S4 Voltage Stabilizer

[Temmck]

I bought a new Voltage Stabilizer from RD. Voltage at the "B" (battery) terminal reads 12.4V and the "I" (Instrument) terminal only reads 3.8V. I thought I had the wrong stabilizer but it has what appears to be the correct part number on it - 26B 6032 1000. Has anyone else seen this and do I need to buy yet another stabilizer?

Terry

[billwill]

Was this with the instruments connected to its output?
Otherwise it might show the wrong result.

But on the other hand it might have the wrong chip inside, if it is an electronic one.

If it is a mechanical vibrator (old-type) it's never actually 10 volts, it vibrates on and off and the AVERAGE output voltage is 10 volts. It does this by varying the length on the ON period versus the OFF period. Itt is intended only for slow instruments, such as those which bend a bi-metallic strip to move the needle; the sensor (e.g.fuel tank portion) is a variable resistor feeding a heating coil around the bi-metallic strip.

[Temmck]

I believe it to be an electronic one. I measured the same voltage with the fuel sensor connected and disconnected.

[Andy8421]

Terry,

This has come up on another thread recently. The voltage stabilisers are '3 terminal' devices - they have an input, output and an earth. Without the earth, they don't work. The earth can either be via a separate terminal, or through the case of the stabiliser where it mounts onto the speedo. This of course then implies the speedo also needs a good earth.

The electronic stabilisers have a current limit, about 1 amp, so measure the voltage with nothing connected to the 'I' terminal. If it is 10v, but drops when you connect the gauge, then possibly there are issues with the gauge or its wiring. If the stabiliser is still showing a low voltage with nothing connected to 'I', and you have the case (or separate terminal) earthed properly, then you probably have a dud.

The old bimetallic stabilisers didn't care about polarity, but the electronic versions do. They are now supplied in two flavours, one for negative earth, the other for positive. It is always possible you have received the wrong version. If it has markings, you could check it is the right one.

Andy.

[Temmck]

Thanks Andy,

I took the measurements with the stabilizer hanging from the wires, no ground wire to the case. I will ground the case and try again. I am very aware that some stabilizers have a ground connection and others ground through the case, I have no excuse for not grounding this circuit.

Terry

[Temmck]

Well, it looks like I jumped the gun on my last post. I grounded the case but still reading 3.8V both loaded and unloaded. I know that there are 5V stabilizers but I doubt they would be reading 3.8V. Do I have a bad stabilizer?

Terry

[pharriso]

Well, it looks like I jumped the gun on my last post. I grounded the case but still reading 3.8V both loaded and unloaded. I know that there are 5V stabilizers but I doubt they would be reading 3.8V. Do I have a bad stabilizer?

Terry

Sounds like it... give Ray a call.

[nmauduit]

fwiw I believe the vibrating kind stabilizers need a load (preferably close to the actual load) to be tested meaningfully, as their operation is based on a thermal effect (non purely linear) the actual current going through it has to be within an operating window.

[Craven]

A test of the vibrating voltage stabilizer I found satisfactory. Using a bench power supply you really need one with adjustable voltage and a current drawn indicator, connect up with just the input, case and B terminal, with power supply set low say 5 volts a current will flow, slowly increase the applied voltage the current will increase with it, continue to slowly increase the voltage until the current suddenly drops to zero this is the switching point, read off the voltage and you have the stabilizing voltage.

[Andy8421]

fwiw I believe the vibrating kind stabilizers need a load (preferably close to the actual load) to be tested meaningfully, as their operation is based on a thermal effect (non purely linear) the actual current going through it has to be within an operating window.

The original stabilisers are thermal 'vibrating' regulators, but the heating element is not in circuit with the load, so it makes no difference what the load is.

This gentleman meanders around the subject, but has some good pictures of the innards, and does explain how it all works reasonably well.

https://www.youtube.com/watch?v=WlZekK97ntg

For those who care, the variable mark/space approach to voltage control used in these stabilisers is how a modern switch mode power supply works. You will have many of these devices plugged in around your house.

[nmauduit]

fwiw I believe the vibrating kind stabilizers need a load (preferably close to the actual load) to be tested meaningfully, as their operation is based on a thermal effect (non purely linear) the actual current going through it has to be within an operating window.

The original stabilisers are thermal 'vibrating' regulators, but the heating element is not in circuit with the load, so it makes no difference what the load is.

This gentleman meanders around the subject, but has some good pictures of the innards, and does explain how it all works reasonably well.

I'm not sure what "not in circuit with" means exactly (not in *series"? it is actually in parallel with the circuit, since the "hot" end of the sensing wire is connected to the "regulated" lead , which is obvioulsy necessary to assess how much there is to regulate).
My point was these electromechanical systems are not easy to bench test just with simple large impedance digital multimeters, then attempts at calibrating them under testing conditions far from actual operating conditions may lead to unexpected results (regulating hysteresis being different from different overshoot due to different load for instance, which may or may not be compensated by switching frequency variations... also physical orientation in space may play a role due to gravity etc.). That said, I've not messed with that kind of can regulator in particular (only external dynamo regulators, Bosch and Lucas operating differently I believe), from what it seems on the video the principle would be rather simple, and it may actually only be operating in open loop, i.e. according to a thermomechanical reference (the heating wire physical characteristics when enclosed in a can) and not a voltage reference as a solid state would, so fluctuating at least with temperature but probably also with battery voltage... no big deal considering the accuracy level of the fuel gauge, I suppose getting it somewhat accurate towards the 1/4 to empty level is all one would expect (and the canned regulator could only be one-point calibrated).

[billwill]

Well, it looks like I jumped the gun on my last post. I grounded the case but still reading 3.8V both loaded and unloaded. I know that there are 5V stabilizers but I doubt they would be reading 3.8V. Do I have a bad stabilizer?

Terry

Lots of modern logic chips run at 3.3 volts, so I expect stabilizers that have an output around that figure are commonly built-in to electronic equipment.

[Andy8421]

fwiw I believe the vibrating kind stabilizers need a load (preferably close to the actual load) to be tested meaningfully, as their operation is based on a thermal effect (non purely linear) the actual current going through it has to be within an operating window.

The original stabilisers are thermal 'vibrating' regulators, but the heating element is not in circuit with the load, so it makes no difference what the load is.

This gentleman meanders around the subject, but has some good pictures of the innards, and does explain how it all works reasonably well.

I'm not sure what "not in circuit with" means exactly (not in *series"? it is actually in parallel with the circuit, since the "hot" end of the sensing wire is connected to the "regulated" lead , which is obvioulsy necessary to assess how much there is to regulate).
My point was these electromechanical systems are not easy to bench test just with simple large impedance digital multimeters, then attempts at calibrating them under testing conditions far from actual operating conditions may lead to unexpected results (regulating hysteresis being different from different overshoot due to different load for instance, which may or may not be compensated by switching frequency variations... also physical orientation in space may play a role due to gravity etc.). That said, I've not messed with that kind of can regulator in particular (only external dynamo regulators, Bosch and Lucas operating differently I believe), from what it seems on the video the principle would be rather simple, and it may actually only be operating in open loop, i.e. according to a thermomechanical reference (the heating wire physical characteristics when enclosed in a can) and not a voltage reference as a solid state would, so fluctuating at least with temperature but probably also with battery voltage... no big deal considering the accuracy level of the fuel gauge, I suppose getting it somewhat accurate towards the 1/4 to empty level is all one would expect (and the canned regulator could only be one-point calibrated).

The heating element is in parallel with the load, and therefore the load makes no difference to the amount of 'heating' done by the element, it only depends on battery voltage. It would pointless having a voltage stabiliser that depended on load. Have a look at the video I linked to, it shows how the internals are connected. When the internal contact is closed, the heating element is connected directly to the battery, the load doesn't play a part.

I would agree that the regulation provided by this device is likely to be terrible, and at best only work over a narrow range of battery voltage. To be fair, the gauges are terrible as well, and the sender in the tank isn't exactly a precision instrument.

There is a suggestion above about how to calibrate the stabiliser - as you say it isn't straightforward. Frankly, I wouldn't bother. The modern electronic regulators are cheap, and if you fancy having a go yourself, a pack of 10 LM7810 (the electronic bit in modern stabilisers) cost £2.55 on eBay with free postage.