Hi.
I have been storing up my chip amp projects for what seems like an age. Mainly as I have two 300 VA toroid with 24V secondaries which is slightly higher than the 22 V secondaries which keep more margin of safety from overvoltaging the amplifier IC.
So eventually I have gotten round to just trying building with what I have and measured some voltages etc
I have a dual secondary in a centre tapped configuration, rectifier bridge, then a bank of 13600 uF on each rail.
Primary mains input is a measured 254 V...
Open circuit voltage = +/- 41 V rather too close to the max supply voltage.
But...
Output with each rail loaded to 0.25 A to simulate the quiescent current draw and output voltage is 38 V or so.
Ripple measured at ~120 mV p-p
Now...my question is...
Is this kind of ripple acceptable for an unregulated power supply for audio?
I did hesitate and initially design for double the capacitance per rail, but settled for less due to the added size mainly.
I have been storing up my chip amp projects for what seems like an age. Mainly as I have two 300 VA toroid with 24V secondaries which is slightly higher than the 22 V secondaries which keep more margin of safety from overvoltaging the amplifier IC.
So eventually I have gotten round to just trying building with what I have and measured some voltages etc
I have a dual secondary in a centre tapped configuration, rectifier bridge, then a bank of 13600 uF on each rail.
Primary mains input is a measured 254 V...
Open circuit voltage = +/- 41 V rather too close to the max supply voltage.
But...
Output with each rail loaded to 0.25 A to simulate the quiescent current draw and output voltage is 38 V or so.
Ripple measured at ~120 mV p-p
Now...my question is...
Is this kind of ripple acceptable for an unregulated power supply for audio?
I did hesitate and initially design for double the capacitance per rail, but settled for less due to the added size mainly.
I should also say...
I am open to regulating the supply.
I could regulate using something like a 317 IC and series pass BJT (2N3055 or simlar), or I have some LM12 which might make good regulators.
If anyone has anything to contribute regarding regulating using these devices I'd be interested to hear.
Thanks
I am open to regulating the supply.
I could regulate using something like a 317 IC and series pass BJT (2N3055 or simlar), or I have some LM12 which might make good regulators.
If anyone has anything to contribute regarding regulating using these devices I'd be interested to hear.
Thanks
I hadn't thought of doing this - mainly as I'm unsure where I can fit then (I etched a simple PCB as compactly as I could with large pours for rails and ground.)
It is however by far the simplest solution to try out...
Cheers.
The easiest way to add the R's is to run a blade through the trace and use SMD resistors across the gap.
Use 5 or so in parallel and the dissipation is not an issue.
If you've never used SMD components before, buy a few and try various techniques with a scrap of Veroboard. They aren't as scary as people make out.
Use 5 or so in parallel and the dissipation is not an issue.
If you've never used SMD components before, buy a few and try various techniques with a scrap of Veroboard. They aren't as scary as people make out.
Last edited:
I agree.
I have stated in the long distant past that designing a combined regulated power supply+Power Amplifier is a much more onerous task than designing a Power Amplifier.
I do not recommend that any Builder takes on the role of designing a regulated PSU for a power amplifier.
Thanks for the great advice 
I must admit designing a regulated PSU, whether discrete or IC plus transistor series pass device, does appeal to me.
But I also did wonder about stability issues from load return currents in the ground upsetting the stability.
So I agree, perhaps a regulated supply is better left as a separate project.
If anyone has any good references, papers,books or other that can help with the design of CRC filter that would be very helpful
I've read around, having never included a R in the filter bank in previous PSUs and I've come to the conclusion that C (4700uF) - R (0.22 - 0.68 Ohm) can get me a Fc of 35-50Hz.
Ive not found a equation for the second C though.
Going higher than 0.68 Ohm just creates a little too much dissipation at max current and perhaps affects regulation adversely.
So maybe an option is CRCR? Although I'm not sure I'll gain much that route other than spreading the dissipation across two Rs rather than one.
I must admit designing a regulated PSU, whether discrete or IC plus transistor series pass device, does appeal to me.
But I also did wonder about stability issues from load return currents in the ground upsetting the stability.
So I agree, perhaps a regulated supply is better left as a separate project.
If anyone has any good references, papers,books or other that can help with the design of CRC filter that would be very helpful
I've read around, having never included a R in the filter bank in previous PSUs and I've come to the conclusion that C (4700uF) - R (0.22 - 0.68 Ohm) can get me a Fc of 35-50Hz.
Ive not found a equation for the second C though.
Going higher than 0.68 Ohm just creates a little too much dissipation at max current and perhaps affects regulation adversely.
So maybe an option is CRCR? Although I'm not sure I'll gain much that route other than spreading the dissipation across two Rs rather than one.
Thanks for the great advice
I must admit designing a regulated PSU, whether discrete or IC plus transistor series pass device, does appeal to me.
But I also did wonder about stability issues from load return currents in the ground upsetting the stability.
So I agree, perhaps a regulated supply is better left as a separate project.
Going higher than 0.68 Ohm just creates a little too much dissipation at max current and perhaps affects regulation adversely.
So maybe an option is CRCR? Although I'm not sure I'll gain much that route other than spreading the dissipation across two Rs rather than one.
I must admit designing a regulated PSU, whether discrete or IC plus transistor series pass device, does appeal to me.
But I also did wonder about stability issues from load return currents in the ground upsetting the stability.
So I agree, perhaps a regulated supply is better left as a separate project.
Going higher than 0.68 Ohm just creates a little too much dissipation at max current and perhaps affects regulation adversely.
So maybe an option is CRCR? Although I'm not sure I'll gain much that route other than spreading the dissipation across two Rs rather than one.
Thanks for all your help
So as suggested by Katie and Dad I cut the trace between the two banks of Capacitors and installed a low value R. In this case 0.47.
Using a dummy load again to simulate quiescent current draw of two LM4780 IC's I now have a reduced ripple of 37mV per rail compared to 60mV on each rail with no series R.
At roughly 2.2 amps draw this increases to 270 mV. At this stage the dissipation become noticeable and the R probably hits 55 degrees C.
Time for heat sinks or a reduction in the series R to 0.33.
So as suggested by Katie and Dad I cut the trace between the two banks of Capacitors and installed a low value R. In this case 0.47.
Using a dummy load again to simulate quiescent current draw of two LM4780 IC's I now have a reduced ripple of 37mV per rail compared to 60mV on each rail with no series R.
At roughly 2.2 amps draw this increases to 270 mV. At this stage the dissipation become noticeable and the R probably hits 55 degrees C.
Time for heat sinks or a reduction in the series R to 0.33.
The LM4780's I've used benefit enormously from stable power supply.
Before we got kids I had dual mono (so two of them, one for each channel) 48v battery banks powering 2*1 channels of LM3875 and 2*3 channels of parallel LM4780.
Did not need that much power, but it sounded a bit like:
Seeing the vast amount of infinite specks of dust on a clear star filled sky when you are way out into the wilderness and there are no sources of artificial illumination for several 10's of kilometers in any direction. The amount of detail is just increased by a ridiculous amount because the noise floor is so low.
These chips have good PSRR. But a good PSU and a rock solid grounding scheme makes them much better.
Maybe I'll do the same thing when the kids are bigger, or maybe just make a foolproof case, but that'll probably just stop myself from any maintenance.
Before we got kids I had dual mono (so two of them, one for each channel) 48v battery banks powering 2*1 channels of LM3875 and 2*3 channels of parallel LM4780.
Did not need that much power, but it sounded a bit like:
Seeing the vast amount of infinite specks of dust on a clear star filled sky when you are way out into the wilderness and there are no sources of artificial illumination for several 10's of kilometers in any direction. The amount of detail is just increased by a ridiculous amount because the noise floor is so low.
These chips have good PSRR. But a good PSU and a rock solid grounding scheme makes them much better.
Maybe I'll do the same thing when the kids are bigger, or maybe just make a foolproof case, but that'll probably just stop myself from any maintenance.
Last edited:
Yeah I'd considered LM338 and the somewhat higher rated LM12CLK which I have a bunch of but after my latest tests I may just stay with an unregulated supply.
Yesterday I messed around with CRC after addimg heatsinking to the power resistors and the results are now a lot better.
Using load resistance drawing nearly 4 amps per rail and series R of 0.66 Ohms I now have around 250mV RMS ripple per rail and a output voltage of 31Vdc.
Power dissipation of around 10 W in the power resistors at 4 amps per rail load. Using a thermometer the heatsinking stabilise at about 55 degrees after an hour.
Surprisingly the small heatsinking on the 35amp rectifier bridge gets hotter!
I found this for consideration:
High current dual rail regulator kit for power amplifier or bench power supply ! | eBay
Regulated, high current, room for larger capacitors.
High current dual rail regulator kit for power amplifier or bench power supply ! | eBay
Regulated, high current, room for larger capacitors.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.