HELP needed with TL783 Regulator Circuit

[GoGreenElectricsLTD]

@ronv
? we must be using different datasheets
View attachment 71936

@GoGreen
Can you define 'negligible'? The IC can put out only ~3mA at close to 5V, so a buffer stage might be needed? Or were you thinking to use the FET to drive the SSR?

OK, well I just measured the current flowing in the existing circuit controlling the SSR and I got 26mA, but thats is at 24v !. Although it doesnt need it I put the 150 Ohm resistor in there just as a saftey precaution. The SSR will switch on with anything over 5v.

I tried using the FET to drive the SSR once before and it didnt seem to work. I think it needs a load accross it to operate correctly ?.

Incidentaly its a Mager SSR, I am sure there is a data sheet somwhere ?.

 

[alec_t]

I tried using the FET to drive the SSR once before and it didnt seem to work. I think it needs a load accross it to operate correctly ?.

I'm sure we can find a way.
In the meantime, if my calculations are correct, these are suggested changes/checks to uprate the controller and allow 80-95V adjustment:

D2 - ensure suitable voltage rating (e.g. use 1N4003 or better),
R3 - 7k5, >= 1.5W,
R4 - 560k,
R6 - 4k5 ( or use 8k2 in parallel with present 10k),
R7 - 60k ( or use 120k inparallel with present 120k.

 

[GoGreenElectricsLTD]

I'm sure we can find a way.
In the meantime, if my calculations are correct, these are suggested changes/checks to uprate the controller and allow 80-95V adjustment:

D2 - ensure suitable voltage rating (e.g. use 1N4003 or better),
R3 - 7k5, >= 1.5W,
R4 - 560k,
R6 - 4k5 ( or use 8k2 in parallel with present 10k),
R7 - 60k ( or use 120k inparallel with present 120k.

Do these changes allow for the correct voltage feed to the IC and to the LED's ?.

I found the specifications for the SSR I am using:



Load Current: 80A

Load Voltage: 5-220V DC Actual (24-220V DC Marked)

Control Voltage: 3-32V DC
Control Current: 5-25mA DC

On Voltage: ≤2.3VDC

Off Leakage Current: ≤2mA

On-off Time: ≤10mS

Dielectric Strength: 2000VAC

Insulation Resistance : 500M ohm/500VDC

Ambient Temperature.: -30℃～75℃

Working indicators: LED
Certification: CE

 

[alec_t]

Do these changes allow for the correct voltage feed to the IC and to the LED's ?.

Yes. R3 determines current and zener D6 (looks like a 5V1 in the photo, not the excessive 6V2 as shown on the schematic) determines voltage.

On-off Time: ≤10mS

That could be a big problem if the controller is doing PWM at anything over ~ 20Hz.

 

[GoGreenElectricsLTD]

Yes. R3 determines current and zener D6 (looks like a 5V1 in the photo, not the excessive 6V2 as shown on the schematic) determines voltage.
That could be a big problem if the controller is doing PWM at anything over ~ 20Hz.

Well, we can try it with the existing FET first, I can check the frequency when its operating. Also I was a bit concerned that the zenner may not be able to handle the higher voltage level. Since it has no markings on it I have no way to check the specifications.

If you think those component values should work. What I will do next is send a copy of that list to Mari at SWEA and see if he can confirm if they will be OK. Then I will order the components and make the mod's.

Before all that though I will see what difference this new enormous cap makes to my present system. Unfortunately I have to wait for the next round of stormy weather to see if anything has improved.

 

[alec_t]

Also I was a bit concerned that the zenner may not be able to handle the higher voltage level. Since it has no markings on it

The zener sets its own voltage. The present current through it is ~ 12mA if R3 is 3k9 and the turbine voltage is 58V. It would still be ~12mA if R3 becomes 7k5 and the turbine voltage is 95V. I'm sure I can make out '5V1' marked on the zener in your second photo? That is wholly consistent with the IC supply max of 5V5.
95V is uncomfortably close to the FET 100V rating. Personally I'd look for a higher rated logic-level FET.

 

[GoGreenElectricsLTD]

The zener sets its own voltage. The present current through it is ~ 12mA if R3 is 3k9 and the turbine voltage is 58V. It would still be ~12mA if R3 becomes 7k5 and the turbine voltage is 95V. I'm sure I can make out '5V1' marked on the zener in your second photo? That is wholly consistent with the IC supply max of 5V5.
95V is uncomfortably close to the FET 100V rating. Personally I'd look for a higher rated logic-level FET.

Im sure the FET will be fine. All these devices have some tollerance on voltage, plus when the circuit is operating corrently it will never get past the 100v mark anyway. The alternative is that I use the SSR which I am already using.

 

[alec_t]

If you wanted to try using the FET to drive the SSR this might work, using just two additional components to limit the SSR input current to ~ 8-11mA @ <= 8.2V. The R1 here would be dissipating ~1W.


Edit: However, I suspect the SSR turn on/off is too slow if the micro output is PWM. That could account for why your previous attempt to drive the SSR was unsuccessful.

 

[GoGreenElectricsLTD]

If you wanted to try using the FET to drive the SSR this might work, using just two additional components to limit the SSR input current to ~ 8-11mA @ <= 8.2V. The R1 here would be dissipating ~1W.
View attachment 71973

Edit: However, I suspect the SSR turn on/off is too slow if the micro output is PWM. That could account for why your previous attempt to drive the SSR was unsuccessful.

I think it would probably be easier to put somthing like two 10w resistors acting as a potential divider accross the output of the FET. Letts say a 750 Ohm resistor and a 250 Ohm resistor. The voltage accross the 25 Ohm resistor then would never exeed 25v and this could be used to drive the SSR.

 

[alec_t]

Letts say a 750 Ohm resistor and a 250 Ohm resistor.

At 95V, current > 95mA. Power > 6.8W in the 750R, 2.3W in the 250R. That would mean bulky resistors and a lot of heat to get rid of .

 

[GoGreenElectricsLTD]

At 95V, current > 95mA. Power > 6.8W in the 750R, 2.3W in the 250R. That would mean bulky resistors and a lot of heat to get rid of .

And the problem is ?????

 

[alec_t]

The problem would arise only if you wanted to pack the controller into a small case, or if you were building this down to a price (two x 10W resistors cost more than 1 x 2W resistor + 1 small zener)

 

[GoGreenElectricsLTD]

The problem would arise only if you wanted to pack the controller into a small case, or if you were building this down to a price (two x 10W resistors cost more than 1 x 2W resistor + 1 small zener)

True..... However for the sake of reliability and a certain degree of being bullit proof. Surley the pair of resistors are the better bet ?. If the zenner were to pack up and the SSR didnt fire, the turbine would over speed and destroy its bearings within a few hours.

Ive not had much experience putting zenners into a circuit, but I have had to replace a lot of damaged ones !... So im not realy a big fan of them. Do you realy think such a circuit could work, and would the zenner survive these kind of high voltages ?. I have looked over the existing circuit board and I think I could fit the two components in that you have suggested. All I would need to do is scratch out one part of the existing tracking on the board.

Would a BZX85C24 be any use as I have some of these in my drawer ?

 

[alec_t]

However for the sake of reliability and a certain degree of being bullit proof. Surley the pair of resistors are the better bet ?.

Not sure where I'd place my money if I did bet on it! On the one hand a beefy resistor may be more reliable in itself than a zener (I have no evidence one way or the other). On the other hand the heat generated could make solder joints in the vicinity more likely to fail due to expansion/contraction of board and components.

Would a BZX85C24 be any use

Probably. If the 8k2 were retained that would limit the SSR current to ~ 9mA. Would that be enough? Don't know what voltage the 9mA would actually drop across the SSR, but it would be 24 max.

Edit: Can you test the SSR with 5V input and measure the current it draws? At the moment we're just guessing.

 

[GoGreenElectricsLTD]

Not sure where I'd place my money if I did bet on it! On the one hand a beefy resistor may be more reliable in itself than a zener (I have no evidence one way or the other). On the other hand the heat generated could make solder joints in the vicinity more likely to fail due to expansion/contraction of board and components.
Probably. If the 8k2 were retained that would limit the SSR current to ~ 9mA. Would that be enough? Don't know what voltage the 9mA would actually drop across the SSR, but it would be 24 max.

Edit: Can you test the SSR with 5V input and measure the current it draws? At the moment we're just guessing.

I did test it once with a 9v battery and that worked ?...

I do think 9mA is a little on the low side even though its within spec for the SSR. I would realy like to see some more current than that. Perhaps somthing like a 4.7K Ohm 2w resistor would work, that would make the current 16mA and would keep the wattage within the tollerance of the 1.3w zenner.

My other concern at this point is wether the 16ma current flow is sufficient for the FET on the control board to operate correctly ?. Its the main reason I suggested using a couple of 10w resistors as a PD in the first place.

I could easily fit the zenner and the 2w resistor to the existing board without too much difficulty. But what I would like to do is try it out with the existing voltage settings, if it fires the SSR at 58v, then I know its deffinately OK at the higher voltages. What im concerned about is making sure the SSR is fully on. As a low input current to the SSR will cause it to only be part way on and not work properly. Luckily it has a little LED on the front and when its porperly on it lights up brightly.

You mentioned before that the 10ms delay could be problem if operating over 20Hz. I worked out that 20ms equals 50Hz ???....... However what I wanted to know is letts say the controll circuit did go above that frequency and was operating at 500Hz for example. Surely that would just make the SSR stay contnuousley on ?, which is OK ?.

 

[alec_t]

Perhaps somthing like a 4.7K Ohm 2w resistor would work, that would make the current 16mA and would keep the wattage within the tollerance of the 1.3w zenner.

Sounds fine. That = 1.2W heat in the 4k7 and < 0.4W in the zener.

My other concern at this point is wether the 16ma current flow is sufficient for the FET on the control board to operate correctly ?

I'm not aware that MOSFETs have a minimum current requirement. The 16mA is much greater than the drain-source leakage current (max 250 microamps as per spec) and the SSR certainly wouldn't be triggered by the leakage current.

As a low input current to the SSR will cause it to only be part way on and not work properly.

I've never used an SSR but I'd like to think they were either on or off, not half way. Other forum members could give you a more informed view.

You mentioned before that the 10ms delay could be problem

The 20Hz was surmise (2 x 10mS would be a very significant 40% of a 20Hz PWM signal). If you could monitor the gate voltage of the FET with a scope when the dump load is active we would have some useful info. Or perhaps your pal at SWEA could tell you what sort of gate drive waveform the PIC outputs (unless that's a trade secret)? Let's suppose the PIC is programmed to pulse the gate high for only 1mS then low again for 1mS (i.e. at 500Hz). Far from being on continuously the SSR would not then activate at all as far as I can tell from its spec.

 

[GoGreenElectricsLTD]

Sounds fine. That = 1.2W heat in the 4k7 and < 0.4W in the zener.
I'm not aware that MOSFETs have a minimum current requirement. The 16mA is much greater than the drain-source leakage current (max 250 microamps as per spec) and the SSR certainly wouldn't be triggered by the leakage current.
I've never used an SSR but I'd like to think they were either on or off, not half way. Other forum members could give you a more informed view.
The 20Hz was surmise (2 x 10mS would be a very significant 40% of a 20Hz PWM signal). If you could monitor the gate voltage of the FET with a scope when the dump load is active we would have some useful info. Or perhaps your pal at SWEA could tell you what sort of gate drive waveform the PIC outputs (unless that's a trade secret)? Let's suppose the PIC is programmed to pulse the gate high for only 1mS then low again for 1mS (i.e. at 500Hz). Far from being on continuously the SSR would not then activate at all as far as I can tell from its spec.

The SSR does have the ability to be part way on as ive done it myself when messing around with different size resistors on the input.

I have sent a question the the supplier of the SSR about what would happen when fed with a 500Hz pulse.

 

[alec_t]

The SSR does have the ability to be part way on

That's something new to me. Thanks.

I have sent a question the the supplier of the SSR about what would happen when fed with a 500Hz pulse.

I'd like to know their answer.
In the meantime, can you check the SSR input current when fed with 9V?

 

[GoGreenElectricsLTD]

That's something new to me. Thanks.
I'd like to know their answer.
In the meantime, can you check the SSR input current when fed with 9V?

I cant on the one I am using as its in use today, with it being windy and all. But I do have another of similar spec. So I will test that one !.

 

[GoGreenElectricsLTD]

That's something new to me. Thanks.
I'd like to know their answer.
In the meantime, can you check the SSR input current when fed with 9V?

OK, well I tested it with a 9v battery which was probably a little on the low side at 8.18v no load, although it did blow a 500ma fuse when I put the meter directly accross it. Anyways using that battery I got a reading of 13.44 mA