HELP needed with TL783 Regulator Circuit

[alec_t]

On the circuit. FETs running in linear mode like this can't be paralleled with just a small source resistor like a BJT. The threshold voltage variation can be several volts and the gain and thermals don't like it. It looks ok in spice because the thresholds match, but if you put 2 or 3 diodes in series with one gate you can see the problem.

Point taken. My bad. Sorry, GoGreen, ignore my circuit. Back to the drawing board

Edit:
Ronv makes a very valid comment that a linear regulator would need to be set below the minimum of the ripple (say 60V). So if you used one the existing dump load would never activate. In effect the regulator would become a new dump load since it would have to dissipate a large amount of power. For a load current of 14A and an input voltage of 130V the dissipation would be (130-60)*14 = 980W. Hence 2 x 300W FETs alone wouldn't cut the mustard. You would also need some hefty resistors to share the power dissipation.
As Ronv said, an SMPS would be a better bet.

 

[alec_t]

Here's a revision of the circuit. Current balancing between FETs is now achieved by use of a long-tail pair Q3,Q4 (and works even for low values of R9,R11 in the sim).
R9,R11 have been increased in value to help share the heat load. I have deliberately used two different FET models to check that Vgs threshold variation doesn't affect current balancing.
For > 80V in, 60V out, Load current = 10A, power dissipation is as follows:
Load 600W
M1 or M2 up to ~200W,
R9 or R11 ~50W,
R3 up to ~7W,
TL783 (modelled as LM317) up to ~3W


Edit: I think you'd be pushing your luck to try this circuit with the FETs you posted the link to. You'd probably need more than 2 FETs 'parallelled'.

 

[GoGreenElectricsLTD]

Here's a revision of the circuit. Current balancing between FETs is now achieved by use of a long-tail pair Q3,Q4 (and works even for low values of R9,R11 in the sim).
R9,R11 have been increased in value to help share the heat load. I have deliberately used two different FET models to check that Vgs threshold variation doesn't affect current balancing.
For > 80V in, 60V out, Load current = 10A, power dissipation is as follows:
Load 600W
M1 or M2 up to ~200W,
R9 or R11 ~50W,
R3 up to ~7W,
TL783 (modelled as LM317) up to ~3W
View attachment 71859

Edit: I think you'd be pushing your luck to try this circuit with the FETs you posted the link to. You'd probably need more than 2 FETs 'parallelled'.

60v is way too low !... The turbine needs to be making at least 85v to produce 600w, 90v for 1Kw. Below 85v and the turbine just would not be able to make the power even to make one inverter pay its way. Up to 70v the turbine is doing little more than producing enough power to run a stereo. But then things realy start to take off and within the space of another 15v (85v) it will go from 100w - 600w output, another 5v (90v) and we hit 1Kw output.

I bought an 82000uF 100v Cap off Ebay a couple of days ago. So im going to stick that in when it arrives. Then its just a waiting game for the next storm when there are winds strong enough to realy drive the turbine hard. With any luck that will cure most of the ripple and get it to remain under 100v peak. If the low point in the ripple remains above 85v, then problem solved !. My concern is that it may not do so.

 

[alec_t]

60v is way too low !..

But if that's what the minimum of the ripple drops to then that's what a linear regulator would need to be set at for ripple elimination. Which is why it was suggested earlier that a linear reg probably isn't what you need . Good luck with the cap.

 

[GoGreenElectricsLTD]

But if that's what the minimum of the ripple drops to then that's what a linear regulator would need to be set at for ripple elimination. Which is why it was suggested earlier that a linear reg probably isn't what you need . Good luck with the cap.

Enter the large Capacitor !...... I was already aware of the point you make. Which is why I am hopefull the large capacitor will raise the low end of the ripple. However if we could fix all the problems associated with ripple, then why did anyone bother invent the regulator ?.

 

[ronv]

Enter the large Capacitor !...... I was already aware of the point you make. Which is why I am hopefull the large capacitor will raise the low end of the ripple. However if we could fix all the problems associated with ripple, then why did anyone bother invent the regulator ?.

You already have a pretty good regulator in the inverter. The problem is once the dump load turns on you don't have anymore power for it. Think about 95 volts and 4 ohms - 2200 watts from a windmill rated at 1000.

I tried using a larger resistance value to do as you have said and soften the rate at which it works. However this only resulted in it taking far too long to knock back the speed of the turbine during gusts. Also the dump load even though its rated at 400w got extremely hot !

The best you can do with your system is to use the largest resistor you can that keeps everything together and set the trip point for the dump load as high as you safely can. The larger resistor should have produced less heat. (p=I^2R).

You will have the same problem with the bigger inverter. It will just happen when the wind is blowing harder.

 

[GoGreenElectricsLTD]

You already have a pretty good regulator in the inverter. The problem is once the dump load turns on you don't have anymore power for it. Think about 95 volts and 4 ohms - 2200 watts from a windmill rated at 1000.



The best you can do with your system is to use the largest resistor you can that keeps everything together and set the trip point for the dump load as high as you safely can. The larger resistor should have produced less heat. (p=I^2R).

You will have the same problem with the bigger inverter. It will just happen when the wind is blowing harder.

Actualy your wrong !

Even a very small wind turbine can produce far in exess of its rated power for a split seccond. There is a certain amount of inertia in the rotating blades. When the dump load is first connected its job is to quickley reduce the speed of the blades. In that brief moment the current levels can be very high indeed !.. To give you an idea of how high, the solid state relay I have controling the dump load has to be rated at 80 amps !. It is entierly possible that for just a brief moment the power within that circuit could exeed 3Kw. Think about it P=V/RxV !. So the current in the circuit = 110v / 4 Ohms x 110v = 3025w. But this IS what we want it to do, this is how the braking system works. In fact the manufacturers of the turbine have suggested I use a 2 Ohm load !. If the resistance was too large more of the power would end up in the dump load and less of it in the inverter.

A modern family car may only have a 95Kw engine and takes 10 secconds to accelerate to 60mph. But the 4 large disc brakes usualy fitted to a modern car are capable of absorbing energy at a rate which 4 or 5 times the output power of the engine. Which is why it only takes 2-3 secconds to stop from that speed. However the heat dissipation ability of car brakes is poor. If you fitted less agressive drum brakes they would be in-effective. If you continued to drive the car in the normal way, braking just as if you had disc brakes fitted, the drums would get very hot and most likely catch fire.

My dump load will absorb large amounts of power for very short periods, such as strong wind gusts. Its low impedance ensures that it will slow the turbine rapidly. Thankfully strong wind gusts are not constant, so the heat dissipation ability of the dump load is not requied to be huge. Thus its 400w rating is more than enough. If I were to increase the value of the dump load resistors, it would have the same effect as fitting old drum brakes to your car. It would be in-effective at reducing the turbine speed and would likely overheat, possbly catching fire too.

The same problem will NOT occur with a larger inverter, because there is a point where the turbine simply cannot produce any more power, no matter how hard the wind blows. If I make the inverter larger it will absorbe all the available power when the turbine is running and the dump load will be redundant. However it will still need a dump load, to restrain the turbine incase of inverter failure.

 

[alec_t]

Actualy I have around 40v of ripple right now. From 70vdc to 110vdc

One possible cause for the ripple is not so much the dump controller's response time, more likely I think is that the controller has considerable hysteresis between its cut-in threshold and drop-out threshold, and the 4Hz ripple is due to the turbine acceleration/deceleration times between those thresholds. Does the controller have just the one adjustment? Or can you set upper and lower thresholds (or cut-in plus hysteresis)?

 

[GoGreenElectricsLTD]

One possible cause for the ripple is not so much the dump controller's response time, more likely I think is that the controller has considerable hysteresis between its cut-in threshold and drop-out threshold, and the 4Hz ripple is due to the turbine acceleration/deceleration times between those thresholds. Does the controller have just the one adjustment? Or can you set upper and lower thresholds (or cut-in plus hysteresis)?

Well, your on the right track with that thought. But like many power producing items with a large coil such as a transformer. The voltage can be easily dragged down by a heavy load and the inverter is certainly that !. The wind turbine speed could not possibly change in 0.25 of a seccond. Yes there is a degree of hysteresis, but that in itself is not a problem. The delay in the controler's ability to react to the changes in voltage is a problem, but I have not found an alternative circuit to do the job more effectivley. It doesnt need a setting point for upper and lower limits as like I said although there is some hysteresis, its not a problem. Infact its actualy the basis upon which most dump load controllers operate.

 

[alec_t]

The voltage can be easily dragged down by a heavy load

If a 4Ω load (ignoring the inverter loading) drops the voltage from 110 to 70 almost instantaneously with no appreciable change in turbine speed that implies the turbine (plus wiring) source impedance is ~ 2.3Ω. Seems a tad high. Is that typical for turbines of the size you have?

Edit:
Is the 4 Ohm load resistor readily detachable from its controller? If so, we might be able to come up with a faster control circuit for it.

 

[GoGreenElectricsLTD]

If a 4Ω load (ignoring the inverter loading) drops the voltage from 110 to 70 almost instantaneously with no appreciable change in turbine speed that implies the turbine (plus wiring) source impedance is ~ 2.3Ω. Seems a tad high. Is that typical for turbines of the size you have?

I am sorry I couldent tell you the answer to that question. Its three phase I know that and it has around 50ft total of 6mm2 cable with a bridge rectifier around 15ft from the turbine head. Perhaps I will do an impedance measurment next time there is no wind.

Also, I didnt say there was no change in turbine speed !.. What I said was that its not going to change the speed in 0.25 of a seccond. But certainly within 1 seccond it does do so.

 

[alec_t]

But certainly within 1 seccond it does do so.

Ah, that makes a difference. So the impedance is somewhat < 2.3Ω.

 

[ronv]

Mr. RB stated the problem. It is due to the bang bang nature of the load dump and the maximum power point of the inverter. Changing the dump to a more linear mode might fix it, but why bother if the only power lost is from a gust. That is what the dump is there for. That and the inverter being offline. Kinda like having your foot on the gas and the brake on.

 

[GoGreenElectricsLTD]

Mr. RB stated the problem. It is due to the bang bang nature of the load dump and the maximum power point of the inverter. Changing the dump to a more linear mode might fix it, but why bother if the only power lost is from a gust. That is what the dump is there for. That and the inverter being offline. Kinda like having your foot on the gas and the brake on.

I think your starting to understand the operation of the dump load. Your correct that the dump load has a Bang-Bang type of operation and that is the way it is supposed to be. Many dump loads have even less impedance than the one I am using !. The real cause of my problems are that the controller cannot operate any faster than 4Hz... If I could fix that then even a small capacitor with no regulator would do the job. But alas, I cant fix that issue because the manufacturer of the controller is in China and despite my best efforts to explain the problem to him, its like talking to a brick wall !!!.. So I have to find a different solution and I belive a whacking great Capacitor and a Voltage regulator is the answer !.

 

[alec_t]

The real cause of my problems are that the controller cannot operate any faster than 4Hz...

Hmm. Not sure that's the case, having done some more sims. I simulated switching the 4Ω dump load in and out at several kHz and the result was the same amount of ripple but at the higher frequency. The real problem as I see it now is that the source impedance drops the volts down too much when the dump current flows. That's fine for the braking effect but not good for ripple. I find myself in agreement with ronv (post #33) and Mr RB.

 

[ronv]

Do you have a link to your turbine and inverter?

The cap will help a bit. The voltage will only go down 67% of where it is going now in 0.272 seconds. Of course the turbine won't stop as fast nor recover quite as fast either for that matter. So maybe that's okay?

Alec's idea for a pwm load would be a fix if we knew more about the turbine so it could be made to balance the load to the voltage and the wind speed.

 

[GoGreenElectricsLTD]

Do you have a link to your turbine and inverter?

The cap will help a bit. The voltage will only go down 67% of where it is going now in 0.272 seconds. Of course the turbine won't stop as fast nor recover quite as fast either for that matter. So maybe that's okay?

Alec's idea for a pwm load would be a fix if we knew more about the turbine so it could be made to balance the load to the voltage and the wind speed.

PWM only works as far as bleeding off some of the power to the dump load. So that the higher the voltage the longer the ON pulse, lower the voltage shorter the ON pulse. This is exactly how a lot of dump load controllers do work. My 58v SWEA dump load controller (which I dont use) makes a fast buzzing noise at around 1Khz when the voltage goes way beyond its set point. But if its only just over the mark at say 59v it makes more of a rapid ticking noise.

However I cant use the SWEA controller because my turbine operates at much higher voltages.

I think you can forget trying to ballance the load and turbine, its simply not possible !.... You would need to employ some fuzzy logic, because of the ever changing voltage, current and inductance properties. Setting up the inverter MPPT points and getting the best out of the turbine is no simple task and takes a lot of patience.

 

[ronv]

PWM would not be difficult, but without power curves for your turbine it might be.
If your dump load is not good for 1KW it is to small. Should the inverter shut down in a high wind bad things may happen.
Running your dump load where it is now is probably hard on your turbine.
Do you have the curves? Or is it a home brew?

 

[GoGreenElectricsLTD]

PWM would not be difficult, but without power curves for your turbine it might be.
If your dump load is not good for 1KW it is to small. Should the inverter shut down in a high wind bad things may happen.
Running your dump load where it is now is probably hard on your turbine.
Do you have the curves? Or is it a home brew?

The Turbine power curves are not important and have no connection with the purpose of building a dump load controller. The only thing a dump load controller needs to know is what voltage to start limiting !.... yes a PWM circuit would work and SWEA make a very good one that can handle between 52v and 58v dc and up to 30A current to the inverter. It will also operate with a dump load of just 1 Ohm and controlls the voltage to the tubine with good precicion. I have a photo of one which I used with an older turbine that operated at a lower voltage and higher current, C2 capacitor is mounted seperately as it is too large to fit on the circuit.....see below.

Unfortuntely I have been unable to locate a similar circuit which can operate at higher voltages.

 

[alec_t]

Looks like it shouldn't be too difficult to uprate that controller if we can find out a bit more about it. The FET is only rated at 100V so could be swapped out. Can't read the numbers on IC1 ?
Don't suppose you have a schematic for the controller?

Edit: IC1 seems to be a 10F204 (PIC micro).
BTW, Spice thinks that with the present bang-bang dump load and one inverter you'd need ~ 0.4 Farads smoothing cap to hold ripple within 10%.