It's because I could not find the pinout of both of them in the library of components so I improvised a littleThen what led you to number their pins identically to each other?
I would carefully check the pin outs. ;-)
Another point, mentioned by Douglas Self: if one of the regs starts up before the other from e.g. An unbalanced load, it will pull the output of the slower start up reg to te wrong polarity and this will lead to self destruction.
The way to cure this is to put some reverse clamp diodes between the output and ground of each regulator.
Are you plugging your load into the reg board? Might also be an opportunity for this to cause the situation I described above.
BTW, if you accidentally blew your zeners, then you would not get the reverse polarity protection I described above. I would remove these and replace with 1N4007.
Another point, mentioned by Douglas Self: if one of the regs starts up before the other from e.g. An unbalanced load, it will pull the output of the slower start up reg to te wrong polarity and this will lead to self destruction.
The way to cure this is to put some reverse clamp diodes between the output and ground of each regulator.
Are you plugging your load into the reg board? Might also be an opportunity for this to cause the situation I described above.
BTW, if you accidentally blew your zeners, then you would not get the reverse polarity protection I described above. I would remove these and replace with 1N4007.
Initially I placed the zeners to protect against excess voltage on the opamps. I did not think of reverse voltage, but I think they're OK since I checked them (while in the board) and they seem to behave like a diode.
As for unbalanced voltage... I am using a center tapped secondary transformer with a single bridge rectifier with ground at the center. I think this is an unbalanced setup right? So maybe I need to use two bridge rectifiers to get balanced input voltage? I never did this before. Does it make sense?
Good points Bonsai. With +/-38V going in there's plenty of opportunity to exceed the maximum in-out voltage spec on power-up. A sure-fire way to get regs to either explode or let out the magic smoke... If you have decoupling capacitors on your opamp board between the two power rails (as opposed to to ground from each) then this problem is much more likely to occur.
I have to agree here.
If the slower VR gets just a few Volts in opposite polarity on its output, even if it's only a short while, the Vin - Vout = 40 V max. will be exceeded.
The regulators may take this abuse a few times, but will eventually fail.
Make sure all decoupling is between power rail and ground.
Perhaps the low current draw of a few opamps may also cause a less than ideal startup of the regulators. One of our customers has a near identical design based on LM3x7 for a +/- 18 V PSU, the only real difference with your circuit is two 1k5 1W resistors between each output and ground to ensure sufficient loading of the regulators at all times.
Last edited:
Minimum load current is 10 mA (worst case) according to the datasheets. Typical values are lower, but it's safer to assume you need the 10 mA.
With R2 = 150R and R3 = set to 1k77 for Vout = 16 V, load current is just over 8 mA. Even with just one opamp connected would you already exceed worst case min. load requirements, so this rules out problems with the load current.
Why our customer chooses to load the output of the PSU with an additional 12 mA, I'm not really sure about. Perhaps the additional load is just to make sure the regulation will absolutely always be stable.
My advice to elic is to install clamping diodes or, if that doesn't work, a pre regulator to drop the +/-38 V to a lower value before feeding it to the LM3x7s. Good luck!
With R2 = 150R and R3 = set to 1k77 for Vout = 16 V, load current is just over 8 mA. Even with just one opamp connected would you already exceed worst case min. load requirements, so this rules out problems with the load current.
Why our customer chooses to load the output of the PSU with an additional 12 mA, I'm not really sure about. Perhaps the additional load is just to make sure the regulation will absolutely always be stable.
My advice to elic is to install clamping diodes or, if that doesn't work, a pre regulator to drop the +/-38 V to a lower value before feeding it to the LM3x7s. Good luck!
Last edited:
I'm putting my money on the zeners. The datasheet says these regs can even be used at 3-digit regulated voltages, with no dire start-up warnings. Doing the formula correctly(!) using 150 and 2k resistor gives a higher voltage than the 16.4V of the 2 zeners. And the zeners are "odd man out" in an otherwise relatively easy circuit. What kind of regulation battle occurs when the LM3x7s take on the zeners?
Since elic drew R3 and R6 as potentiometers in his schematic, and he wrote that his PSU was set to +/- 16 V, I'm assuming the zeners are doing no battle with the LM3x7s.
Edit: come to think of it, since zeners act more or less as normal diodes in forward direction, shouldn't they protect the outputs from reverse voltage like clamping diodes?
Edit: come to think of it, since zeners act more or less as normal diodes in forward direction, shouldn't they protect the outputs from reverse voltage like clamping diodes?
Last edited:
Correct, the potentiometers are set to provide +/- 16V. I just got home from work. I am going to do some testing and will update you guys later.
If they are rail to rail then the cap in question can pull the output of one rail toward the other at switch on and switch off and cause undesired effects. It could pull the positive rail negative or vice versa and momentarily exceed some parameter of the IC.
I don't think this is possible. One rail can't be favoured over the other because of symmetry.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.