HELP needed with TL783 Regulator Circuit

[GoGreenElectricsLTD]

I have an existing adjustable TL783 based PSU which I bought from Ebay. I intend to use this to limit voltage and provide a smooth power output. I hope to set it at 90vdc. The input power source being a wind turbine rages from Zero vdc to 130vdc. However there is a dump load controller set to 95vdc with a large capacitor and this diverts much of the unwanted power to a large resistor bank. The TL783 will be used after the capacitor and dump load.

Even so when the dump load is in operation there is significant ripple produced by the dump load pulsing on-off-on-off at 4Hz. This ripple is still faily large and needs to be smoothed out. The purpose of the regulator is to stabilise the final output to a steady DC ceiling when at maximum power. Anything below the 90v threshold is not a problem as the dump load will not be activating below its 95v set point.

I believe the TL783 PSU circuit could do this. But I would also need to connect a couple of 300w MOSFETS to handle the high level of power, I think these will work -

**broken link removed**


Please could someone help me to design a small ADD-ON circuit to work as I have described above ?.

Thanks
Rob

 

[ronv]

FETs can be tricky to run in parallel. Something like fig. 20 might be what you want. What is the maximum current?

https://www.electro-tech-online.com/custompdfs/2013/03/tl783.pdf

 

[GoGreenElectricsLTD]

The maximum operating current for the turbine at full power should be around 11A on paper, however with wind turbines its a good idea to be conservative as they can suddenly go much higher than expected. So the maximum current at the present is around 7A but that is likely to rise to 14A when I can afford a seccond inverter. Untill then the power difference is sent to the dump load and it is the controler which is causing the heavy ripple in the supply.

I already have a TL783 regulator circuit which I think can do the job very well. However as I explained at the top of this thread, the TL783 on its own has a very low power handling ability of around 40 watts. I am looking to regulate a supply which eventualy could exeed 1Kw !!...

I looked at Fig 20, it suggests using some TIPL762 transistors which are only 120w each. I have built regulator circuits in the past using MJ2955's and 2N3055's, but that is old fassioned tech and ide need a lot of them !. Actualy I have a brace of 3 MJ2955's on the shelf fitted to a large heatsink. To make the circuit work as intended and have some headroom I would need 10 of them, thats both bulky and expensive.

However I realy do want to do this with some high power MOSFETS like the ones I have listed in the link above. But im not that clued up on connecting MOSFETS in a regulator and realy need some help in the form of a simple circuit diagram.

P.S.

I do have another question though ?...... My smothing capacitor is presently 4700uF, when I remove this the voltage fluctuation problem becomes much worse and adds a further 10v to the ripple. So this got me thinking if I fitted a monster capacitor of say 68,000uF and got that ripple right down to lets say 10v peak-peak (90vdc -100vdc). My Inverter would then consume all power up to 85v effectivley eliminating the need for the regulator up to that point. My dump load then takes most of the power produced over 95v. So.... what would the power rating need to be for my regulator circuit when set at 90vdc ?.

I realise this is confusing and its not meade any easier by the fact that the voltage, current and power are all constantly variable. However what is constant is that the speed of the turbine (which is what I realy want to control) is direclty linked to the voltage produced from it. So if I can limit the voltage it produces, then I am also limiting the speed at the same time. The best way to limit the voltage is to apply a load and controlling the amount of that load is the hard part.

Rob

 

[ronv]

Actually I don't think I read you first post close enough.

But. Why do you care about 4 volts of ripple. It would seem to me the inverter would have a large input voltage range and the load dump is only there to keep it from going to high under low load.

 

[GoGreenElectricsLTD]

Without the capacitor there is around 40v of ripple and yes I do care about that !... The small 4700uF capacitor brings that down to around 30v of Ripple. I am hopefull that a much larger cap of 68000Uf or even 82000uF will bring this down to 10v of ripple.

Whilst the inverter can cope with almost any ripple and be un phased by it. It causes problems with its ability to correctly adjust its loading on the turbine. I have found that during very strong winds when the dump load is operating, the output from the inverter is actualy only at around half power. When the wind backs off a bit then the power output goes back up again. When I put a scope of the feed to the inverter. I was able to see that this was caused by the pulsing effect of the dump load. So the quest is how to smooth out that ripple when the dump load is in operation.

 

[GoGreenElectricsLTD]

Actually I don't think I read you first post close enough.

But. Why do you care about 4 volts of ripple. It would seem to me the inverter would have a large input voltage range and the load dump is only there to keep it from going to high under low load.

I am sorry to point this out, but I think that you failed a seccond time to read my initial post correctly. I said that the ripple was at 4Hz thats the frequency of the ripple, not the voltage !.

 

[alec_t]

If you could split the dump load into sections and switch them in individually as needed you might have a smoother control (at present its all or nothing) and possibly less ripple.

 

[GoGreenElectricsLTD]

If you could split the dump load into sections and switch them in individually as needed you might have a smoother control (at present its all or nothing) and possibly less ripple.

Unfortunately that would be very complex and at 0.25 of a seccond to perform the switch, timing between the different switches could become a problem. I think a simple voltage regulator is the best answer. I just need a simple circuit which can incorperate a TL783 and at least 2 x 300w Mosfets.

 

[ronv]

Hmm.
Maybe a few questions:
What are the input/output specs on the inverter?
Can you adjust the trip point on the load dump or is it fixed? What is it's load in amps?
Where does the 4Hz. come from? Is that from the turbine or the dump?
What is the actual voltage of the lowest point of the ripple?
Is the inverter load constant or can in vary while the load dump is on?
Knowing this we can see if a regulator would work, but it would need to be set below the lowest voltage plus some head room so you might see only 55 - 60 volts or so into the inverter. From that standpoint the caps are much better, but I would question if 4700 Ufd. would be anywhere near high enough to filter out 10 volts worth of the 40 volts of ripple at 11 amps that you have. So do you think something else is going on?
The bad thing about the regulator is that you throw the power away (as heat) all the time. So lets say we set it at 60 volts to get rid of the ripple. When the turbine is putting out 90 volts and the load from the inverter is 11 amps you are loosing 330 watts. From this standpoint Alec's idea of matching the load might be easier as the fets could just be on full of off rather than in linear mode. But let me look around for a switch mode supply. I think that would be better than the linear.

 

[GoGreenElectricsLTD]

I had though that my initial post explained most of this already... but here we go anyway !.

The inverter can handle a maximum input voltage of 150v, but like I said I am not worried about this because the turbine will never reach that voltage even if allowed to run free !. The main point of concern is the potential for the turbine to overspeed and damage the bearings.

Yes I can adjust the trip point.... It is digitaly programable.

the 4Hz is produced by the dump load being switched in and out. As the voltage climbs above the preset 95v, the dump load is switched in and this then pulls the voltage down sharpley slowing the turbine speed. However it has a tendency to overshoot the mark because the digital circuit cannot operate any faster than 4Hz. So the dump load remains connected for too long bringing the voltage down too far before disconnecing again. Once it does disconnect the same effect then happens with the rising voltage and it will again go past the mark and rise to an unaceptably high voltage. The small 4700uF capacitor did help to some degree and has kept the maximum voltage below the point at which it would cause damage. However the level of ripple between the low point and high points is still too great.

The inverter like most Grid Tie Inverters varies its load according to an algorithim based upon the voltage it detects.

I am happy to throw away some of the power as heat. In-fact this is actualy what I am trying too do. This is the purpose of a Dump Load or what is somtimes called a diversion load.

When you talk of matching the load you are refering to circuits with fixed voltages and fixed currents. This is not what happens with a Wind Turbine at all. Since the wind varies and the power output from the turbine increases exponentialy according to wind speed, it is entiley possible for the turbine to double or quadruple its power output within a seccond or two.

The turbine to load matching is perfomed by an algorithim within the inverter. This algorithim is based upon the maximum power output of the turbine at particulay voltages. Since power is related to the speed of the wind and increased wind speed will cause the turbine to spin faster and produce more voltage. Controlling that voltage also controls the speed and the amount of power being produced..

A voltage regulator realy is the best way to do this. Whats more, the more in-efficient it is when the incomming supply volatage is 'above' the set point the better !!. If the regulator can absorb any additional energy and dissipate it via a large heatsink then that is the best I could hope for. However at voltages 'below' the set point it should be as efficient as possible and not be producing heat.

 

[alec_t]

However it has a tendency to overshoot the mark because the digital circuit cannot operate any faster than 4Hz.

That seems an incredibly slow response for a digital controller. I could understand it if turbine mechanical inertia meant the turbine took 0.25 sec to respond a given amount to a change in wind speed. Perhaps the 4Hz is related to alternator output frequency (hence rpm)? What rpm does the turbine/alternator do at max?

 

[GoGreenElectricsLTD]

That seems an incredibly slow response for a digital controller. I could understand it if turbine mechanical inertia meant the turbine took 0.25 sec to respond a given amount to a change in wind speed. Perhaps the 4Hz is related to alternator output frequency (hence rpm)? What rpm does the turbine/alternator do at max?

Arrrghhh.... All I want is a simple TL783 Voltage regulator Circuit using a couple of 300w Mosfets !!.

But in reponse to your comment - yes the 4Hz is slow and I have taken this up with the manufacturer of that particular digital circuit. However to date I have not found a suitable alternative. Most readily available Wind Turbine Dump Load controlers apear to be made for people who are charging batteries and so the devices include a PWM charge controller too. But since I am NOT charging a battery, I am simply connecting directly to a GTI, I do not need PWM charge control.

Max safe speed on the turbine is around 1200 rpm, but if un-controlled this could exeed 2000 rpm.

 

[alec_t]

All I want is a simple TL783 Voltage regulator Circuit using a couple of 300w Mosfets !!.

I'll see what I can come up with.

 

[GoGreenElectricsLTD]

I'll see what I can come up with.

Thankyou very much

What I would like to know is "does the power rating of the mosfets need to be equal to the power being drawn from the regulator. Or does the mosfet rating need to be equal to the power being dissipated as heat ?."

IE at full power the Turbine produces 1Kw... One inverter can proces up to 600w. So that leaves 400w of power which need to be dissipated as heat.
Two Inverters can process up to 1200w. So that leaves 0w of heat to be dissipated.

 

[alec_t]

Something like this perhaps?

R7 and R8 help current-balancing between the FETs. Further FETs could be added to share the current and power dissipation. The TL783 and FETs would need big heatsinks. R7 and R8 would also be dissipating considerable power.

Suppose Vin is 100V, Vout is 85V, Iload is 10A.
Load power = 85*10 = 850W.
TL783 power ~ 6W (from simulation).
R7 (or R8) power = 10*10*0.1 = 10W
Total FET power = (100-85)*10 = 150W.
Power per FET = 150W / number of FETs = 75W in this example.

Edit: Increase R3 to 27 Ohms (minimum 10 Ohms) to reduce TL783 dissipation.

 

[ronv]

I'm not sure GoGreen really wants a regulator. I know he thinks he does.
Right now he has 30 volts of ripple that for some reason makes him loose power from the inverter. We can only guess that the ripple is from 95 to 65 volts since we don't have that info yet. So if you apply the regulator set to 90 volts you will only reduce it from 95 to 65 to 90 to 65.
It sounds like the resistors in the load are to small. That's why I ask about it's current at 95 volts.
The other downside is that 50 watts or so will be lost all the time even when the load dump is not on. Maybe that's OK.


**broken link removed**

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.

 

[Mr RB]

The problem seems to be that you have two conflicting closed loop systems.

The commercial inverter is a MPPT type that tracks maximum power, with your wind generator that will occur with maximum input voltage. So the MPPT is adjusting it's PWM until the input voltage rises to the max.

However when the voltage rises enough, your other closed loop system (the overvoltage protection circuit) clicks in and the added load rapidly drops the voltage. The causes the MPPT to re-adjust rapidly, giving your nasty 4Hz loop oscillation.

What you need is a very gentle curve for the overvoltage protection to come on, over (say) a range of 20 volts. So it draws zero current at 80v and is fully on at 100v. Then the MPPT controller in the commercial inverter will allow voltage to rise until your limiter JUST starts to waste some current, which the MPPT will detect and then keep the voltage just below this point (where it can track maximum power).

 

[GoGreenElectricsLTD]

Something like this perhaps?
View attachment 71836
R7 and R8 help current-balancing between the FETs. Further FETs could be added to share the current and power dissipation. The TL783 and FETs would need big heatsinks. R7 and R8 would also be dissipating considerable power.

Suppose Vin is 100V, Vout is 85V, Iload is 10A.
Load power = 85*10 = 850W.
TL783 power ~ 6W (from simulation).
R7 (or R8) power = 10*10*0.1 = 10W
Total FET power = (100-85)*10 = 150W.
Power per FET = 150W / number of FETs = 75W in this example.

This sounds like it could work !..... But I do want to use the 300w FET's shown at the top of this thread. Would they be a direct replacment in this circuit ?.

I dont suppose you would be able to produce a heat dissipation graph for the FET's and the TL783. I would be interested to know what the level of heat dissipation is when below the voltage regulation set point. Also I am assuming that the yellow line for load, represents the load on the regulator produced by the inverter. What I am also interested in the the load on the supply. I am sure the heat energy which is inevitably going to be disipated in the heatsink comes from somwhere. Knowing the load in amps on the supply would also be usefull.


There is one more point that is specific to Wind Turbines and not other high power circuits. The greater the load you put on the turbine wether thats from the inverter, regulator or heat dissipation or the dump load, the lower the maximum voltage will be.

The only reason the pulses can reach such high voltage right now. Is because during the 0.25 of a seccond where there the dump load is switched off, the turbine is under loaded. It behaves much like any other inductive coil. However in the future when I install a seccond inverter (cost £350.00) then this will no longer be an issue, because 2 inverters are potentialy able to put more load on the turbine than the turbine can produce. The output voltage will not be able to exeed 90Vdc, even if the regulator and dump load were dissconected. At that time the dump load circuit will only remain in situe as a precaution against malfunction of one or both inverters.

 

[GoGreenElectricsLTD]

I'm not sure GoGreen really wants a regulator. I know he thinks he does.
Right now he has 30 volts of ripple that for some reason makes him loose power from the inverter. We can only guess that the ripple is from 95 to 65 volts since we don't have that info yet. So if you apply the regulator set to 90 volts you will only reduce it from 95 to 65 to 90 to 65.
It sounds like the resistors in the load are to small. That's why I ask about it's current at 95 volts.
The other downside is that 50 watts or so will be lost all the time even when the load dump is not on. Maybe that's OK.


**broken link removed**

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.

Actualy I have around 40v of ripple right now. From 70vdc to 110vdc and if I remove the 4700uF capacitor it gets even worse !. The inverter looses power because it needs to calculate the correct load to apply dependent on the voltage. If it sees a voltage that has large ripple, it will apply a load apropriate to the lowest point in the ripple, this being 70v. As a reult of this the dump load is requierd to absorb more of the unused energy and getts hotter than it should. If the ripple could be stabilised the inverter will then take the lion share of the power and the dump load will only take a minimal amount.

 

[GoGreenElectricsLTD]

The problem seems to be that you have two conflicting closed loop systems.

The commercial inverter is a MPPT type that tracks maximum power, with your wind generator that will occur with maximum input voltage. So the MPPT is adjusting it's PWM until the input voltage rises to the max.

However when the voltage rises enough, your other closed loop system (the overvoltage protection circuit) clicks in and the added load rapidly drops the voltage. The causes the MPPT to re-adjust rapidly, giving your nasty 4Hz loop oscillation.

What you need is a very gentle curve for the overvoltage protection to come on, over (say) a range of 20 volts. So it draws zero current at 80v and is fully on at 100v. Then the MPPT controller in the commercial inverter will allow voltage to rise until your limiter JUST starts to waste some current, which the MPPT will detect and then keep the voltage just below this point (where it can track maximum power).

Yes I would agree with this..... SWEA manufacture a great PWM dump load controller and I have one of them. However it can only cope with a maximum voltage of 58vdc. That inverter uses PWM to increase the amount of dump loading the higher the voltage goes above the set point and it works realy well !!. Unfortunately I cant use it because I am operating at almost double that voltage.

Right now my Dump load is 4 Ohms and this seems to be the best. 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 !. By reducing the resistance back to 4 Ohms it does its job much better serving as a brake. Infact now that I think of it, brakes on a Car are capable of bringing a car to a halt quickly if needed. But we dont drive around with our feet continuousley on the brake, we only use it when necesary. The level of heat dissipated in a car braking system is only appropriate to the level of energy lost. If you removed all your modern disk brakes of your Volvo V70 and replaced them with drum brakes from a 1970's mini. You would still be able to slow the car down and stop it. But it would take a very long time and cause significant overheating, possibly even catching fire. In the end the brakes would use more energy doing it. Well I dont want drum brakes on my Wind Turbine, I want the equivalent of disk brakes !!.