To make life easy for you guys and not scramble anymore eggs, My regulator works a little different than what you guys are using. Mines workes like an ac transformer. Your calculated amp draw would only exist on the seconday side of the regulator. What I mean is this, I will use 10 in my example so everyone can understand. My regulator is based on V*A principles. So lets say the gauge was 5 volts and 2 amps to reach full power. The primary side would not draw 2 amps at 10 volts. It would draw 1 amp. (10/1=10) (5*2= 10). In your case that festering resistor would have less of a voltage drop than using lets say a standard lm7805 chip or equivilant. This makes the secondary side more solid. So in this case lets plug in some real world values:
You guys are calulating 5.5 volts at .9 amps.Thats about 4.95 watts at 3 fully loaded gauges. Let say the on average the alternator puts out 14.5 volts. The draw on your resistor is now .34 or so. That helps the incoming voltage not fall. Thats Unlike the solid state chips you guys are using, there are some benifits here. For instance: the chips you are using for example 12v in at 2 amps and 5.5 at 2 amps, the difference in power is 24 watts coming in, and 11 watts coming out. Thats means 13 watts of power is lost as heat. As we all know that pretty hot. If you would like to know how hot that is we are looking in the realm of about 160 degrees. Those high tempertures is what causes the chip to fail over time. What you see to everytime is someone adding a big heat sink to keep the chip from over heating and turning off. But its no different than your engine, you constantly run it a 210 degrees and you are sure to rebuild the engine a lot sooner than at 185. Now with the example I gave you, about the heat lost with the chip, try running your calcs and see how much heat is lost on mines. With three gauges fully pegged its not even starting to get warn.
Another thing to consider is grounding, grounding, grounding.
Pay alot of attention here. For you all that like to build your own regulators, remove theground on your regulator when its being fed 12v.
You will get a ferocious 12v right into the gauges. It is no different than placing a jumer wire across the chip. If your lucky you catch this in time. With the truck running at 50 psi and full tank of gas, with a warm engine. You are likely to blow all your sensors within 10 seconfs. The gauges themselves are more resiliant and can often not be damaged.
Again, since mines is built on ac principles, you lose a ground, you loose your power, it turns off.
Another problem is common failure, and nobody is immune here. But the chips you guys are using soild state and are physically bound together. Your chip decides to let go, again, yoy get the full 12v, and to the degree of placing a jumper wire.
My reguator has a transformer inside, If you feed it 12v constantly, the transformer cant work properly, it loses it ability to oscilate and nothing makes it to the secondary side. Again, it shuts down.

 

I know this thread is long and convoluted at times....
At first we were talking about using the linear regulators 780x and LM317 etc. Not sure if a working version ever got made with these?

Later on to OP ditched that idea and went with an off the shelf variable switching regulator. (Pin compatible with the 780x series).
This device also has a transformer as it is a switching supply. Much less power lost as heat etc.

This has pictures of the insides:


This is the 25W version with better picture of the inside:


Here is the data sheet
http://www.dimensionengineering.com/...s/DE-SWADJ.pdf

Fixed voltage devices are also available from that same company:
Is this any different than what you have?

 

This is not the same thing, this is a switch mode chip. This is switching on and off at high frequencies. Your load is directly connected to the source 12v. But its filtered throught the capacitor. Hence the reason its wrapped in a heatsink with filtering capacitors.
Mines is not connect to the input voltage, its completly isolated. You remove the ground on this one and you will still shoot 12v into your loads. its based on lt3573 thats simular to switch mode but with true isolation.I forgot to mention, that regulator will still draw the same coming in as it draws coming out. The only thing your doing is pulsing like the original IVR.

 

ic237... man, you need to keep the ending [ / quote] (without all the spaces) for the original text to be in a gray background.

BTW thanks for posting.

 

Sorry it aint easy on a mobile phne.

 

First, the results of todays run: Perfection! The regulator didn't even get warm. And, the gauges worked exactly like they should. The Temp gauge stayed exactly between the R and M of NORMAL for all 5 hours of driving, just as the mechanical temp gauge stayed exactly at 197. The oil pressure gauge shows N when the mechanical gauge shows 20 PSI, O when it shows 30 PSI, R when it shows 40, and M when it shows 50 PSI - each and every time. And the fuel gauge has no jumps to it the way it has always had - some times it would go up, some times it would go down. So, as far as I'm concerned this project is a wrap.

Wrong. The wave form cannot change if you don't change the components. So, as long as there is the stock 8 - 9 ohm input resistor, standard ICVR, gauges and senders the shape of the wave form cannot change. Yes, its frequency can change, so maybe movement changes that. But not the wave form. And, since the gauges use the output voltage and induced current in the form of heat, changing the frequency has no effect as long as the duty cycle, the ratio of off to on, doesn't change.

Remember that the shop manual says that a 10 ohm resistor will show full scale, with some words around exactly where the needle will be, a 22 ohm resistor will show mid-scale, and a 75 ohm resistor will show at the bottom mark of the scale. And sure enough, as shown in the pictures, that's exactly what both sets of gauges I tried showed with an ICVR and at a regulated 5.4 volts. So, while there may be a difference somewhere between each of those points, 5.4 volts gets exactly the readings as a working ICVR at bottom, mid, and full-scale.

Having said that, I did try other voltages and they did not work. Anything less than 5.4 volts didn't get as high of readings as the ICVR did. Any higher voltage than 5.4 gave readings higher than a good ICVR gives. And, in my opinion we want the gauges to read exactly like a working ICVR gives - without the bobbles that seem to come with them. Not higher, and not lower.

Go ahead, chicken, say it: As all of Gary's threads are.

Bingo. And, by the way, it is the one you found. GOOD JOB! IMHO this is a winner. An extremely simple solution that most can do on their own. As said, this project is a wrap. But, I'm happy to answer any questions.

 

Gary
all I can tell you is this, you can believe me or not

Every single "Waveform" is different. No two waves are the same. Especially with a resistor on the feed. I will tell you why, Each gauge is not in the same spot all the time, you have a combination of 3 different currents always changing. The voltage drop on the resistor is a direct reflection on what is going on down steam to it. You have a network of gauges and sensors always changing. Thus each peak is different and each pulse is spaced differently. You will never see the same wave twice. I promise you this. I cleaned out my garage and cant find my old regulators. I would be more than happy to show you what I mean. I have fords and jeeps with limiters also. They are all the same. I can asure you those vehicles have no resistor in series with the limiter and each peak is different and each pulse is spaced differently. Now you want to see the typical waveform that you are expecting, buy an RTE voltage limiter and you will see that perfect square wave over and over again. Use a mechanical limiter like the O.E and you will only see many spikes at different intervals. The first pic of the oscope in this thread is an obviously overheating limiter. The arm has lost its tension allowing it to get so hot that its overshoots and spends too much time in the cooldown period. When it makes contact again it takes too long to heat back up and spends way too much time in the on position. There you would have a square dc thats peak never changes and only frequency does. A properly working one will only make contact thousands of a second. And at that the contact never fully lands so the peaks vary from full battery voltage to as low as 8 vdc. So again all components the same, but the end result is different every time.

 

Yes, they are long and convoluted. I have contributed to some of that. I guess I will need to maybe build one of those for Matt sometime, and maybe for Pete Medlin for his Shelby clone.

 

Yeah just found out that the DE-SWAJD is 'Buck' converter based on the CS51411 chip. Uses a single inductor rather than a transformer.
I think it is entirely adequate for the job it's doing.
Pulsing like the original IVR is what we wanted.....

 

Great! So if a person installs one of these, does that mean our notoriously "unreliable" stock gauges can actually then be more or less trusted as precise (if not accurate) indicators of actual physical values?

If the little regulator didn't even get warm, then I would think it could be mounted directly to the rear of the cluster, such that it doesn't become another component which must be disconnected when removing the cluster from the dash.

 

Yup
It works just fine no questions there. I never doubted it works one bit. A little history on me. A few years ago when I bought my wife her mustang, I replaced the engine right off the bat since I blew on the way home from overheating and the gauge was reading low. True story. I bought new sensors and when it came time to turn the engine over, my gauges worked for literally 15 20 seconds. Then my oil and temp sensor blew. I was so pissed. I started replaced them again and blew them again. I thats when i stumbled on the little gray box. I bought another o.e. limiter and it worked for maybe a couple of days. I started beating my head on the floor over and over saying why me : ). You know how it is on a fresh rebuild you are paranoid with the temp and oil. I then read about the 7805 upgrade. Willing to try something new and being an electronics guy I was all for it. I did it and they worked...untill I removed my cluster and started changing out the bulbs for LED's. I put the cluster back in and wtf why is my temp and oil not working....Aparently the 7805 doesnt like not missing grounds. So I blew yet another set of sensors. Beating my head some more on the floor and saying why me, it hit the drawing board. At the time I got so frustrated I ripped out the system and installed a dakota digital dash. Problem solved. But I still had my jeep. It has a limiter in the fuel gauge. About 9 months later the gauges went. Been down this road before. After tons of trial and error and experimenting I wanted the best protection. There are many styles you can go. The runts drops, the diodes, the RTE, the 7805 and other chips. They all work but all have some sort of limitation. The ones that matter to me is the one that protects my system. Wanna see some picky folks? Talk to classic mustang owners. Some of the pickiest. They all will spend top dollar on original sensors from the 60's. They all believe today's replacement sensors read incorectly. Thats where I come into the picture. I have covered all the possible things that can go wrong and built security around it and maximized the benifits as best I can. RTE limiters are pretty good, but they too have their limits. For example one big flaw they have is the limiter has a start up sequence. Once it sees 12 volts and turns on it outputs 12v all the time for the "gauge warm up". They want to replicate exactly what the original limiter did. But what happens if the car doesnt start, and this is very true for the carb guys out there. We love to crank and crank squirt started fluid and crank some more. Every time you touch the key the limiter resets and starts over on the warmup cycle. The original ones dont have this problem since the arm would retain some heat. Well enough said I need a beer, good luck to all.

 

The first day...blown engine 13 miles after driving

 

Yes, I believe our gauges can be very accurate - as accurate as more modern gauges since the senders are pretty much the same as those used by more modern gauges. The big difference is that our gauges are moved by heat, so they are slow to react.

I do believe that one of these regulators attached to the rear of the cluster would work fine. However, since mine is mounted against a big chunk of aluminum it is possible that it was transferring its heat through the heat-shrink that is around it to the steering column bracket and would get warm without that. I don't know if that is the case but could take it loose from the bracket for the next trip and let it hang by the wire if that would help.

Look. I didn't read all of the previous post and I read none of this one. I will not take the time to read something that someone doesn't take the time to write with reasonable grammer or punctuation or white space. You want to post on my thread then write it so I can read it. Don't and please take it elsewhere.

As for waveform, that means the form of the wave. Each ICVR will have the same wave form. An ICVR is nothing but a switch that gets turned off by heat. So, when it closes you will see the voltage go straight up and level off. Then, when it opens you will see the voltage spike because of the electromagnetic pulse created by the coiled wire of the heating elements and variable resistors, and then it will go straight down. Yes, the amount of spike and the voltage attained will vary due to the position of the resistor's slider, but the form of the wave will remain the same. What will change is the frequency, meaning how often it cycles, and the duty cycle, the % of on/off time. But not the waveform.

If you want to sell your regulators go ahead. But I hope the people reading this thread understand that anything that supplies 5.4 volts to these gauges will make them read accurately. Linear, switching, or whatever, it doesn't matter as long as the regulator is capable of supplying enough current and can shed any heat created. The way the voltage is regulated doesn't matter since any of the approaches will provide a much better source than the ICVR's supply.

Having said that, anything but 5.4 volts will provide a reading that is not accurate. Don't be led astray.

 

I wasnt directing this to you. Actually I was writing to someone else. You dont have to dwell on my grammer, just do as before and dont read. Its okay, i will survive.

 

So, when can I buy one from ya?