15v dual power supply problem. need help

[Scott594]

Originally Posted by poldus

:snip:
The transformer output is 14.6v while connected to the board. 14.75 on the positive and -14,90 on the negative rail (relative to ground). :snip:

You need more than 14.75 out of the transformer (I think you mean rectifer output to ground though) for the regulator to function at all - they need at least 17vdc in to produce 15 out reliably. The regulators cannot increase the voltage. As for the rectifer configuration, the half-wave gives AC X 1.414=Vout. Thus a 12.6V transformer will give about 17.6Vdc best case. Subtract from that the .6 - .7v foward drop of the rectifer and you get 16.9-17Vdc which is right on the regulator's dropout spec. If you could check the transformer secondary voltage (leg to leg) it may shed some light on the problem. With this minimum of headroom, balance between legs is critical to prevent one side from dropping out. In this design ground floats between the 2 DC rails. Simple component tolerances would be enough to kick it one way or the other. This document from Hammond may help.
http://www.hammondmfg.com/pdf/5c007.pdf

[poldus]

Sorry for not being too clear on the numbers.

14,6v AC at the transformer secondary ( as scott594 says, seems like I need a higher voltage transformer then. What rating would you recommend?)

+14,75 and -14,90 relative to ground after the rectifiers, when the unit is working, but:
-0,72 (this on the positive rail!!!) and -14,92 on the negative relative to ground when in faulty condition, with 2,200uf capacitors added.

18,06 at positive regulator input and -17,86 at negative regulator input.

[AndrewT]

14.6Vac from the transformer makes it seem like this is a 12Vac transformer.
Normally a transformer voltage for a regulated supply is chosen to equal the DC output voltage.
eg.
for a 15Vdc 200mA supply choose a 15Vac 800mA transformer. This 12VA transformer may have a regulation of about 20%. Measuring across the secondary when the primary is fed with an accurate mains voltage supply will be ~18Vac.

If this were a 230:15Vac transformer fed with UK mains then expect ~16.9Vac when mains is at 216Vac and expect ~19.9Vac when mains is at 254Vac.

Your regulator MUST be designed to work reliably within that range of supply voltage and over the range of output currents between 0mA and 200mAdc.

[Mooly]

Originally Posted by poldus

14,75 and -14,90 relative to ground after the rectifiers, when the unit is working, but:
-0,72 (this on the positive rail!!!) and -14,92 on the negative relative to ground when in faulty condition, with 2,200uf capacitors added.

18,06 at positive regulator input and -17,86 at negative regulator input.

After the rectifiers is the input to the regs
Your readings contradict each other...

[poldus]

The transformer is 15v secondary.I checked before doing the connections. The measured AC is with it connected to the power supply, which seems to reduce the voltage.
I should have said regulator instead of rectifier. However I was referring to the regulators´input or output as per the circuit diagram on the ESP site which marks them as in, out and ground legs.

[AndrewT]

a measured 14.6Vac from the loaded secondary does not indicate you are using a 15Vac transformer.

From that example I quoted, the worst case low voltage will be ~20.5Vdc and worst case high voltage will be ~27.4Vdc at the input to the regulators.

This is not what you are measuring.
Your voltages are very much lower, indicating you have selected the wrong transformer.

[Mooly]

You have something odd happening to cause the start up issue you have... and I'm sure Andrew will agree that whether the tranny is right or wrong... there is something else going on. If nothing gets hot when it's faulty, such as the regs, diodes, caps etc then it seems more like an intermitent wiring error.

Could you post a picture of it all

[poldus]

I have no camera so can´t post a photo. However, I´ll get another transformer tomorrow and see if it fixes the problem.
The fact is I don´t know my transformer´s amperage.
AndrewT mentioned 15v 800mA. Should I try an 18v or so for a safer bet? Let´s see what´s available at my electronic store.
BTW. I´ve noticed no high temperature at the regulators or anywhere else for that matter.

[Mooly]

It won't be the transformer... but try it to "prove" that to yourself.

Do this and copy and paste this with your results.

Connect your meter negative to the left hand ground in Rods circuit (the transformer) and set the meter on DC volts. You don't move that negative lead now, leave it where it is
Now measure these these voltages in this order and write them down and copy and paste the results here.

1. Cathode (stripe end) of D1 ...... Volts
2. Anode (non stripe end) of D2 ...... Volts
3. Pin 1 of the 7815 ...... Volts
4. Pin 2 of the 7815 ...... Volts
5. Pin 3 of the 7815 ...... Volts
6. Pin 1 of the 7915 ...... Volts
7. Pin 2 of the 7915 ...... Volts
8. Pin 3 of the 7915 ...... Volts
9. Positive end C2 ...... Volts
6. Negative end C3 ...... Volts

Now switch the meter to AC volts and measure on the anode/cathode junction of D1 and D2 ...... Volts AC

[poldus]

Mooly, the units powers on correctly now, so it´s a finicky fault whatever it is.
The thing is both boards show the same condition always. That pretty much rules out a bad solder joint.
Thanks for the voltage check-list.As soon as the fault reappears I will measure the values on your table and post.

I won´t change the transformer for now.