|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
76.4.180.69
In Reply to: RE: power supply question who can help me ? posted by horny on March 10, 2015 at 15:24:52
Hello,I had a chance to look at your supply, first stage, in PSUD2 for you. I did several iterations, and the "numbers" are summed up ( below ) in the same sequence as charts ( above ).
My general approach was to reduce C2's 47 uF amount, after seeing the odd looking decay on a 1.15% current-step at 4.10 seconds, starting ( "Stock" - the top simulation ), with your stated choice of 47 uF for C2:
Stock .. 47 + 47uF, 2.25 mVAC, 460 mS. settle, 9.8 VDC drop
Delta 1 20 + 47uF, 5.30 mVAC, 310 mS. settle, 9.8 VDC drop
Delta 2 15 + 47uF, 7.20 mVAC, 275 mS. settle, 9.8 VDC drop
Delta 3 15 + 47uF, 6.00 mVAC, 343 mS. settle, 8.1 VDC drop
Delta 4 20 + 47uF, 4.51 mVAC, 347 mS. settle, 8.1 VDC dropDelta 2 is my favorite as a smooth settle, but there is a question in my mind as to IF 7.2 mVAC will be good at all, ripple wise, to a low mu input stage. In a high mu stage, I always shoot for below 2 mVAC, prefer below 1 mVAC, to get the utmost in diction and clarity. I just don't know on a mu of 4 driver tube, WHAT we can get away with in ripple. Would have to HEAR that to tell for sure.
I can tell you this, a 2A3 with a 14 Watt plate dissipation, is a NASTY thing to listen to, because it is being thermally stressed, and, the tube will SOUND stressed all 'de time !!! I would shoot for 42 mA. at 250 VDC to run the 2A3 in my amp. That is Delta 3, using 15 uF as C2, and you will "play" with the value of the Rk ( increase its Ohm value ) to bias at 42 mA. and possibly decrease the value of C1 slightly, to get B+ value you want, and the VDC P-K you seek across the 2A3.
The interstage transformer will have a better time with 42 mA. across it also, instead of almost 60 mA. of unbalanced current !!
You are not finished with this amp my friend. The entire through-gain ( mu 4, times 2, times mu 4, or 32 times ), assuming full mu, is rather low, and it will likely need to be "pushed" and played loud, to sound its best. What if you wanna play it at medium or lower SPLs, and it simply dies ?
I can PSUD Finals, later, if there is a need to do so. Likely in the Final's B+ filter, that same C2 uF value will also need to be reduced, which is EASY cause there is little need for lowest ripple on the Finals. You are "hammering" the 300B too ... stressed sounding presentation is the result.
Have fun. If your moniker is Horny, and if you use Horns, you don't need to PUSH the tubes, one or two really good watts, well executed, will usually do !!
Jeff Medwin
Edits: 03/10/15 03/10/15 03/10/15 03/10/15Follow Ups:
Not sure if anyone is aware but there is a neat capability in PSUD2 which allows you to look in minute detail at the ripple voltage:
Generate a plot of voltage vs time with your chosen power supply design As drlowmu has done in the above examples). Then zoom out using the zoom- icon until you are showing your full voltage range on the vertical axis. You should see the exponential rise from zero up to the stable straight line final voltage; usually a second or two to reach stable voltage.
Then use your mouse and hold the mouse cursur just a little over the top of the straight line stable voltage section. Hold down the left mouse button and a rectangle will appear which you can size to just cover a little above and a little below your voltage line and perhaps a second or so of your time axis. When you release the mouse button you will see a magnified version of your straight line voltage and any ripple should be easily visible. You can do this magnification repeatedly as many times as you want, to really focus in on the ripple peak to peak. You may find that you go"off scale" on the vertical voltage axis in which case you wont see your trace and you have to drag the graph window up or down; do this by placing the mouse cursor anywhere inside the graph window then right click and hold. By holding and moving the mouse up and down you can make the view window 'slide' up and down the the vertical axis of the graph and you will eventually find your trace (may have to zoom out first to see where it is on the axis).
Apogs if you all knew this already but I have found it very useful
Cheers
Sean
Sean
Wow thank you very much drlowmu!
i wil try it
and let it now here
is it also possible to get c1 2 c2 47 and c3 100?
will this give a good supply?
or maybe no cap choke input or a very small cap as c1?
I will try also the 2a3 on the lower setting
but i can't say the amp sound stressed the sound is very good
yes i use horn speakers 116 dB with one watt
and active bass below 220 hertz
the minus 3dB point is on 20 hertz
the active bass amp is connected on the output of the 300b amp
and it is totally quiet, no hum.
There is also a preamp for this power amp with a gain of 7,5 and my source give 2 volt out
so this is 2 volt x7,5x8 is 120volt that i can give the 300b
OK,Ask and ye shall receive......
Delta 5, 47 + 100uF, 1.08 mVAC, 530 mS., 8.1 VDC drop.
Notice, the drop is only 8.1 VDC and with less wigglies.
This MEANS to me, to "fit" two series L/Cs together best, have a smaller C2 than C3, with these equal 10 HY high DCR chokes. In initial post, " Stock " 47 plus 47 uF gave us the wigglies, but also, a 9.5 VDC drop !!!!
530 mS. full settle time is a degrade.
Recovery is predictably slower with larger storage caps. Graduate engineers like John Hasquin will not use over 50 uF in a supply, for sonic reasons. Graduate engineers like Bill Swenson might wanna see a 50 mS. as a filter settle time.
In 2015, I personally am interested in filters with a smooth recovery in about 180-200 mS. maximum, using equal sized chokes.
Jeff Medwin
Edits: 03/11/15 03/11/15
Hello
i try first then the 15 uF as second cap
is it possible to leave the first cap away
like a choke load supply?
and who can i get 180/200mS recovery or is this not possible with this supply?
and a stupid question what is the VDC drop meaning ?
No, you need the small first cap with the value of your power transformer's secondary, to get the high enough voltage to run the 2A3. The 222 VDC with no C1 won't be adequate, as you can see above, and the supply is no longer looking as nice, it overshoots.
With large chokes, high in DCR, having large storage capabilities, and large caps, you can not get a fast-settling supply in a double L/C series arrangement, as far as I can tell.
WE have not even looked at the Finals' supply, have we ?? !!
The voltage drop has to do with the impedance of the supply, the less the voltage drop, for a given current change, the " better" the supply is, the lower in Z it is. Has to do with Ohm's Law R equals V over I.
HTH.
Jeff Medwin
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: