external power supply, rectifier in or out?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,
with multiple taps and such a big transformer. I would go for something a lot more flexible.
Two AC three wire outputs, one fixed at 50+50Vac and the other switchable.

Two DC three wire outputs, again one fixed from the 50Vac winding and the other one switchable.

For the DC 3wire supply put the transformer+rectifier+smoothing all in a very short small area loop.
The 100Vac to DCsupply will need 200Vdc smoothing caps !
 
This thread is almost a year old but in the interests of limiting threads I thought that maybe it would be better to add to this rather than start a new one as the subject is touched on briefly.
As I'm in the (long) process of designing a 'universal' external power supply that could be used with a number of power amp modules (hence saving cost) for comparative listening tests I finally got to the power output connections.
As it seems that the 'audiophile standard' is to use copper busbar to interconnect the rectifiers/caps et al, how does one thus connect the PSU with the amps without similarly using busbar interconnect 'cables'? Using the thickest wire available will still be 'insufficient' in comparison - and then there's flexibility to consider ...
Allied to this is finding high current output sockets that are reasonably priced if one does go the 'thin wire' route. Cheapest alternative would be to use 3 pole mains outlets - but somebody, somewhere, sometime would probaly connect mains voltage into the outlet sockets - by hook or by crook - in my absence (temporary or permanent).
So exactly how does one get past this conundrum?
Right now I'm thinking 'stuff it' (the cost savings of one PSU and multiple smaller amp chasis) and putting the PS into a box with the power amplifiers rather than external.
It was originally intended to put caps and voltage control in each power amp box but the 'wimpy interconnect wire' with the PSU problem still remains a 'gotcha'.
What at first sight seemed the 'easy part' of the projects ...

Any suggestions/experiences welcomed!
 
Pol, you have the theory all mixed up.

When using a remote PSU, it supplies DC current to the active equipment.
It's job is to turn AC mains voltage into safe (isolated from mains) DC voltage/current.

Don't get all worked up about level of DC current sent to meet the DC current demand of the active equipment.
Just design the Mains to DC PSU to do it's own job properly.
 
Thanks for the quick reply Andrew.
Actually I'm housing the main Caps (2 by 4,700 uF per supply rail) in the PSU box too.
Further line conditioning would be done in the amps boxes.
Unless I'm completely mistaken by putting the 'primary' caps in the PSU supply in the first place?
 
Last edited:
that's where you have it mixed up.
The transient peak demand must be met by local capacitance or local low impedance regulated supply.
The decoupling must all be right next to the amplifiers highest current demand devices. Particularly those that can suddenly turn on demand or turn off demand from a quiescent state.

The remote PSU tops up the decoupling with slowly changing DC current. These recharging currents alter slowly and never become large.

Locate the transformer and rectifier and smoothing caps close together.
Locate the amplifier with it's decoupling.

If amp and PSU are separated then add some Low Frequency decoupling next to the amp.
 
Thanks yet again Andrew,

Actually been struggling (contemplating) whether or not to add the main caps to the amp boxes or the PSU. Appreciate that the current variations are supplied by the caps, but have lots of space in the PSU box. Additionally trying to limit variations in (pulses including hum et al) of current/voltage from the PSU through to the (relatively) clean amp box and any problems that might occur from this. more or less a 'battery type of supply' for want of a better description. Maybe this is not audible anyway?
Ultimately I was trying to create a 'perfect power supply' (yes I know) for comparing different amplifier topologies to remove the 'power supply variances' from the equation.
Thanks for your help. Guess I'll use the simpler torroid/rectifier PSU and put the power where it's needed - 'on board'.
BTW, have ordered a pair of the JLH 10W amps (MJ15003 version) from SiliconRay as the first amplifier comparison point.
 
Last edited:
Administrator
Joined 2004
Paid Member
Place the main smoothing caps with the rectifiers and transformers as Andrew has suggested, you do not want to be sending large charging currents across umbilical wiring - substantially worsens the regulation of the supply due to IR drops in the wiring (and the large ripple current) and also radiates lots of EMI into surrounding space.

Additional capacitance in the range of 10 - 20% (or more if you'd like) of the smoothing capacitance should be placed within the amplifier enclosure to supply local variations of demand. (local decoupling)

Make sure to place bleed resistors across both the smoothing caps and the local decoupling caps so that they have independent discharge paths if they are disconnected. (And don't reconnect until you are sure they are fully discharged as the arcing may ruin your connectors.)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.