I am new to this and have spent the last few months reading manuals and postings. I have a few questions about power supplies in power amps.
Aleph service manuals show (for as far as I can make out) only capacitors between bridge and amp circuit.
KK PCB Design uses a CLC design:
ALEPH 3 POWER SUPPLY
In the F4, F5, BA-1 and BA-2 manuals a CRC design is shown.
Why the different approaches?
Aleph service manuals show (for as far as I can make out) only capacitors between bridge and amp circuit.
KK PCB Design uses a CLC design:
ALEPH 3 POWER SUPPLY
In the F4, F5, BA-1 and BA-2 manuals a CRC design is shown.
Why the different approaches?
CRC is used becuse of its price. and one thing is left out on the KK-PCB circuit. the bleeder resistors in paralell with the capacitors.
Sorry guys, I am still confused.
The power supplies in the original Aleph amps (according to the service manuals) used only capacitors.
I cannot imagine that the cost of the inductors was the reason they were omitted because I cannot imagine Nelson using expensive "heatsink cases" for these commercial products but then not spending a bit extra for sound-improving inductors. So why did Nelson use C only instead of CLC (or CRC for that matter)?
The power supplies in the original Aleph amps (according to the service manuals) used only capacitors.
I cannot imagine that the cost of the inductors was the reason they were omitted because I cannot imagine Nelson using expensive "heatsink cases" for these commercial products but then not spending a bit extra for sound-improving inductors. So why did Nelson use C only instead of CLC (or CRC for that matter)?
Thanks for that Atilla. I know the original Aleph 3 sounds great (I own one) but it is interesting to see that KK PCB Design used CLC in the Aleph power supplies and Nelson uses C in the Alephs and CRC in the F4/5/BA-1/2.
The original Aleph 3 had 8 x 22000uF and 25V rails.
The Aleph 30 had 2 x 6 x 10000uF and 26V rails.
I do not think the 1V difference on the rails is an issue.
In the Aleph 30 manual, Nelson wrote:
I understand the increase in the number of caps: From 8 to 12. I also understand that the higher voltage rating would be beneficial. What is difficult to get my head around is that the total capacitance was reduced from 176000 to 120000. Could this just be because the parts were that much better that less "nominal" capacitance was needed?
Please understand I am new to this. I have worked in IT for the last 30 years where everything is much easier to understand than this analog world! I am trying to understand how and why these analog things work the way they do. In order to do that I am trying to find patterns or rules. For power supply caps I have come up with:
a) higher voltage rating = good
b) better quality caps = good
c) more caps is better than few caps (I know, not always but I assume that if you used caps from the same manufacturer/series you would you would rather use 8 x 10000uF than 2 x 40000uF)
Perhaps my question can be rephrased:
How do you design a power supply? Is it just a matter of trial-and-error and with a lot of experience you'll get a reasonable choice the first time and take it from there (artistry) or is there a more (dare I say) scientific approach?
KK's schematics can hardly be of any relevance
anyway - go to Nelson's site and download all pdfs and read them
or rephrase your question once more , keeping one particular amp/PSU as object of asking
shortly - that issue is pretty complex ; to have fastest PSU , you need to decrease amount of capacitance and series losses ;
in same time - gto have smallest ripple and highest TES possible , you need great amount of capacity
my practical preference is pure and huge C bank , but xformer must be heavy chunk ;
my ideal preference is choke input filter (LC) but that's expensive in both money and space , and demands proper mileage
anyway - go to Nelson's site and download all pdfs and read them
or rephrase your question once more , keeping one particular amp/PSU as object of asking
shortly - that issue is pretty complex ; to have fastest PSU , you need to decrease amount of capacitance and series losses ;
in same time - gto have smallest ripple and highest TES possible , you need great amount of capacity
my practical preference is pure and huge C bank , but xformer must be heavy chunk ;
my ideal preference is choke input filter (LC) but that's expensive in both money and space , and demands proper mileage
Well, first of all, let's point out that the Alephs have a LOT of capacitance in the power supply. Quite a bit, really. I'd be more than happy to leave it there and not touch it any further.
Next - more smaller capacitors, connected in parallel, are indeed better than 1 big one. This is because of the Equivalent Series Resistance of every capacitor, a characteristic you can find in the datasheets. When you parallel caps, this reduces severely. Better and bigger caps tend to have better ESR ratings too.
You can calculate the performance of you power supply yourself, if you've got all parameters. You can also simulate it in simulation tools, there's even a dedicated tool for that - Duncan PSUD.
You can read up online in different places on how that's done. Or, if you're interested a good start is the 'Practical Electronics for Inventors' book. It's not very advanced, but it's a good intro to ... everything, really.
As for the art part .. well. You know the saying about opinions and ********, right?
Next - more smaller capacitors, connected in parallel, are indeed better than 1 big one. This is because of the Equivalent Series Resistance of every capacitor, a characteristic you can find in the datasheets. When you parallel caps, this reduces severely. Better and bigger caps tend to have better ESR ratings too.
You can calculate the performance of you power supply yourself, if you've got all parameters. You can also simulate it in simulation tools, there's even a dedicated tool for that - Duncan PSUD.
You can read up online in different places on how that's done. Or, if you're interested a good start is the 'Practical Electronics for Inventors' book. It's not very advanced, but it's a good intro to ... everything, really.
As for the art part .. well. You know the saying about opinions and ********, right?
I agree that KK's choices are irrelevant, but I was intrigued why he would go to CLC when Nelson himself used C only (and the service manuals were available when he made the boards). Why not just copy Nelson's design (like he did with the Aleph circuits)?
To be honest, I only really started thinking about this issue because I had made a mistake and bought 32 x 1000uF ELna SIlmic II to build 2 x stock F4.
I asked about it on the F4 thread and buzzforb answered that I needed much more capacitance. I then looked things up in the Aleph and other manuals, and well, did not understand why things were done they way they were done and thought it better not to foul up the F4 thread, hence this new thread.
I would not dream of modifying my original Aleph 3. I have never opened it. I have been curious, but must admit that being confined to a wheelchair and unable to lift the thing ensures I leave it alone!
Thanks for that. I am downloading PSUD2 as we speak!
I assume you mean this book?
Practical Electronics for Inventors : Paul Scherz : 9780071452816
I'll get started on my (PSUD2) homework!
Albert
PS Some might even say it is magic.
I went to Convergent Technologies (a Unix hardware manufacturer) in Santa Clara (Silicon Valley) in 1985 for a Unix training course. I met Carl Reisinger. His business card said:
I have downloaded PSUD2 and have started on entering my PSU, but am afraid I am stuck.
First step is setting the transformer parameters. I have an Amplimo transformer and found 18V, 8.33A and 5% regulation as parameters to work out the source impedance.
http://www.amplimo.nl/images/downloads/ds standardrange/77014.pdf
Next step is the rectifier. Mine (Vishay GBPC3502-E4/51) is not in the list. It is also not in the online database. So, I need to enter a number of parameters. I think the values for the following are:
- Name = GBPC3502-E4/51 (anything would do of course)
- Vpiv = 200
- Ipks = 400
- Ipkr = 35
- Dres = ?
- Vlaw = ?
- Vfac = ?
http://www.vishay.com/docs/88612/gbpc12.pdf
Same goes for the other two.
Any help is appreciated.
Can someone check my PSU?
Hi,
I do not have any F4/F5 PSU boards, so have decided to use (I think the term is) perfboard. I will be using 14AWG silver plated stranded copper wire for the connections shown.
The main caps are 15000uF/63V Panasonic (ApexJr). I wanted to use some Elna Silmic II (1000uF/35V) as well, as they were so highly praised by Nelson. I included 4 x 1000uF per rail. Don't know if it will make a difference, but worth a try. The resistors are 3W Panasonic. Thermistor is CL60.
Can someone please check my layout?
Thanks,
Albert
Hi,
I do not have any F4/F5 PSU boards, so have decided to use (I think the term is) perfboard. I will be using 14AWG silver plated stranded copper wire for the connections shown.
The main caps are 15000uF/63V Panasonic (ApexJr). I wanted to use some Elna Silmic II (1000uF/35V) as well, as they were so highly praised by Nelson. I included 4 x 1000uF per rail. Don't know if it will make a difference, but worth a try. The resistors are 3W Panasonic. Thermistor is CL60.
Can someone please check my layout?
Thanks,
Albert
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.