Counter-intuitive to the extreme, or a bit of wishful thinking?
Also with an LT4320 ideal rectifier setup and no blasting diodes, wouldn't the 'mains harmonics' be eliminated?
What? Since when is one cap better than several smaller, esp for low ESR...?
Ya you can just parallel a bunch of cheaper smaller values and get the ESR down, but I've seen some of those installations that look awful messy to me. I kind of like his idea of just one high quality large value single cap with super low ESR instead of a herd of several. The big PEH169 and that FTcap are like 4 mOhms or something..
Redbone and William2001 , Thanks for pointing this out .
I recently bought these caps with 10mm snap in terminals for the DIYAudio Universal power supply board v.3
https://www.cde.com/resources/catalogs/380-382.pdf
380L223M050A052 22,000uF 50Vdc ESR ( 0.023 max - 0.018 min ) 35 mm dia x 50 mm height
But having a look at a PEH169 beast , the ESR is between 4 and 5 mOhms .
https://www.mouser.ca/datasheet/2/447/KEM_A4035_PEH169_85C-3316717.pdf
PEH169KU610VM(1) 100,000uF 40 Vdc
PEH169MV582BQ(1) 82,000uF 63 Vdc
It might be wise to cover the terminals with an insulator
.
Last edited:
I've based this on Jan Diddens post.
1/5th of 68,000µF is 13,600µF, so 10,000µF should be ok.
The LT4320 bridge eliminates trafo secondary ringing which is something different.
The second capacitor at 100,000µF is separated by resistors on plus and minus pole.
This one is not easily discharged, and will only have gently top-up when needed
Sure you can get low ESR by using several small capacitors, but the wrong kind of wiring or PCB layout can easily increase both the ESR and the dreaded inductance to unwanted values.
Hello,
No an input choke is not automatically a swinging choke! Hard to find a swinging choke nowadays anyway.
As i have posted before you must take care there is a minimum current being drawn. With a properly designed choke input voltage will NEVER go up to1,4 . If the dc voltage on the first cap is higher than the ac voltage of the transformer you dont have a choke input.
So use the right choke to start with, take care of the minimum current ( know how to calculate it) and then you can tell people it will be beneficial. It will take a serious investment and another power transformer.
If you said you tried a choke input using the same transformer you did not make a true choke input. If you make one better do it right from the start.
Greetings , Eduard
If you start a supply with an L you will loose a fair bit of voltage as far as I know.
If you want to learn more about this I will suggest you look at Eduard's posts in this thread and others. He knows a lot about this.
rebone
Hello,
The key element will be the input choke. Better have LCRC with a big choke ( lots of mH/ H and low dcr ) then two smaller low grade chokes in LCLC. Especially with high current 1A the input choke will cost 100€ or more. But it will work for decades to come.
Try to simulate with Lundahl LL2733 or LL2771 and you will see what you can accomplish.
Greetings Eduard
The current through all your parts will be at the same level. Making C1 bigger without an input choke and it will only get worse!!!
The key element will be the input choke. Better have LCRC with a big choke ( lots of mH/ H and low dcr ) then two smaller low grade chokes in LCLC. Especially with high current 1A the input choke will cost 100€ or more. But it will work for decades to come.
Try to simulate with Lundahl LL2733 or LL2771 and you will see what you can accomplish.
Greetings Eduard
The current through all your parts will be at the same level. Making C1 bigger without an input choke and it will only get worse!!!
Maybe it's a good idea to post this information every couple of years or so :
https://www.hammfg.com/electronics/transformers/rectifier?referer=968
Just remember that you have to compare apples to apples, so the same rectification scheme with a change only of the first element following the rectification - capacitor, resistor, or choke.
Also notice that both peak voltage and available current change with the different configurations.
Finally, you can tweak voltage and current capability of a "choke input" filter with the addition of a (small? large?) capacitor before the choke. It's a continuum of sorts, but what you gain in terms of higher voltage is offset by a reduction in "choke input-y goodness".
https://www.hammfg.com/electronics/transformers/rectifier?referer=968
Just remember that you have to compare apples to apples, so the same rectification scheme with a change only of the first element following the rectification - capacitor, resistor, or choke.
Also notice that both peak voltage and available current change with the different configurations.
Finally, you can tweak voltage and current capability of a "choke input" filter with the addition of a (small? large?) capacitor before the choke. It's a continuum of sorts, but what you gain in terms of higher voltage is offset by a reduction in "choke input-y goodness".
I'm not mentioning that you are wrong, but an 8Ohm choke sounds too high for a moderate power class A amp (bias target between 1 and 2 amperes per channel).
Looking at the power supply schematic of the WHAMMY and the circuit board , its clear why you went down the road of adding an L .
https://www.diyaudio.com/community/threads/whammy-pass-diy-headphone-amp-guide.317803/
Because its a simple addition that would improve the noise floor of the power supply rails ( quieter than a church mouse ) .
But I don't quite understand because I thought those 3 pin voltage regulators chips are pretty noisy .
https://www.tnt-audio.com/clinica/regulators_noise2_e.html
Unless you left those noise generators out .
Long story short , in this application , instead of inductors , suggest looking at a 4th order LPF .
Just a 4 pole ladder filter C R C R C R C R C C .
Also , suggest simulating in the frequency domain , playing around with R values , and look at the roll off .
.
Also note what rebone has done by putting " The Beasts " as the last caps before the power amp .
Its ESR of those caps is in the ball park of 4 or 5 mOhms . So the current can get to the load ( speakers ).
So the last caps in the ladder filter should have a low ESR .
.
Last edited:
I agree, 8 ohms is way too much for that application. The example showed what was done for a class A headphone amplifier so the resistive drop for 0.5 amps is only about 4 volts.
As to your other point, my stuff uses the most ordinary wirewounds Mouser has. Any inductance is a feature, not a bug.
Hello,
If you ask Lundahl to adjust airgap( for free) of the LL2733 to a current rating around 600mA you will get more than 400mH with total dcr of 0,85 ohm!!
I have a friend in Vietnam who used the LL2733 in a Nelson Pass single ended classA power amp, yes for choke input.
It can be done with carefully chosing the choke. And like i said if you know the current being drawn you can ask Lundahl to adjust the airgap. 200mH at 2A will turn into 800mH at 500mA . The minimal current necessary to make it work as a true choke input is MUCH smaller at 800 mH and filtering is much better.
800mH Lundahl in a LCRC will be better than let us say LCLC with two cheap 50 mH chokes. Compare the minimum current needed for 50 and 800mH. Extra current wasted through the " load resistor" will give you bigger ripple.
YES, always use this resistor if something goes wrong in your circuit voltage on the caps will not be 0,9 but 1,4 times ac that is a bog difference. With proper choke input the voltage will be kept at a certain maximum and of course will be a bit lower once your circuit is drawing current ( pretty quickly)
Greetings Eduard
If you ask Lundahl to adjust airgap( for free) of the LL2733 to a current rating around 600mA you will get more than 400mH with total dcr of 0,85 ohm!!
I have a friend in Vietnam who used the LL2733 in a Nelson Pass single ended classA power amp, yes for choke input.
It can be done with carefully chosing the choke. And like i said if you know the current being drawn you can ask Lundahl to adjust the airgap. 200mH at 2A will turn into 800mH at 500mA . The minimal current necessary to make it work as a true choke input is MUCH smaller at 800 mH and filtering is much better.
800mH Lundahl in a LCRC will be better than let us say LCLC with two cheap 50 mH chokes. Compare the minimum current needed for 50 and 800mH. Extra current wasted through the " load resistor" will give you bigger ripple.
YES, always use this resistor if something goes wrong in your circuit voltage on the caps will not be 0,9 but 1,4 times ac that is a bog difference. With proper choke input the voltage will be kept at a certain maximum and of course will be a bit lower once your circuit is drawing current ( pretty quickly)
Greetings Eduard