Motorcycle Voltage Regulator ---- AC regulation

Thread Starter

abuhafss

Joined Aug 17, 2010
307
Any shunt regulator will involve components getting hot, since the shunt action has to dump excess generated power somewhere
Agreed, but the SCR in picture 4 or 5 in post #1 only gets warm, though it is also acting as shunt regulator.
 

Alec_t

Joined Sep 17, 2013
14,280
If you built the reg in pic 4 or 5, at what rpm did you test it? Did you have as much as 50V amplitude from the generator when not driving the reg?
My sim of that pic 4 reg shows about 20V (!) rms across a 6Ω load with a 100Hz (6000rpm) 50V amplitude input to the reg, and the SCR dissipating ~2.2W.
 

Thread Starter

abuhafss

Joined Aug 17, 2010
307
If you built the reg in pic 4 or 5, at what rpm did you test it? Did you have as much as 50V amplitude from the generator when not driving the reg?
My sim of that pic 4 reg shows about 20V (!) rms across a 6Ω load with a 100Hz (6000rpm) 50V amplitude input to the reg, and the SCR dissipating ~2.2W.
The reg. in pic 4 is the Chinese unit (I haven't yet built that but I do have a spare unit).
I have been checking physically at 6000-7000 RPM.
Here is my sim for pic 4.
2016-12-06 (2).png
I am wondering if this regulation provides any protection to the lamps. I think, the reg should maintain maximum 14-15V rms across the load at high RPMs.
 

Thread Starter

abuhafss

Joined Aug 17, 2010
307
Well, it's better than nothing. But with over 17V rms @ 6000 rpm the lamps will have a short life if they're rated for 12V.
In that case, my design at post #15 gives much better protection. The only issue is that triac is getting very hot.
Working of my design is almost similar to the Chinese except that shunting is being done by triac.
Do you think that the triac shunting on both positive and negative cycles is dissipating more heat? In that case, I think, the heat dissipation should be doubled as compared to the SCR, though in simulation the level of heat dissipation in both designs is almost the same.
I have attached the LT spice file of my design.
 

Attachments

Thread Starter

abuhafss

Joined Aug 17, 2010
307
I built the Chinese design with 3.9K (for 3.6K), 10K (for 9.1K) and 5.1V zener.

Without regulator the generator out was 9V rms at idle and 35V rms at 6000RPM.
With regulator and lights on, 6.8V rms at idle and 9V rms at 6000RPM.
Replacing the zener with 6V, I got 6.5V rms at idle and 18V rms at 6000RPM.
The SCR was barely warm in both conditions.
 

Alec_t

Joined Sep 17, 2013
14,280
What did you use to measure the regulator RMS voltages? A normal DMM is likely to lie, since it isn't receiving pure sine-wave input.
I note that in your sim file V1 has a 150sec delay parameter, so the frequency is fixed at 10Hz (600rpm) for the whole run-time. Is that intentional?
 

Dodgydave

Joined Jun 22, 2012
11,285
Your only regulating half of the output, the Thyristor only conducts on the Negative cycle and shorts the alternator out, on the positive cycle it isn't conducting, so the full voltage is allowed out, why dont you put the thyristor across the bridge rectifier output, so it conducts in both cycles, then adjust the zener to suit.
 

Thread Starter

abuhafss

Joined Aug 17, 2010
307
What did you use to measure the regulator RMS voltages? A normal DMM is likely to lie, since it isn't receiving pure sine-wave input.
I note that in your sim file V1 has a 150sec delay parameter, so the frequency is fixed at 10Hz (600rpm) for the whole run-time. Is that intentional?
Yes, I had used a DMM.
But, here are the measurements with a mechanical voltmeter:
Without regulator, 7V rms at idle and 75V rms at 6000RPM.
With regulator and lights on, 6V rms at idle and 15V rms at 6000RPM (zener 6V).

Regarding V1 or V2, I simply copied (without any change) from your sim which you forwarded me earlier. ;)
 

Thread Starter

abuhafss

Joined Aug 17, 2010
307
Your only regulating half of the output, the Thyristor only conducts on the Negative cycle and shorts the alternator out, on the positive cycle it isn't conducting, so the full voltage is allowed out, why dont you put the thyristor across the bridge rectifier output, so it conducts in both cycles, then adjust the zener to suit.
I was referring to this design which had a triac.
2016-12-03 (6).png
I will check your suggestion and get back.
 

Thread Starter

abuhafss

Joined Aug 17, 2010
307
why dont you put the thyristor across the bridge rectifier output, so it conducts in both cycles, then adjust the zener to suit.
Putting the SCR or triac across the bridge's output will make the design same as Sgt.Wookie's design see post #14. In that case, the bridge needs to be rated more than 15A which makes the size of actual regulator very big; and that is not feasible.
 
Last edited:

Thread Starter

abuhafss

Joined Aug 17, 2010
307
After playing with the Sgt.Wookie's design at post #14, I came up with following design with exactly same results.

2016-12-07 (2).png
2016-12-07 (1).png

The output (n002) is steady about 14V at high RPMs.
The heat dissipation of BT151 is about 24W.
For D1/D2, 2 x 10A diodes or 1 x 20A (TO-220) could be used.

Please let me have your valued comments. Tomorrow I am planning to test it.
 
Last edited:

Alec_t

Joined Sep 17, 2013
14,280
You still have the 150s delay for V1 (that was an error in the .asc file I posted).
Zoom in on your waveforms and you will see that the "steady about 14V" is in fact a series of 14V spikes with a 2.77V RMS value.
The average dissipation in the SCR is around 3.3W, with brief spikes of 24W.
D1/2 dissipate an average 2.4W each and wipe out half the AC cycle.
BTW, you can get average/RMS values by pressing the Control key and left-clicking the title of a displayed waveform.
You could improve the sim by replacing the 2 Ohm resistor with an inductor having appropriate inductance and DC resistance vlues.
 

Dodgydave

Joined Jun 22, 2012
11,285
So set the zener to give that voltage then!

Better still, use a TL431 zener and a variable resistor to give the precise voltage.
 

Alec_t

Joined Sep 17, 2013
14,280
I can't replicate the post #35 results. I'm seeing an average dissipation of ~60W in the Darlington at high reves, with L1 resistance = 2 Ohms. What resistance did you assume for L1?
It would help if you assigned labels to each node that so that we know which waveforms you're plotting, and if each time you change the circuit you post the new .asc file.
I was thinking that spikes are repeated so quickly that the lamps would have 14V.
No. Incandescent lamps have thermal inertia and "see" the RMS voltage.
 

Thread Starter

abuhafss

Joined Aug 17, 2010
307
So set the zener to give that voltage then!

Better still, use a TL431 zener and a variable resistor to give the precise voltage.
But I feel that will not make a big difference because the regulated AC voltage keeps on increasing with the RPM.
 
Top