cascading power supplies

[SheldonD]

Hello:

I will have 3 different voltages from Xfrmer 2x 22V AC and

LT1083 regulated to say +/- 24V DC.

The following supplies will use LM317 TO-220 in a standard layout
( I have a few of LM317)

This in turn will feed +/_ 15V DC this will have only a few mA current

and +/-9V DC. this will need considerable current say 0.5A

The first stage will be well Regulated and filtered.

Question: does 3 stages of regulation matter?

One of the considerations is that the 15V will either have to feed the 9V or enough bleeder current will be needed to keep the LM317 for 15V operating properly.

If it does not matter then the 9V will be enough drain to keep the 15V LM317 happy.

Thanks
I would like some advice and reason why.

 

[AndrewT]

Hi,
the low current +-15Vdc regs can be fed from the regulated +-24V supplies. The lm317/337 will not require heatsinks for 9Vdrop and a few tens of mA.

The hi current +-9Vdc supplies should not come from the 15V.
They can come from the +-24Vdc or direct from the unregulated smoothing caps. Heat dissipation will be a problem from either voltage source. A floating pre-reg could be used to bring the voltage drop down to manageable levels, otherwise the sinks are going to become very big.

Alternatively, one could add a pair of dedicated 10Vac windings (about 11Vac when measured open circuit) to the transformer using 0.6mm or 0.7mm diameter enamelled wire. Each winding needs to be rated at 10VA to give a half amp output.

Are you sure 22Vac when on maximum draw and on minimum mains input voltage can achieve sufficient voltage to give an adequate drop across lt1083 and still get 24V from it (remember to subtract ripple from the unregulated voltage when calculating/measuring)?
The last thing you need feeding your circuit is a smooth DC supply with notches due to ripple causing drop out.

 

[SheldonD]

CASCADING POWER SUPPLIES

Andrew: Thank You for reply.

Andrew wrote:

"Are you sure 22Vac when on maximum draw and on minimum mains input voltage can achieve sufficient voltage to give an adequate drop across lt1083 and still get 24V from it (remember to subtract ripple from the unregulated voltage when calculating/measuring)?
The last thing you need feeding your circuit is a smooth DC supply with notches due to ripple causing drop out."

I believe it is enough, I am presently getting 30 to 31 volts under load playing real music loudly. The Lt1083 are low dropout, (down to one volt). Tthis gives me about 6 to 7 volts to play with. I could drop another 2 volts if required.

The 3 stage idea was to :

1. drop the voltage gradually, minimising heating of 9 Volt
supply i.e. smaller heat sinks required
2. preclude the use of a bleeder on the 15 Volt stage. this will definitely be needed, as LM317 may need load of max 10 mA (avg 3.5 mA). So I will design for 10mA to ensure proper operation.

Use of added secondary windings is out of th e question.

I realize small or no additional heat sink for 15V if not supplying
the 9Volt regulator.

I still do not know why the 3 stage idea is bad, I recall reading not to do it but no reasons mgiven. To my mind the attenuation of ripple and noise should be accumulative.

Sheldon