JSR04, +12V, +5V (from the negative side)

[builder brad]

Hi P-A,

I hope you had a relaxed and enjoyable break!

any ideas regarding this PS project?

Brad

[peranders]

Hi P-A

Finally got to around to having a closer look at this project again......

The faulty component was the T4 transistor and I now have 4.13 volts regulated output which drops to 2.7 volts when I connect a 10 watt 47 ohm load resistor. The H2 LED is not illuminated and only has .3 volts across its terminals. R32 is supplying 2.61 volts to the LM431 regulator.

Can you measure the voltages around the AD825? The LM431 should have 2.5 volts and 2.61 is a tiny bit too much. Have you paid attention to the polarity of the LM431? Anode and cathode must be switched and ref must be tied to cathode.

Which transistor types do you have right now?

Could you possibly take a couple of closeup photos? (not more than 2 MB please)

[builder brad]

Hi P-A

Thank you!

when I chacked the Op amp voltages I immediately noticed that it was not getting enough voltage, then I remembered that I had cut its supply during some of the previous tests and had re-routed a supply from the +12v op amp side.

As soon as I corrected for this i got exactly 5.00 volts!!!!!!

quick test show a drop in output to 3.66 volts when a 12 ohm resistor is connected across the output, is this ok?

Brad

[peranders]

Good that you finally manage to get 5 volts out of the regulator :D

The voltage should be constant within a couple of mV's when it regulates properly.

When you increase the load and get a voltage drop you can have this:

Not enough input voltage and this means including the voltage ripple.

Not enough headroom for the opamp output due to the DC shifter (the LED or zener). I'll guess you feed the opamp with more than 5 volts so this should be a problem.

Not enough base current for the output transistor

[builder brad]

Thank you,

The voltage drop is still a problem and when the Power supply is connected to the squeezebox the voltage also fluctuates.

The AC supply is 10v which produces 9v unregulated. I am still concerned about the voltage drop across R12 which is 1.48v and the amount of heat that is generated by R12 - R11 on the 12v+ side only drops .3v and is cool to touch.

The Opamp used is an NE5534 which is powered by 5.7 and this drops to 4.8v when the load is connected.

The LED is still not illuminated - I am getting .32v across the LED terminals.

I know that I am almost there now, just need to sort this final thing

[peranders]

5,7 volts for NE5534 is a bit low accordning to datasheet which says 6 V but ought to work. Remember though that a NE5534 isn't verified by me. AD825 works.

You must also unsure that the incoming voltage is 8 V (can be measured by an oscillscope) at least as a very minimum and if you don't use the preregulator if if you do it's 11 V.

[builder brad]

Hi P-A,

I am using a 9-0-9 transformer with the PS, configured to provide a nominal +9 and +18v AC suppply for the +5 and +12v regulators. Is it possible to power the 5v PS from the +18 turn of the transformer, this would certainly deal with the low voltage preoblems!

[peranders]

9 VAC should give you 11-13 V I should guess. One way to find out if 9 VAC is to little for you is to test the 5 V regulator with 18 VAC in.

The value of R10, R12 is dependent of how much current your load take. If you can't "afford" much there, just short them.

[builder brad]

I tested with the 18v AC supply and saw quite a large improvement. The 5v output only dropped by about .2v with the 12ohm load, so I have now removed/shorted the R12 resistor and now see a similar voltage drop with that 12 ohm load.

Just trying the 12 ohm load on the 12v side also causes a large voltage drop down to 8v. i have shorted the 12v side equivalent resistors, but do not see an improvement.

I think I may have the answer, but just want to check if I am on the right track.

On the original schematics you indicated the required parts for this project and included one load resistor and diode for each half of the power supply ie. R8 and D8, which only utilises one half of the AC cycle. Surely I need a full bridge rectifier here for each half of the PS?

I could easily modify the board to provide full wave rectification - would this help with the voltage drop problem?

[peranders]

You can use halfwave rectification. The only thing to think of is maybe more smoothing capacitance since you have only one voltage pulse during a mains cycle.

Your AC source, is it weak? How many VA is the transformer rated?

[builder brad]

The transformer is rated at 50VA/2.78 amps. our mains supply is nornally pretty good here in terms of voltage, always around the 230v mark. I do not have access to a scope  so there ius no way of really knowing if there are any other problems with the mains.

I do want to get this perfect as I am planning on powering the JSR06 with balanced AC power at a slightly lower mains voltage - around 225volts

[peranders]

The disadvantage of halfwave rectification is:


More (double) smoothing capacitance

More harmonics going out to the mains

More losses in the transformer ... but this is more present if the load is heavy in relation to the size of the transformer.

In your case you have 50 VA but how heavy is your load really?

[builder brad]

Well......it would be very easy to add an extra 2 diodes and resistors for full wave rectification, so I guess i will do that anyway, your PCB is v.flexible in that area. I will just change the pcb tracks to suit, is it absolutely neccesary to include the capacitor across the resistor and diode in the rectifier 1/2 that I am going to add?

I originally wanted to build a PS that could supply battery-like currents, if demanded, to cope with transients ect. I guess you are right about the "actual" power requirements, although the original voltage drop I was seeing would not provide enough power in any situation.

[peranders]

If the transformer has two separate secondary windings it's possible to fullwave rectify, otherwise not. Winding with center tap will not do.

[builder brad]

The Transformer used has 2 x 9v windings that are separate, although thay are joined at the centre tap for the 9v supply, with the second winding connected in series to give 18v. The start of the first winding is 0v and it would be this point, or somewhere on the ground plane that I would  connect to a 4watt 1 ohm resistor and diode for the 2 x additional sides of the bridge rectifier.