|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
24.140.192.106
Reading drlowmu's posting about adding the LSES filter to a "flywheel" gave me the motivation to try something similar tonight for my RH84 type test/project amp.Ended up with LCLCLC (to finals) UF4007 isolation diode + R dropper + cap//bleeder ( to driver stage) .
Filter after rectifier is:
35mH -> C (value determined by simulation) -> 2H -> 50uF -> 320mH -> 50uF -> UF4007 isolation diode + 1K R dropper -> 20uF//200K bleeder for safety.
This 3 stage LC filter to finals allowed me to trim a lot of capacitance out of the filter and get quieter operation than the 2 stage flywheel.
The AC ripple measured at the speaker posts dropped to 3mV.
The dynamic Z (JLH method) of my filter was around 300 ohms.
Edits: 10/07/16Follow Ups:
If you are looking for a really quiet B+, you can get rid of the hash chokes. The little chokes are suggested for psychoacoustic "dynamics" or something.
4H -> 50uf -> 4H -> 100uf given a critical load of 100mA to 150mA will give you a good starting point for an actual choke regulated supply. Use good chokes rated for 2X the current and low ESR caps. Download PSUII and play with it to get you to a starting point.
There is an advantage in ripple reduction in 2 LC section critical L input filters.
But that is not a huge win because the ripple is reduced by the OPT turns ratio, so the flywheel also performs well enough in that regard.
The 3 stage LC flywheel filter perform better in PSUD2 step current tests in terms of settling time and regulation at point of use.
If there are any marked differences that would appear in the baseband Hz's that would effect THD, THD+N, IMD or beat note distortion or any other operation parameter of note, I have not gone that far in testing or comparing critical vs. sub-critical L filters to make any statements pro or con.
As of now, I will stay with the 3 stage flywheel.
Thanks for the comments.
Hi Deathtube.So glad that worked for you.
As long as you measure 5 mVAC or less of ripple, with the inputs shorted, at the output jacks, you are "good to go" on our sensitive high efficiency speakers.
Just came back from RMAF where I listened for three full day's continuously to Serious Stereo's latest 2A3 DC amp, and it sounded to me like a superb amp, I am guessing, the world's best 2A3 amp effort, ever.
Unfortunately ( IMHO ) the RMAF show-goers listened to ONLY " redbook " digital, but in after-hours, we hooked up analogue, etc and it was "all there", in spades. Lovely. I was pleased, impressed.
The newest Serious Stereo amp is constructed with refinements : reduced wire lengths, and the two filter chokes have been reduced in DCR from about 10 Ohms ( Stancor NLA part ) to ( custom-made my Magnequest, to D.F.'s spec )..... 4 Ohm DCR chokes. Configured as Dennis' LSES filter.
There is a question, unanswered in my mind at THIS point in time, ( 10-2016 ) Deathtube, as to "WILL a three stage filter sound less-direct than a two stage". However, either filter is superior, music-playback-wise, IMHO, to conventional designs.
( We are talking about LSES : L1/C1/L2/C2 OR a Flywheel : L1/C1/L2/C2/L3/C3. In both, all Ls sub critical and under 20 Ohms DCR and all Cs under 50 uF. ) Messers John Swenson and John L. Hasquin , Forum Members and EEs, posted their theory and listening results about this, on this Forum, some years ago.
All one needs to do is build it, wire it well, and compare... BY EAR. Then the discussion ends !!
No need to study 100 year old EE text book theory. Use your 10,000 year old ears and decide, after all....listening is what matters.
Use PSUD to design it, check with a step-test and check for ripple. Typically, I would suggest, the output stage B+ needs to PSUD at up to 800 mVAC of B+ ripple, and the input stage B+ needs 2 mVAC or less, prefer 1 mVAC. See attached PSUD2 shot as just one example of a Flywheel-inspired filter doing two JJ 2A3-40 s in a DC stereo amp, 274BX PT.
Regards, have fun.
Jeff Medwin
Edits: 10/10/16 10/10/16 10/10/16
"As long as you measure 5 mVAC or less of ripple, with the inputs shorted, at the output jacks, you are "good to go" on our sensitive high efficiency speakers."
I am surprised you can tolerate 5mV of output ripple. I tested that on 98dB speakers and it was beyond anything I would be willing to put up with.
Chris
Thanks to a shielding issue on my TX-output linestage, I get 3-4 on the outputs terminals. With shielding, I can easily get it much lower( or at least adequately lower ). But that is just me, doing non-critical listening...if I had to tell conclusively what the amp is up to, it has to be lower...:)
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
The more sensitive our speakers, the lower the hum level needs to be.
Jeff,
I sent an email to your gmail account.
Take a look when you get time.
Glad to hear that RMAF went well for you and DF/SS.
DT 667
The more filtering stages you implement, the more ripple you will (typically) remove. What a critical choke power supply gets you is a power supply with regulation. This is a parameter of note and is easily measurable.
I'm not sure how important voltage regulation is when the PS is feeding a Class A circuit.
What's more important, in my mind, is the high charging currents caused by a cap input filter or a filter with less than a critical inductance input choke.
With either the so called LSES or Flywheel supply the input choke's Hy value is so much smaller than critical that the supply will, for the most part, act like a supply with a cap input filter.
With a cap input filter the diodes are both off most of the time and each diode in turn is only on for a very short time each cycle. This necessitates that the current peaks be high.
The power transformer still has to deliver the same amount of power to the load circuit. With the diodes only conducting for a short amount of time each cycle they have to "make hay while the sun shines" and the charging current peaks are high.
The charging current peaks through the power transformer, diodes and first cap can be 6 times the load circuit current draw and all three (the transformer, the diodes and the first cap) are being continually "hammered" on.
With a critical inductance choke input filter the charging current is a sine wave and the RMS current is about 75% of the load circuit current draw and nothing is being "hammered" on.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I have run a flywheel filter on a SE 45 stereo amp for a few years with no component failure.
The DC load was around 96mA, B+ over 500V since it was a parafeed circuit.
The Triad C56U L1 choke is rated for 2A, it is a robust part.
I also doubled up the rectifier diodes and used motor run caps for C1.
Reports of a "psychoacoustic" effect or something that will make a sound like a guitar effects processor or AM radio by using a flywheel filter in your amp is simply not based in reality , IME.
I didn't say it would cause failures.
Although it could if the "minimum required plate supply resistance per plate" in series with the rectifier is not met. Please see the thread linked below and study Rating Chart III of the 5ar4 datasheet linked. There seems to be little understanding of this by those who espouse the LSES and other such supplies.
As for me, I just like the idea of a calmer PS in my amp.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Although it could if the "minimum required plate supply resistance per plate" in series with the rectifier is not met. Please see the thread linked below and study Rating Chart III of the 5ar4 datasheet linked. There seems to be little understanding of this by those who espouse the LSES and other such supplies.
I would not use 83 MVR with the low DCR input chokes, IME.
UF4007 diodes or 5U4G, never an issue after several years of use.
Never used additional resistance with 5U4 types.
"Never used additional resistance with 5U4 types."If the DCR of the power transformer is high enough to limit the peak current to below spec of the rectifier tube used then there's no need for external added resistors/resistance.
But even when that's done the charging currents are still much higher than with critical choke input filter.
Again, I like calm.........
I placed a 35mH choke with a 5uf cap after my 5u4 and in front of my critical LCLC filter in my supply that delivers 60ma to my 300b.
The diodes, each in turn, abruptly turns on for a very short time, go very quickly to over 330ma. and then the current (through the diodes and 35mH choke) falls just as fast and that diode abruptly turns off. Each diode is only on for a short time. It's like "turn on surge" 120 times a second.
By contrast with the LCLC connected to the 5u4 each diode is on for it's entire half cycle. While on the current never falls below 15ma. and peaks at just over 100ms.
After the peak, when the rest of the filter is charged against the load, the current falls for the remainder of that diode's half cycle. Then the other diode turns on resuming, current wise, just where the other left off and continues to fall until the rest of the filter is discharged against the load and then rises to just over 100ma. where the filter is fully charged against the load.
The current in that diode then starts to fall until it's half cycle is over and the other diode picks up what that one left off. Rinse and repeat.
Very calm........
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 10/10/16 10/10/16 10/10/16
I don't think he's using a "LSES" supply. According to his first post, the supply is configured like this:
35mH -> C (value determined by simulation) -> 2H -> 50uF -> 320mH -> 50uF
The 35mH does practically nothing at 2X line frequency, so this is actually a C-L-C-L-C filter with smallish values of L. Output ripple might be higher than is typical, but that depends on a lot of factors, including ESL of the caps, DCR of the chokes and the load. This isn't how I would do it, but nothing to get too tangled up about. Output Z over frequency is probably reasonably good.
--------------------------
Buy Chinese. Bury freedom.
Atta Boy! (That is what you were looking for, correct?)
"Atta Boy! (That is what you were looking for, correct?)"
For mentioning that I tried some of drlowmu's power supply ideas, it was more like I was expecting to be treated like a North Korean political prisoner or a member of a doomsday cult from Waco, Texas.
You were not looking for suggestions for improvement. You stated that the (marginal) power supply filter you constructed was improved by adding another filter stage. Yes, that is how these things worked.
I stated that if your goal was a truly quiet supply, a choke input filter was better. I am not attacking you or the filter. Your...
"Reports of a "psychoacoustic" effect or something that will make a sound like a guitar effects processor or AM radio by using a flywheel filter in your amp is simply not based in reality , IME."
was the post that was uncalled for. Where did you come up with this crap and the North Korea stuff. Easy there pardner. Yes, the effect of the flywheel is psychoacoustic. I don't think you know what the term means. Read the link.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: