AC regulation and active rectification circuit

[spec]

It's 18 poles and a bike wheel can rotate as little as say 3 times per second - 180RPM, although as stated I'm not expecting much watts until a higher speed.

The hub will give out 50V with little load at high speed. I know this by experience as it blew a LM2576 buck (45v rated) with a phone as a load. I don't know how much further the voltage will go, only that 50V is a minimum and 100V is what is talked about.

Yes it would, but there would have been ways to protect the chip

Unfortunately I don't have much kit to to more assertive measurements.

Got you. No problems. Ron or I will be able to sort it all out. I, and probably Ron, have a fair idea of the cycle alternator characteristics now.

Where are you going to put the circuit when it is built?

 

[ronsimpson]

ACharnley,

Please send the schematic for your circuit that blew up. And include any information on the inductor you used.
Pictures if you can.

 

[spec]

Hi Ron,

I may have some developments on the rectifier front. What is the maximum voltage you would like for the SMPS. I am assuming 38V at the nonent.

 

[ACharnley]

Originally I was just using one of these: **broken link removed**

But afterwards I found out the efficiency isn't great for 5v out, around 77%, however it does have a relatively high input voltage at 45v.

I then protected the same circuit using an LM317 in open mode, which drops up to 40v from vIn-vOut. This worked for protection, but the dropout was relatively high and I noticed some heat. The LM2596 requires min +7v in so with the dropout it frequently wouldn't supply enough juice.

The circuit I now use is;

**broken link removed**

It's not LM2576 as the title suggests but a CN chipset. It works from 5.5v and is more efficient, 90-95%, however the max input voltage in in 24V.

Presently I use a KBP307 Bridge (1.1v dropout) to rectify the voltage, then a single darlington TIP142 with zenier to ensure the chipset never goes above 24V. No heat is produced, but the TIP142 drops another 1V in operation. The TIP142 is in series with the chipset (no load, no wastage).

 

[ACharnley]

The idea for the MOSFET bridge is here: **broken link removed**

I have to admit to wondering why there's not an integrated MOSFET bridge chip. If the turn on voltage is too low will that mean they are off, so gaps in the voltage? I guess what I mean is, will they be less efficient than a KBP307?

Having searched the end of the earth, I finally found a twin MOSFET DIP-8 package with 20v gate handling - one I can solder. I've ordered 100! **broken link removed**

Not realising the gate voltage mattered, I now need to check the spec.

 

[ACharnley]

I just found this which is very similar to what I'm trying to achieve.

https://www.ktverkko.fi/~msmakela/electronics/dynamo5v/linear/index.en.html

Curiously the zeners dump the excess voltage >7v without the darlingtons, which is Shimano's design at 6V and which can get fairly hot.

The diode D3, I'm wondering if I can do away with by use of the switch mode instead of linear regulator. That has only required a 100uF capacitor in front to be stable, and I can live with that discharging into the hub.

 

[spec]

Rectification is done by the two power BJTs. The two diodes are for protection only.
I haven't checked the protection completely- all too difficult.

The power trans are MAG6332/MAG9942, only because I know them. They would be ideal but somethig a bit cheaper could be used on the day.
The Zeners are BZX84_18V
Q2 & Q2 are 100V VCE small signal.
The other two transistors can be any small signal.
The two protection diodes are 1N400x (100V)

Circuit Function

The circuit both rectifies and limits the output rectified voltage at 18V6 *2= 37V2. It relies on the low VCEsat of beefy power BJTs (see attached image) for good rectification efficiency.

Conduction starts when the alternator reaches around 600mV and stops when the output voltage reaches 18V6. Unlike the shunt clamp, which dissipates a load of power, this circuit dissipates little power.

Note that there may be major flaws in this circuit implementation- I havent checked- but it is the principle that counts.

https://www.farnell.com/datasheets/866312.pdf

ERRATA
(1) There is a major flaw in the circuit which needs sorting. No flaw in the principle though.

​

 

[ronsimpson]

I have questions: Does this seem right to you?

Using a 6V, 3W 0.5A Dynamo.

The voltage is clamped to a set level. (Zener Diode)

Clamp voltage is 3V you get 0.5A and 1.75 Watts. Half power at 4Km/h, Full power at 7Km/h.
Clamp voltage is 6V you get 0.5A and 3.5W, 7Km/h, 15
Clamp voltage is 9V you get 0.5A and 5.25W, 10Km/h, 20
Clamp voltage is 12V you get 0.5A and 6.9W, 12Km/h, 20
Clamp voltage is 15V you get 0.5A and 9W, 15Km/h, 25
Clamp voltage is 18V you get 0.5A and 11W, 18Km/h, 30

There is little to no increase in power at increased speed?

 

[spec]

If you are asking me:

Did you mean '..little to no increase in current...'

If so, yes it does make sense, otherwise in normal usage, when a filament lamp were connected to the dynamo, the lamp would blow at high speed.

Also any shunt clamping scheme would dissipate clossal power, as in my previous joke post.

 

[ronsimpson]

Did you mean '..little to no increase in current...'

Asking all....
If the voltage is clamped, there will be no increase in voltage, because of the clamp. And no increase in current, and thus no increase in power.
I have seen 6V alternators used to charge 6V batteries. The battery is some what of a voltage clamp.
I have seen 6V alternators used to charge 12V batteries.
I have seen car alternators used to make 120V. With out a low voltage clamp.

There are people that claim to have removed their 6V bulbs and replaced then with twice as many 12V bulbs. (bicycle)

Some of the bicycle USB chargers I have seen do not have a voltage clamp.
Some of the bicycle hub alternators have a built in Zener voltage clamp. (I do not know what voltage)
I have seen 6 and 12V external voltage clamps.

It appears that a alternator, clamped at a high voltage, should be harder to turn.

 

[spec]

Asking all....
If the voltage is clamped, there will be no increase in voltage, because of the clamp. And no increase in current, and thus no increase in power.
I have seen 6V alternators used to charge 6V batteries. The battery is some what of a voltage clamp.
I have seen 6V alternators used to charge 12V batteries.
I have seen car alternators used to make 120V. With out a low voltage clamp.

There are people that claim to have removed their 6V bulbs and replaced then with twice as many 12V bulbs. (bicycle)

Some of the bicycle USB chargers I have seen do not have a voltage clamp.
Some of the bicycle hub alternators have a built in Zener voltage clamp. (I do not know what voltage)
I have seen 6 and 12V external voltage clamps.

I too have seen a 6V alternator used on a 12V system

It appears that a alternator, clamped at a high voltage, should be harder to turn.

True, energy must come from somewhere, in this case the rider, but I suspect it is not significant in the order of things.
I still think clamp around 40V and use your SMPS- job done!

 

[ronsimpson]

I have a alternator in the barn that is 100 years old. I checked and it is frozen and will not turn. The question is: it it easy to turn if shorted out? Is it hard to turn if clamped at 40V and full current? Math make me think so.

Anyway, A 6V shorting regulator eats less power than a 12V shorting regulator. There fore a 40V Zener might get very hot. There seems to be some high voltage point where the current is down. One design I saw used a 68V 1/2 watt voltage limiting Zener doide.

I have seen several designs using a LM7805. The only limiting I can see is the leakage current in the input capacitor. (bad idea to over-voltage a cap)
One option is the simple LM7805 and use a 3 watt over voltage limiter at 8V.

There appears to be many 3W alternators on the market. I did not see a 6W alternator. (hub type) There are many 1.5W "bottle" type. Some have internal limiting.

 

[spec]

That all figures. The bulbs on UK were 2.5V at 300mA (I think). The bulb was quite dim at low speeds and got brighter up up a certain point and didn't get much brighter over a certain speed. It was a bit rough and ready and bulbs were always blowing, normally with black internal staining on the bulb glass.

Bottle alternators had a more powerful output but they reved at a hell of a rate and were a drag

The dynohub output was much weaker. There was no difference in the drag with light on or off. (wheel spun by hand)

I can't see why not clamp at 40V. I feel it would give the best USB PSU performance at all speeds. Am I missing something?

 

[spec]

https://www.bike2power.com/dynohub-powered-bike-phone-chargers

30 bucks- why mess about?

 

[ronsimpson]

I can't see why not clamp at 40V.

If they are really constant current, then 1A at 3V compared to 1A at 40V, it is better to clamp at low voltage. (3W vs 40W)
If we use the voltage doubler, the cap needs to be very large at low frequency/low voltage AND with stand 40 volts.
10,000. uF 65V is large, compared to 10,000uF 16V. or 10,000uF at 6.3V

 

[spec]

Ah yes- hadn't thought of that.

 

[ACharnley]

So, to get back on track...

The 40V buck IC's like the LM29** series typically start at Vin of 7+ volts. This won't be attained at low speeds. Also, the more common ones are only 80% efficient @ 5v out.

And really, I'm after the best method of preventing dropout voltage loss, which was why the original circuit using MOSFET's to replace the bridge was put forward. However, unless it's made a bit more complicated it won't withstand more than 20V. This roughly aligns with the MC buck chipsets (90%+ efficiency @ 5v) which have a max input of 24V.

The rest I intend to drop as heat. Probably it'll be on a decline anyway so won't be an issue.

I have since found a German guy that had success with two zenier 5W diodes instead of a smaller pair and two darlingtons. Simplifies it even further.

 

[ronsimpson]

I would like to see your circuit when done.
The MOSFETs in place of Diodes idea is complicated. I have done it with both diodes on ground, using 20 to 100s of amps.
The voltage doubler circuit like in post #35 only has current passing through one diode at a time. So the voltage loss is 1/2 as much.
The voltage doubler should work at a lower speed. I wish I had time to try it for real.
I agree with using Zener Diodes to keep the voltage down.

I worry that your PWM may have died because of the coil you chose. We should look at that.

 

[spec]

The approach outlined by the circuit of post 47 will achieve a low forward drop and has a lot going for it. It has a forward drop of 30mV at 1A!

It can probably be converted into a shunt clamp if that is the way required.

What is the best clamp voltage for the alternator? Would it be around 40V?

As you imply, reverse blocking MOSFETs are awkward

 

[ronsimpson]

The approach outlined by the circuit of post 47

Have you tried this?