if it reduces voltage?

08-30-2014, 04:44 PM
I have 2 regulators that read nothing
And a third that pegs hot..

I dont know anymore.

08-30-2014, 05:56 PM
Are you talking about the small, bimetal regulator that mounts on the back of the instrument cluster? I've heard lots of people replacing them with solid state ones that are more reliable.

08-30-2014, 07:01 PM

Yea thats what im talking about.

I adjusted the screw on the back of the regulator and adjusted it till my volt meter read 5 volts at the gauge.

I put it back together and my gauges are working.
I have about 7 gallons in the tank and it reads just under half a tank. 16 gallon tank.

The coolant temp gauge moved as the car warmed up.
Problem is as soon as my mechanical gauge said 180 the temp gauge pegged hot and stayed there.

Guessing my coolant temp sensor is bad.

I'll try replacing it with another one.

So as of right now everything is working.

I'll order a solid state regulator for a spare.

09-02-2014, 11:02 AM

They test "best" with an analog volt meter, but I've used digital volt meters (DVM) too with OK results. The analog needle is "dampened" and does not react as fast to changes as digital, so a DVM will see higher spike voltages than an analog needle can physically accomplish. More on that below.

How they work: They are a bimetalic "heater" switch and the idea was to alternate the voltage/current on and off again at a rate that never let it reach a full 12 volt output. Very crude, but apparently worked. Internally one side is a fixed point contact. The other contact is at the end of a bimetalic spring. When the spring is cold it closes the switch - full voltage at the two terminals (12 volts in, 12 volts out). But when it closes the switch and current is flowing this very quickly heats the bimetalic spring which causes it to move in the opposite direction of closed to open the switch - zero voltage. The spring now cools again, closes, full voltage, heat, open, cool, close, rinse, repeat - on/off/on/off. The average on-off voltage created (though seeing 12 volts) is about 5 volts.

To test: ground the body to -12 volts and apply +12 volts to the input terminal. Measure the output (body ground to output terminal). With an analog volt meter the needle swings from 0-5 volts (+1/-0). With a DVM the reading will be higher than the 0-5 volts. The digital reading is more reactive so you will see slightly higher than the 5 volts +1 reading. I saw up to +3 (8 volts) using a DVM. You'll see a "glued-in-place" screw that originally set the bimetalic spring tension to get the on/off cycle just right; this to make it output the 0-5 volts.

Suffice it to say this design is VERY FRAGILE and any attempt I made to adjust them (if they worked, but gave the wrong reading) or look inside one and make a bad one good (there is a set of contacts inside that wear from the constant ON/OFF cycle) and so my final analysis of them is - if they work, great. Otherwise they are trash.

I did pick up a few of the solid-state versions they offer now and put them in various cars I've touched of late, all with great results. They cost a little more, but will last a long time. Transistors can handle this switching action quite readily. I did find a site showing how to make them too, but I've not got around to making any, though I did get the parts to do so.

Real Soon Now.

Good luck!

09-02-2014, 12:44 PM

I had three one being less than a year old.
Two did nothing including the new one.
The third unit pegged the needles.
Since that unit did something I adjusted it.
Following the test procedure in the manual I saw I was getting way to much voltage. So I adjusted the screw on the back until my digital volt meter read 5 volts

This procedure got my temp gauge reading cold and my fuel reading half tank which it was.

Put everything back together started car and as the engine warmed up the temp moved.

Once the temp hit 180 the needle pegged hot. Im guessing my temp sending unit is damaged or incorrect.

09-02-2014, 01:21 PM
I think with the DVM you should shoot closer to an 8-volt reading as I saw with them.

I don't know the math that says what makes things read what; like, does the higher CVR setting make for higher or lower gauge readings. I didn't want to play with working units (at the risk of breaking them) to find that out. I do know most senders have a pretty reliable resistance reading range when new.

The new one... was it solid state? I think most being sold today are , whether they say so or not. Can't prove that, but such a no-brainer to use a transitorized 5-volt regulator and solder it in a can than to make that fine-wire heater wrapped around a bimetalic spring doo-dad. The transistors cost all of $.50 to buy (I think I paid $.10 a piece for mine on-line). Shipping cost me more.

09-02-2014, 09:36 PM

The new one had a set screw on the back just like the original.

The shop manual says probe the ignition side of regulator and the ground side of the gauge and it should read 5 volts.

So that's what I set it to. I'll see what happens.

09-03-2014, 10:53 AM

Keep in mind that back in 196X they only had analog volt meters. Because this is a crude on/off switch, the 5 volts was based upon how quickly the needle could swing from 0-5-0-5-0-5 etc. My digital volt meter saw much higher peak voltages on the same part - closer to 8 volts at times. I have both analog and digital volt meters, so it was based on this I got transistors that were higher than 5 volts.

I can't recall if I actually measured on the "purchased" solid-state parts. I think I still have one uninstalled, so I may try to check that.