Hi to all,
my question is rather simple:
I am bulding a CD-Player with DAC and I could screw the TO220 voltage regulators to the heatsinked,
very massive chassis - instead of leaving them on the servo pcb, close to the load (and the electrolytic caps btw).
The setup would run much cooler as the heat is dissipated directly via the cabinet.
Length of the wires will be about 23cm, 9 inches.
The voltage regulators for the seperate DAC circuitry could be kept on the pcb as the DAC´s regulatoprs do not run hot.
Now I remember Philips did this with it´s early CD-players.
All voltage regulators were on a seperate board (i.e CD-104 / CD-304) and connected
with unshielded wires to the main boards.
Length of the wires must be about 4-5 inches.
With this setup, one regulator did feed
the servo as well as the audio circuitry.
To make things "worse" there was even no seperation of digital and analog circuits.
But playing a silent signal from a test CD shows no audible noise at all
with the players mentioned.
So I assume it is not critical to keep the voltage regulators away from the load,
maybe with some extra decoupling close to the load?
All the best,
Salar
my question is rather simple:
I am bulding a CD-Player with DAC and I could screw the TO220 voltage regulators to the heatsinked,
very massive chassis - instead of leaving them on the servo pcb, close to the load (and the electrolytic caps btw).
The setup would run much cooler as the heat is dissipated directly via the cabinet.
Length of the wires will be about 23cm, 9 inches.
The voltage regulators for the seperate DAC circuitry could be kept on the pcb as the DAC´s regulatoprs do not run hot.
Now I remember Philips did this with it´s early CD-players.
All voltage regulators were on a seperate board (i.e CD-104 / CD-304) and connected
with unshielded wires to the main boards.
Length of the wires must be about 4-5 inches.
With this setup, one regulator did feed
the servo as well as the audio circuitry.
To make things "worse" there was even no seperation of digital and analog circuits.
But playing a silent signal from a test CD shows no audible noise at all
with the players mentioned.
So I assume it is not critical to keep the voltage regulators away from the load,
maybe with some extra decoupling close to the load?
All the best,
Salar
Last edited by a moderator:
It is an old design from 1986, a Sony CDP 103. 11Vin, 5v out, 7805, decoupled with 2200µF.
Original heatsink on the main pcb dissipates 14k/W but becomes very hot.
Load (ICs) is not decoupled.
Very common with japanese manufacturers back then.
Not with the Philips CD 304 I mentioned before:
Power regulators outputs on the power pcb are decoupled with values between 22µF and 33µF.
On the remote servo / DAC bords, ICs are decoupled again, with the combos of 4.7Ω/47µF/47nF.
I do not now if the "heavy" decoupling of the Philips design was a result of having longer wires between boards.
Original heatsink on the main pcb dissipates 14k/W but becomes very hot.
Load (ICs) is not decoupled.
Very common with japanese manufacturers back then.
Not with the Philips CD 304 I mentioned before:
Power regulators outputs on the power pcb are decoupled with values between 22µF and 33µF.
On the remote servo / DAC bords, ICs are decoupled again, with the combos of 4.7Ω/47µF/47nF.
I do not now if the "heavy" decoupling of the Philips design was a result of having longer wires between boards.
Last edited:
Just checked a later (and cheaper) Philips CD-160. It uses the same chipset
as the CD-304/104 but the the power regulators are now on the main board,
as well as their load, with much shorter traces.
Decoupling in the CD-160 is the same as described above in the CD-304,
where the distance between regulation and load is much longer
and even on seperated boards.
I assume long distances between regulation and load were a
no-brainer for the engineers back then as well as today...
as the CD-304/104 but the the power regulators are now on the main board,
as well as their load, with much shorter traces.
Decoupling in the CD-160 is the same as described above in the CD-304,
where the distance between regulation and load is much longer
and even on seperated boards.
I assume long distances between regulation and load were a
no-brainer for the engineers back then as well as today...
Last edited:
Regulators need to be well decoupled or they can oscillate.
If you have a long pcb track to the load you can get voltage variations at the load as the load changes so you need to decouple the load as well.
LDO's are by far the worst to work with as the output is from the collector and not the emitter on the pass transistor. I have seen me add up to 47uF on the output to tame them.
Distant regulator and good local decoupling should work fine I suppose. Be carefull, too low ESR of decoupling capacitors might cause instability.
On the other hand if you need local regulator but have no room for heatsink, you can use distant regulator (or just transistor plus zener diode) to drop voltage first to 7V and input that to the local regulator, which would not dissipate much now, because of small difference between Vin-Vout.
On the other hand if you need local regulator but have no room for heatsink, you can use distant regulator (or just transistor plus zener diode) to drop voltage first to 7V and input that to the local regulator, which would not dissipate much now, because of small difference between Vin-Vout.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.