This is a question from a newbie who will be borrowing a scope for this experiment.
I have 5 different PS to test, all set to about the same output voltage. What am I looking for to select best PS for a pre-amp or phono stage or a headphone amp that run on 15-18v bipolar supply?
What's the criteria, can I use a scope to see it and generally is this the right approach of selecting a power supply prior to doing the listening tests?
Thanks!
I have 5 different PS to test, all set to about the same output voltage. What am I looking for to select best PS for a pre-amp or phono stage or a headphone amp that run on 15-18v bipolar supply?
What's the criteria, can I use a scope to see it and generally is this the right approach of selecting a power supply prior to doing the listening tests?
Thanks!
Well, output noise is the obvious measurement, but if the power supply is well designed the noise will be beyond the oscilloscopes input range, a dedicated instrument capable of noise measurements will be required.
I don't see the point of listening to a power supply, surely if the power supply is so poorly performing it would be obvious on a oscilloscope. Additionally a well designed amplifier will also have its own inherent pssr (power supply rejection ratio) further reducing any impacts a power supply will introduce. So that leaves the criteria for a supply to power a pre-amp, phono stage or a headphone amp to be the appropriate voltage regulation, current capability and low noise (more important for a phono stage).
I don't see the point of listening to a power supply, surely if the power supply is so poorly performing it would be obvious on a oscilloscope. Additionally a well designed amplifier will also have its own inherent pssr (power supply rejection ratio) further reducing any impacts a power supply will introduce. So that leaves the criteria for a supply to power a pre-amp, phono stage or a headphone amp to be the appropriate voltage regulation, current capability and low noise (more important for a phono stage).
You can build your own out of discrete s, but LM317/LM337 is very hard to beat if well implemented.
If you want closer to discrete s, I have some SG1068 regulators somewhere.
If you want closer to discrete s, I have some SG1068 regulators somewhere.
You should load each power supply to the approximate current that will be drawn in actual use. Use the scope to look at output noise and ripple but beware because scope grounding and connection methods can make a big difference to the measured result.
Also measure the exact output voltage (no load) on a meter and then add a load and see what the percentage change is. You must measure this at the power supplies output points and not at the end of any wire connections to the load.
Also measure the exact output voltage (no load) on a meter and then add a load and see what the percentage change is. You must measure this at the power supplies output points and not at the end of any wire connections to the load.
For these kind of measurements, its advisable to make a little aluminium box, with a BNC out for the scope, two banana inputs for the supply.
But Really, A 7815 fed from clean DC produces less than 50uV of output noise. You need some serious kit to even measure that.
But Really, A 7815 fed from clean DC produces less than 50uV of output noise. You need some serious kit to even measure that.
The oscilloscope may show noise you won't hear - like high frequency spikes beyond the audio band. It may hide noise you can hear - like the noise generated by linear regulators. All in all an oscilloscope is not very helpful in this instance.
You could listen to the supply with some audio amp - any guitar practice amp will do.
Then you may hear some hum (unregulated supply) or noise (regulated supply) - or both. This method only serves direct comparison and does not output any useful technical data. Nonetheless comparing this way different units may be insightful to choose your own poison.
You could listen to the supply with some audio amp - any guitar practice amp will do.
Then you may hear some hum (unregulated supply) or noise (regulated supply) - or both. This method only serves direct comparison and does not output any useful technical data. Nonetheless comparing this way different units may be insightful to choose your own poison.
It can show you ripple, which is always handy, and also if things like rectifier spikes are getting through if you enable peak detect more at slower time bases.
He very careful with your probe, especially is you use a short ground wire, caps hold a lot of energy!
He very careful with your probe, especially is you use a short ground wire, caps hold a lot of energy!
bucks bunny,
>>It may hide noise you can hear - like the noise generated by linear regulators.
Why will this noise not show up on the scope? Can it be filtered?
Is there such a think as sagging (micro sagging) of the power supply rails under music? transients?
>>It may hide noise you can hear - like the noise generated by linear regulators.
Why will this noise not show up on the scope? Can it be filtered?
Is there such a think as sagging (micro sagging) of the power supply rails under music? transients?
"Why will this noise not show up on the scope? Can it be filtered?"
Take a typical LM317 voltage regulator, it has somewhere in the order of 450uV of noise on its output, a oscilloscope probably won't display this noise as it will be below the sensitivity of the inputs. Now lets look at a more recent regualtor the TPS7A49, it has in the order of 20uV of noise on its output, an oscilloscope will not be able to measure this low level signal.
"Is there such a think as sagging (micro sagging) of the power supply rails under music? transients?"
This is the load transient response of the regulator, in simplified terms, since the regulator cannot respond in zero time to an increase in load current, the output voltage will temporarily drop until the internal mechanisms inside the regulator catchup and increase the output voltage to compensate.
Voltage regulator datasheets will specify the load transient response and under what conditions it is measured.
Take a typical LM317 voltage regulator, it has somewhere in the order of 450uV of noise on its output, a oscilloscope probably won't display this noise as it will be below the sensitivity of the inputs. Now lets look at a more recent regualtor the TPS7A49, it has in the order of 20uV of noise on its output, an oscilloscope will not be able to measure this low level signal.
"Is there such a think as sagging (micro sagging) of the power supply rails under music? transients?"
This is the load transient response of the regulator, in simplified terms, since the regulator cannot respond in zero time to an increase in load current, the output voltage will temporarily drop until the internal mechanisms inside the regulator catchup and increase the output voltage to compensate.
Voltage regulator datasheets will specify the load transient response and under what conditions it is measured.
>>the output voltage will temporarily drop until the internal mechanisms inside the regulator catchup and increase the output voltage to compensate.
Is this brief sagging possible to see on the oscilloscope ?
Is this brief sagging possible to see on the oscilloscope ?
You will see it up to a point but this kind of thing highlights why you should not (and I'm not saying you are) use oversized caps on a regulator output.
If you want to evaluate regulator performance then make a simple circuit to add a pulsed load to the PSU. Something like a 555 timer running as an astable and driving a FET and resistor as a load.
If you want to evaluate regulator performance then make a simple circuit to add a pulsed load to the PSU. Something like a 555 timer running as an astable and driving a FET and resistor as a load.
It depends, factors such as the amount of variation in load current and how quickly the load changes from a high to low state. For example take the kind of amplifiers in your first post, they will most likely be operating in class A so the current drawn from the power supply will be fairly constant and any variation in supply voltage due to transient load variations will be in the order of microvolts.
Another example is a high current regulated power supply for a power amplifier, it will be supplying a couple of hundred milliamps in its quiescent state and high current to drive the loudspeaker at high volume. The current from the power supply needs to change from low to high quickly, in this case the voltage may vary by some millivolts or more.
All regulated power supplies will exhibit some load induced voltage variation. This is due to the fact it takes a finite response time internally in the regulator to compensate for any changes in load current.
*The above does not apply to a typical bridge rectifier and filter capacitor type power supply as these are not electronically regulated.