Regulator Woes

[docc]

Just my own... cause your bikes don't even have a metal tank!

 

So, worse even. How to solder a plastic can? As if a "tin can" didn't sound cheap enough!

[Kiwi_Roy]

Come to think of it a simple battery charger might do it. connect it one way to check one Diode SCR pair then flip the leads to check the other pair.

Do you mean I connect 15VDC should give nothing? And then reverse poles on the input. I guess that should work. With a variable supply I could even spot the switch point. No wait, I could put the reference wire to a voltage divider potentiometer! And use a 15VDC or higher source. This might be a plan.

 

Maybe it's better just hooking it up properly to my bike though. This will test all of it at once. It's just that I would need to pull fairing and tank - and then mount the tank again for a test run, and then remove it again, and...

 

The output of a battery charger will be a series of pulses peaking about 18 Volts and dropping to zero but it would only check one pair of the diode SCR bridge

On second thoughts I'm just confusing the issue, If you can connect to the transformer inside the charger or find a suitable transformer do it that way.

No you don't need a variable supply to check the switch point the regulator will take care of that it will just rectify whatever AC you throw at it until the battery comes up to voltage then it will start to skip half cycles and eventually stop firing the SCRs altogether. Obviously you don't want too large a transformer or the regulator will overheat. Another option would be to use say a 20 - 30 Volt transformer with a lamp in series, say a headlamp bulb then as the voltage comes up you will see it cut back by the lamp brightness.

Actually if you can find a reasonable size 12 VAC adapter (Wall Warts they are known as here) they put out a lot more than what it shows on the nameplate.

I use a 1 Amp DC one I found as a battery charger.

 

If you want to test it on your bike a good spot is put it on the steering damper bolt, all the connectors are right there.

[Kiwi_Roy]

... Two of my busted regulators had the alleged fault of giving 17-18 volts of charge but I can't see anything wrong and all resistor/diode/path tests show good....

 

That's reasonable. If you get 18V then all components work as they should, besides the regulator circuit wich obviously fails to open the rectifier at 14/15V. The internals get damp after some years, salt and other nasty things get in as well and as a result migration often starts bridging parts and even wires.

 

Putting mine into the kitchen oven at 120°C for some hours cured the 17V problem. Probably temporarily, but for this summer it works.

 

Hubert

I can't see how the potting could fail to keep moisture out (looked really tight everywhere and no cracks), but maybe you are right. I'll aim for some active tests.

 

BTW I tried re-soldering most components on the PCB's in case there was hard-to-spot soldering faults. My 40W Weller iron is far too weak for the diodes and external wire solders. I need to buy a cheap-ish heavier iron for them. Not sure how much power is needed. I remember my dad had a huge soldering iron that was probably capable of soldering tin cans, that one would have come handy

Are you trying to solder it with the diode mounted or in a vice, if so all the heat will soak away, I did mine with the diode sitting on a piece of wood 40 Watts should be enough.

[Kiwi_Roy]

Heres the Circuit......

It's posted elsewhere but this seems like a good spot for a copy.

Notes:

1. The part in light brown is just for the charging light and not required for the regulator.

2. Note 2 on the drawing shows a value of 575 Ohms, this was measured in-situ. The actual values I just checked were a resistor 820 Ohms on the component side and a trim resistor 4700 Ohms on the circuit side, I think what they do in the factory is instal the 820 Ohm resistor on the component side, install the circuit board then put it on their test rig and substitute different values of "Trim Resistor" until they get the right voltage.

This makes it possible to change the setting slightly if the regulator drifts off with age, you can simply uncover the Trim Resistor and change it for another value, Increasing the resistance will increase the voltage. If you decide to do this, change the value in small steps and do a decent run because it takes a while to reach the new voltage level.

Bear in mind changing the value of the trim resistor by 10% won't change the Voltage much because it has little effect on the combined resistance and also it's only one small part of a Voltage divider circuit. I sugest you solder a couple of short wires onto the circuit board so you can test different values easily or add a trim pot as I show in the photo/

The most voltage you will get is with the trim resistor off altogether.

If you can't get enough Voltage check your diodes, one may be open circuit.

 

I check my Voltage in a set pattern, by going for a ride then turning off the key before stopping so the headlight doesn't drain the battery down any.

 

I attached a photo showing where the trim resistor connects also shows how you could make an adjustable version. In hindsight I would use a 2K fixed resistor in series with a 5K trimpot, this should give an adjustment range of about 1 Volt

Regulator Schematic.pdf

[Kiwi_Roy]

And heres a Diode Fix...........

The Diodes seem to be a common problem, they melt the leads off and go open circuit.

When this happens you loose half of your charging, look at the waveform I show on the right hand side or the regulator schematic and imagine every second pulse missing.

Your regulator will now not be able to keep up the voltage although it will struggle along for years possibly

It's possible with a bit of luck to dismantle the regulator and fix the diodes but you run the risk of doing more damage to the circuit board getting it out.

Fortunately the diodes are directly connected to the outside world so you can simply connect another in parallel. Leaving the bad one in circuit is not a problem.

I like to use one of the potted bridge type rectifiers because they provide a convenient way of mounting to a heat sink (the bike chassis) without having to worry about accidental shorts to ground.

A 20 Amp bridge is big enough because at the most you only use 2 of the 4 diodes so it shouldn't overheat (higher current rating is better)

You could use just one diode to replace the bad one but remember it's case is alive (+12 V connected to the 30 Amp fuse F3) so it needs to be isolated from ground.

Regulator External Diodes.pdf

[luhbo]

Thank you very much. I'll put both ideas on this winter's Probably To Do List.

 

How much would you gain by replacing the diodes by two additional SCRs?

 

Hubert

[raz]

Are you trying to solder it with the diode mounted or in a vice, if so all the heat will soak away, I did mine with the diode sitting on a piece of wood 40 Watts should be enough.

They were mounted, I haven't tried lifting any of the PCB's but I have desoldered the diodes on one. I had to set the iron for 450*C which is not a good thing, but otherwise there was no chance I could desolder them. I don't recall ever having needed a larger iron than 40W for anything. Maybe the potting is very good at transfering heat.

 

Note 2 on the drawing shows a value of 575 Ohms, this was measured in-situ. The actual values I just checked were a resistor 820 Ohms on the component side and a trim resistor 4700 Ohms on the circuit side, I think what they do in the factory is instal the 820 Ohm resistor on the component side, install the circuit board then put it on their test rig and substitute different values of "Trim Resistor" until they get the right voltage.

For the record, the three regulators I've looked at all had a combined value of about 450 ohms, with the trim resistor being 2K4 in one case and 2K7 in two cases. This indicates all three have a component side resistor of 560 ohms (assuming E24) even though I've only confirmed that on one of them. Not that the exact numbers matter much, the important thing is that higher ohms results in higher voltage.

 

Personally I would aim for 14.6 volts if trying to trim the voltage, as that is the high end of the WHB spec (for my bike anyway) and you trade bulb life for good charging. If we only care about battery life, 14.9 volts should be the optimal (for a Hawker AGM, that is).

[Kiwi_Roy]

 

Personally I would aim for 14.6 volts if trying to trim the voltage, as that is the high end of the WHB spec (for my bike anyway) and you trade bulb life for good charging. If we only care about battery life, 14.9 volts should be the optimal (for a Hawker AGM, that is).

When you say 14.6 are you refering to the voltage with bike running, I find as soon as I shut of the voltage drops to around 12.85

[Kiwi_Roy]

Thank you very much. I'll put both ideas on this winter's Probably To Do List.

 

How much would you gain by replacing the diodes by two additional SCRs?

 

Hubert

You wouldn't gain anything, 2 diodes and 2 SCRs is as good as you can do for a single phase alternator.

Actually on my next winter project I am going to use a bridge rectifier (4 diodes) and just one SCR.

[raz]

 

Personally I would aim for 14.6 volts if trying to trim the voltage, as that is the high end of the WHB spec (for my bike anyway) and you trade bulb life for good charging. If we only care about battery life, 14.9 volts should be the optimal (for a Hawker AGM, that is).

When you say 14.6 are you refering to the voltage with bike running, I find as soon as I shut of the voltage drops to around 12.85

Yes, of course. My WHB says the charging voltage should be 14.0-14.6 volts and the Hawker battery spec. says you should never charge the battery with more than 15 volts (but virtually no limit on current, the more the better). Actually, reading that PDF again I think it says 14.7 volts would be better than 14.9 volts.

 

12.84 volts @25ºC indicates a very healthy and 100% charged AGM battery. The reading should ideally be taken after 6 hours rest, after charging

[gstallons]

FWIW..... a fully charged and stable 12v battery should read 12.65v. at rest.

[docc]

FWIW..... a fully charged and stable 12v battery should read 12.65v. at rest.

 

That's true for a wet cell battery, but the AGM battery from Hawker (like the original "Spark") is 100% at 12.84.

[Kiwi_Roy]

FWIW..... a fully charged and stable 12v battery should read 12.65v. at rest.

The instructions that came with my new Jell filled Yuasa calls for 12.8 - 13, 2 hours after charging, the most I have seen is 12.86

I'm happy with that I seem to have lots of cranking power.

[Kiwi_Roy]

OK Here is the latest....

Many thanks to Raz who helped with this project by offering constructive comments and proof read the drawing.

Between the two of us I think we have managed to capture all the salient points.

Roy

Regulator Schematic Sept 1 2010.pdf

[jrt]

Looks good Roy. Thanks to you and Raz!

Jason