The Great KB Regulator & Rectifier project.

[davereid]

Hm. Will this work on single phase alternators (with two less rectifier diodes of course)? Most of my bikes are so old that multiphase hadn't been invented.

Yes, it should be fine. you could even run it off a 24 volt battery if you wanted to do so for testing purposes.

[pete376403]

Couple of questions, please.
Looking at this http://www.farnell.com/datasheets/99091.pdf the Vishay 36MT has a number of voltage codes from 10 thru 160. Which one should we be using?
Also the wattage of the resistors?

Any other component values to be aware of?

Sorry if this is basic electronics but as I said before I haven't a clue about designing circuits

[jonbuoy]

Diodes - standard (i.e. slow) diodes will not be OK on a high frequency system, unless you don't mind wasting power in them.

I'm using them now - and I'm not the only one no problems/huge losses.


The type of alternator in most bikes is a low rpm permanent magnet alternator - unlike a high rpm car one with an adjustable stator field.

What this means is that as the engine RPM changes, the output frequency and power of the alternator changes with it. It's actually a fine balance between having enough charging power at idle to ensure that with ful load you don't run out of juice if you sit idling for 15 mins and making sure you don't generate a bazillion watts of power at high rpm. All of this is solved with a car style alternator but as previously mentioned the complexity and costs are keeping them out of bikes at the moment.

So the voltage AND the frequency of the ripple on the alternator output will vary with engine RPM and load.

I have one with a variable field coil on my 750, frequency of the ripple? You mean the frequency of the AC sinewave? Which will affect the ripple coming from the rectifier diodes.




Not sure what you mean here? You could take the supply for the chip from the battery, but then if the battery is flat the chip will not enable the charging system e.g. if you push start the bike. Also unless you add more complexity with a shutdown circuit the chip will draw power from the batttery while the bike is off. It could all be done, but Dave has designed in a nice simple solution and the cap size given will be plenty, as the chip is not drawing huge amounts of current.


Look at the diagram - Your trying to smooth the entire output of the rectifier not just the supply to the IC, with that one wee cap.

You might be surprised how close to "DC" the output is! It's not battery like, but it's not exactly swinging up and down by 30V. Older style mains powered automotive battery chargers used to just have thyristors that would allow continuous pulses of rectified AC going from 0 to ~14V to charge a battery.

Stick a scope on it and see how close to DC it is.

HTH

EDIT: Dave has been drawing while I composed my thoughts - chip power supply issue fixed! :-) Another potential improvement is to filter the chip suppl via some series impedance, meaning the cap will have more effect.

Anywhoo good luck with the project, be interesting to see how it works out.

[davereid]

...the Vishay 36MT has a number of voltage codes from 10 thru 160. Which one should we be using?
Also the wattage of the resistors?

Any other component values to be aware of?

A voltage code of 20 or greater will be fine.

All resistors can be bog-standard 1/2 or even 1/4 watt resistors, with the exception of the 300 ohm resistor.

This resistor is only ever carrying current when the output transistor is on. So it will have the battery voltage plus any losses across the transistor. Say a max of 16 volts. This means a 1 watt resistor should suffice.

But, I would suggest a 5 watt resistor here, as it will (or a few nanoseconds !) carry a much higher load as the transistors switch.

The TIP41C should be selected.. 41A/B has a lower voltage rating.

[notme]

Anywhoo good luck with the project, be interesting to see how it works out.

Can't work out how to easily quote the quote of a quote.....but anyway to address your concerns:

The rectifier diodes - my point is, using diodes that are too slow for the AC you are rectifiying will be lossy. I don't know for sure the frequency of the AC from most bike alternators, but it's something to be aware of for the constructors if the rectifier starts overheating and popping. If it doesn't, there's no problem.

Regarding the voltage and frequency changes in the alternator output - I'm not sure of your understanding here, and I'm only trying to help clear it up. I am talking about a permanent magnet, low RPM type alternator like on most bikes. If you have a wound field alternator (like a car) on your bike, then non of that applies. In that case, the voltage will be regulated by the alternator, the frequency of the ripple will change with RPM, and all of the info in this thread is useless to you because the RR we are talking about won't work on your system.

With the smoothing cap - my mistake - I have seen so any similar circuit diagrams that I "saw" a resistor where there wasn't one. If there was a filter resistor in the feed path to the chip then the cap is isolated by that resistor and the chip will be the only thing drawing on it. Dave has changed the circuit now so that doesn't apply, as the chip has a much better filter, the battery.

Finally, regarding what comes out of the alternator being DC - and again I am talking about the alternators on most bikes nowadays, so that precludes Ixion's older single phase jobbie - what comes out is pretty much DC, it just depends on your definition. Is DC 100% oure zero ripple? In that case, you will only ever have DC from an electrochemical reaction such as in a battery. The "DC" that comes out of telecoms rectifiers is allowed by the various standards to have around 20mV ripple on it. Is that still DC? The "DC" out of your alternator will have as a generalisation something like a volt of ripple.

One volt out of 14 means about 7% ripple.....and to sum it up:

- true sinewave AC goes below zero the same amount as it goes above zero

- rectified AC goes to zero but not below, i.e. the negative halves of the cycle have been flipped upside down

- DC is a steady voltage that does not vary cyclicly. The disclaimer on that last statement is that most people realise that almost all DC has some ripple on it, if it's not an appreciable fraction of the total magnitude then we just call it DC.

I could scrible some diagrams if that would help, hopefully the descriptions are clear enough though?

[davereid]

Anywhoo good luck with the project, be interesting to see how it works out.

Thanks jonbuoy...
I'm interested in some of your comments...

This is a community project, and I cheerfully confess to this being my first power supply design venture.

So your design assistance is welcomed !

Diodes - standard (i.e. slow) diodes will not be OK on a high frequency system, unless you don't mind wasting power in them.
I'm using them now - and I'm not the only one no problems/huge losses.


What kind of losses do you consider huge ?

I looked at this and thought the losses could be tolerated. In fact, to a certain degree I need to find about 0.5 - 1.5 ohm or equivalent in losses in the stator, wiring and rectifier or the circuit would need a ballast resistor.

As the losses will increase with frequency (rpm) as will output, this was exactly the characteristic I wanted. But Vishay don't publish reverse recovery time data, so I just treated it as being typical of 35 amp rectifier diodes.

Can you suggest a better product that all constructors will be able to access ?

I have one with a variable field coil on my 750, frequency of the ripple? You mean the frequency of the AC sinewave? Which will affect the ripple coming from the rectifier diodes.


This is a 3 phase, full wave rectifier. The ripple will be at 6 times the alternator frequency, so I thought it would be pretty negligible.

Have I F'd up here ?

Look at the diagram - Your trying to smooth the entire output of the rectifier not just the supply to the IC, with that one wee cap.


With the original diagram, with the output transistors OFF, the only thing being supplied by the capacitor was the IC, so I figured it was adequate. Once the output transistors came on, the battery did the levelling.

In the new design, the battery does the job of power supply the entire time, so its pretty irrelevant.

[CookMySock]

Slow diodes are only a problem in high frequency power circuits, when they spend a lot of time partly switched on or off, thereby increasing their dissapation. I don't think this is important here, unless you are rectifying 200kHz.

Yeah he is kinda right about the cap smoothing the entire output of the rectifier. It probably wont last long or do a good job - put a low value resistor in the supply line before it, just for the supply to the chip. I don't think the cap will do jack anyway, as most of these chips have really good regulation, and you can supply them with a crappy supply of any sort and they will make it turn to gold for you.

Interesting !

So who has built one? Can anyone design a PCB for it ?

Steve

[jonbuoy]

With the original diagram, with the output transistors OFF, the only thing being supplied by the capacitor was the IC, so I figured it was adequate. Once the output transistors came on, the battery did the levelling.

In the new design, the battery does the job of power supply the entire time, so its pretty irrelevant.[/QUOTE]

Cheers Dave - haven't been trying to piss in anyones cornflakes on this - just putting my oar in I've never done a comparision back to back with schottkys or the like - I replaced the originals with two square pack rectifiers - nothing got hot battery charged as well as it ever did so I thought the job done. Originally the bike had a relay regulator to control the rotor voltage I looked at a few circuits then found an single chip alternator regulator in a three pin package that looks like it'll do the trick, they are obsolete so I need to put in an order with Dial Electronics. I did do some frequency measurements on my firestorm when I had reg and stator fail - I don't remember what frequency it was running at, stator voltage was 26VAC at idle or flat out so I can only assume it was being clamped by the rect.

I suspect every bike will be a bit different in voltage and frequency so a universal one would be handy. I can't help thinking noise and ripple are going to be the biggest headaches.

[Warr]

Originally I had thought of using a diode array out of a car alternator.. But last night had decided to just use 2x 35A 400V Bridge rectifers from JayCar, only $5 each. And this would give me a spare phase should 1 die!!
Then this morning reading about the frequency issue?? is it??. So I started searching for someone in NZ who supplied 3 phase Rectifiers. **

Came across this site, for them that want to see how it all works check out the simulation. Dont worry that they are SCR's, with "Firing Angle" =0 they are diodes.
http://services.eng.uts.edu.au/~venk...htm#simulation

edit - Link not working now Cant access site....

** Still didnt find anyone yet.

[davereid]

DB might be on to it with regard to rectifiers.
We will only be running at 10k-15k max.. audio frequencies, unlikely to be an issue. If we were trying to do things at 500k, maybe we would have a problem, but for $10 just try your cheapo pair of jaycar rectifiers.

[notme]

Thanks jonbuoy...
I'm interested in some of your comments...

This is a community project, and I cheerfully confess to this being my first power supply design venture.

So your design assistance is welcomed !

Diodes - standard (i.e. slow) diodes will not be OK on a high frequency system, unless you don't mind wasting power in them.
I'm using them now - and I'm not the only one no problems/huge losses.


What kind of losses do you consider huge ?

I looked at this and thought the losses could be tolerated. In fact, to a certain degree I need to find about 0.5 - 1.5 ohm or equivalent in losses in the stator, wiring and rectifier or the circuit would need a ballast resistor.

As the losses will increase with frequency (rpm) as will output, this was exactly the characteristic I wanted. But Vishay don't publish reverse recovery time data, so I just treated it as being typical of 35 amp rectifier diodes.

Can you suggest a better product that all constructors will be able to access ?

I have one with a variable field coil on my 750, frequency of the ripple? You mean the frequency of the AC sinewave? Which will affect the ripple coming from the rectifier diodes.


This is a 3 phase, full wave rectifier. The ripple will be at 6 times the alternator frequency, so I thought it would be pretty negligible.

Have I F'd up here ?

Look at the diagram - Your trying to smooth the entire output of the rectifier not just the supply to the IC, with that one wee cap.


With the original diagram, with the output transistors OFF, the only thing being supplied by the capacitor was the IC, so I figured it was adequate. Once the output transistors came on, the battery did the levelling.

In the new design, the battery does the job of power supply the entire time, so its pretty irrelevant.

Diodes - there are lots out there, such as an MUR1560 @ under $2 each for small quantities - or just choose any SCHOTTKY power diode. The power loss may not be an issue but that was one of the things that stuck out when I was reviewing the design (too many design review sessions at work i suppose). The first constructor will have the answer for us no doubt :-) .

[notme]

......** Still didnt find anyone yet.

Don't buy a 3ph rectifier block - buy 6 diodes....mucho cheaper and more bits to hand solder together!

You could also get dual diodes - 2 in one package. Just as much soldering, but less physical bits to bolt to heatsinks.

[Warr]

Anyone make any progress on this project ?
I currently have no need at present, no bike burnt to a crisp on the side of the road.. I dont think it was electrical !!

[CookMySock]

I think no guinea pigs have vounteered. There was one rectifier query some weeks back, and I directed him here, but he wouldnt take the bait.

My bikes are much too new to muck around with.

Steve

[Squiggles]

Ill do it when my exams are over (3 weeks), have spares of all the components i could kill if i screwup