Why are servo power supplies often unregulated?

[philpoole]

Hi all,

I've noticed on a few CDP schematics, that some circuits, (I think) mainly servo circuits, are powered with unregulated supplies.

Why is this? Is it because motors are being driven, and the current demands require that the supply is unregulated.

Surely, if the supply is regulated, the supply lines will be steadier, and this would offer more precision to the servo driver, not to mention improvements in PSRR and noise reduction in general?

If bigger current demands were required, couldn't a regulated supply with large local decoupling capacitance suffice?

Or, as another alternative, have I just completely missed the point?

I appreciate that some things really don't work well with regulated supplies (i.e. many power amps) but this is using much less power.

Cheers,
Phil

[martin clark]

There's simply no benefit. The positioning is sensed separately, so regulating the supply lines adds nothing in the way of needed accuracy, so manufacturers leave it out.

In a CD player the current demands are nothing plain old 78xx/79xx regs couldn't handle.

Now - in terms of stopping motor noise getting back into a weedy supply, yes regulation or other passive measures (e.g. small value inductors in the raw servo supply lines could make a lot of sense - and there's nothing to stop you if you like DIY.

[philpoole]

Hi Martin,

Thanks for the reply.
That confirms my suspicion (I did see one schematic were the IV conversion was also unregulated! So it does scream cost cutting).

One of my favourite CDP mods is to use a daughterboard with seperate regualtion for different parts of the circuit. Now I know I can extend it to the servo circuitry.

Thanks again!

Phil

[martin clark]

did see one schematic were the IV conversion was also unregulated!

Let us know how it turns out if you try this. You can always lookup the datasheet for the servo drivers and find out the max. current that will supply. I bet it will be under 0.5A - servos are usually only weedy little 3v motors anyway (Mabuchi or similar).

[philpoole]

Hi Martin,

I think as a start I will rearrange the regulation at the power supply.
On this CD player (Philips CD940), all power rails (ignoring the VFD display), regulated or not, are derived from +/- 10V after rectification.
I am, at some point, going to try feeding a +/- 10v regulated supply to the PSU, just after the rectification (and oviously decommision the original relevant secondaries on the transformer). That way, everything should be regulated at least once.
The only regulators on board are +/-5V, so 10V should have ample headroom for them to work properly.

I did replace the 5V supply with seperate regulation to several chips on my, alas now defunct, Marantz CD63, and the difference was actually very surprising.

I'm hoping to not mod this player so much. I just want to use it as a transport with I2S output to feed a DAC. All those regulators made it a real pain to perform maintenance.


Cheers,
Phil

[martin clark]

I am, at some point, going to try feeding a +/- 10v regulated supply to the PSU, just after the rectification (and oviously decommision the original relevant secondaries on the transformer). That way, everything should be regulated at least once.

I think I'd cut tracks as required then rewire so that the original regulated PSU remains to power the servo and control ciruitry, and the new PSU supplies the clock, decoder and SPDIF output devies - really go for the separation between servo and audio you intend. Of course if you only want to use it asa fransport, you can unhook power to the aduio output stage(s) you don't need also...

I did replace the 5V supply with seperate regulation to several chips on my, alas now defunct, Marantz CD63, and the difference was actually very surprising.

Oh I know...but the 5v rail in a standard cd63 is a total mess

[philpoole]

Hi Martin,

I think you're absolutely right, and that's probably the best way to go. From my experience with myCD63, and requiring extra boards with regulators kicking around, I think it will be the last mod I do to it (best to do it properly).
I was running under a stupid assumption that, because its a bitstream CDP, all the ICs in the chain are required (because the clock is fed back from the DAC to the decoder - the only chip I want), I would leave all the power rails in tact. But, you're right, there is a lot that can be switched off, especially the analogue output.
I guess it depends on time, and/or how big a job I want to do on the machine.

Cheers,
Phil