| K-amps |
I am doing some simulations in PSUID... for a 2kVA toroid used in an Aleph2.
Secs are 43-0-43AC...
Question is: for a Toroid that size, what is the approximate DCR of the secondary? (meaured from center tap to any one hot side) My guess is between 0.5-1ohm... maybe lower.
I am sitting in an Airport so I cannot measure it...
Anyone have the vaguest idea of what typical values are??? :D |
|
|
| Stuart Easson |
...a set of useful specs recently on the ksa100 thread.
For a ~1200va, 28-0-28 transformer from Bel Canto it was specified as 0.065ohm...the primary was 0.34(+0.13), depending on input voltage selected.
The secondary was wound with 14gauge wire, so if you are staying in the ballpark for voltages and current, the resistance will likely be proportional to the secondary voltage...ie 45-0-45 will have something like 0.1ohms...
If the voltage is significantly higher the gauge of the wire will need to diminish and the resistance is going to be increasing at a great rate.
HTH
Stuart |
|
|
| K-amps |
Awesome! Thanks Stuart.
Now. since you brushed that one off... let me try and challenge u...
63v 27,000uF Panasonic TH-A series cap. Approx ESR?
:cannotbe: |
|
|
| Stuart Easson |
...since you only want the approx. ESR, I'd use 0.015ohms at 120hz.
I couldn't locate your exact part, but checking the panasonic catalog for all their current ranges of 27000u, 63v working caps shows surprisingly little variation, independant of size, shape, terminal type etc. The numbers seem to range from 0.012 to 0.018 ohms...
HTH
Stuart |
|
|
| K-amps |
Since their Surge rating (for the 63wv parts) is 79vdc....
Am I ok using it with a +/- 60.5 vDC rails?... I know it is cutting close.... but those are the parts I have... |
|
|
| Stuart Easson |
Since they have specified the surge rating, it seems like it should be safe to use it...
Might be wise to check their definition of 'surge rating', if it is limited to events that are on the order of micro or milli seconds, then it might not be the right plan to use all of it, those things are really messy when they pop, not to mention loud.
Stuart |
|
|
| K-amps |
Well....
How do I go about reducing a couple of volts and keep a low impedance rail? I will bias the amplifier in the order of 20 watts pure class-A on those rails. i.e. loading the rails by about 2.25amps.... My gut is that will drop the rails by1-2vdc right there. Then I can use a CL-60 (or a couple) on the primaries of the tranny... in total my guess is I should be able to drop 2 volts...
If not; should I consider building a bridge with disctrete Diodes and use 2 in series... or even 2 bridges in series? |
|
|
| K-amps |
PSU Designer tells me that reducing the load on the PSU from 300ohms (unloaded) to about 16 ohms (approximating a 2.25amp load or 20 watt class-A bias) reduces the rails to 55vdc.... does that seem in the ball park?
Seems the droop is too much... for a 2kVA trannie feeding about 108,000uF per rail in Low ESR Panasonics dropping from 59 to 55vdc with only a 134 watt (2.25Amp) load... |
|
|
| Stuart Easson |
between a pair of 63k cans, on each rail. The choke is wound with 14gauge wire on a steel core and has Isat>5 amps, so not stressed in this scenario at all.
Drops the ripple to very low levels and reduces the DC by a couple of volts (at 4-5 amps total)...IIRC ripple measured ~10mV.
According to PSUd2 I would have needed something like 10x the caps for the same result without the chokes.
Stuart |
|
|
| K-amps |
I thought about doing a CLC, but not having worked with coils before, was about to give up.
Do you have specs on your choke? Turns/ size of donut or core etc.
Did I mention you're a life saver. |
|
|
| Stuart Easson |
Not sure if it matters, but I didn't use a fixed resistor on the output of the PSU in the designer, I used a constant current source, and then set it to 4 amps...
I'm going to redo the work in PSUd2 and see if my numbers end up different to yours or different from what I remember...
Stuart |
|
|
| Stuart Easson |
...The chokes are large-ish, but nothing compared to the caps they are 'helping' or replacing, so for really low ripple it seemed like the right thing to do.
The ones I bought were for speaker work, crossover etc. Parts express had a good selection the last time I checked. Air cored are 'best', biggest and most expensive, especially if they are wound with wire fat enough to have a low DCR. So I went with steel cored, Erse I think made them. 14gauge wire, over a steel dogbone shaped core, not sure how many turns, but each weighs 8-12oz.
Using PSUd2 after I make up the CLC using the numbers we are talking about, ie 46vAC (0.06ohm), 1n5402 bridge, 63k(0.015ohm), 2.2mh(0.5ohm), 63k(0.015ohm). I have a 58.1v DC rail with 10mV ripple, at 5 amps continuous...
With a CLC it's not just the DCR of the transformer that lowers the rail, the almost complete removal of AC ripple makes the average output DC noticeably lower. After all there is 2.5v of ripple removed by the choke...along with 2.5v DC
HTH
Stuart |
|
|
| Stuart Easson |
Quite a while ago now I bought a bagful of these critters...
http://www.partsexpress.com/pe/psho...tnumber=266-904
They are better than I remember, lower DCR etc, I'll update the model in PSUd2 and see if anything changes...
The lower DCR helps the output voltage, a little, to ~59.9
Stuart |
|
|
| K-amps |
| Great, just what I needed. A source for ready made picks! |
|
|
| K-amps |
| How does one calculate the inductance needed for a given supply? Are there rough estimates? |
|
|
| Stuart Easson |
I think I saw values in an article by NP, nowadays I start from these values and vary them in PSUd2 until I get something I can live with...
When I first used a CLC it was for a zen or Aleph amplifier that had relatively poor PSRR, so ripple reduction was critical...
Stuart |
|
|
| Zen Mod |
| quote: | Originally posted by K-amps
How does one calculate the inductance needed for a given supply? Are there rough estimates? |
what you really need to estimate or choose (it's better term for situation ) is max Rdc of choke you can accept;
after that you can try to find on market max inductance for that Rdc.
more mH ,better filter .......
I'll use ~9 mH in my Babelfish,with some 0,15-0,2 ohms Dcr |
|
|
| K-amps |
| Whats the current draw per rail of your fabelfish? |
|
|
| Zen Mod |
| quote: | Originally posted by K-amps
Whats the current draw per rail of your fabelfish? |
800VA donut , 4 x 22Vac
two bridges
5 x 10KUF - 9mH - 5 x 10KUF up and same down
supply is common for both channels
4A Bias (2 a per channel)
if you are willing to wind few chokes ,I can give you a recipe ;) |
|
|
| K-amps |
| I mean at max draw? |
|
|
| Zen Mod |
| quote: | Originally posted by K-amps
I mean at max draw? |
that's it
Babelfish is sorta Aleph J is sorta Aleph 3 is sorta single ended is sorta must be full A class..................
:clown: |
|
|
| K-amps |
Oki,
Now for the recepie you promised. |
|
|
| Zen Mod |
| quote: | Originally posted by K-amps
Oki,
Now for the recepie you promised. |
current you need is ..........? |
|
|
| K-amps |
| 37 amps at Peak power... :) |
|
|
| Zen Mod |
| quote: | Originally posted by K-amps
37 amps at Peak power... :) |
ya nutz?
:clown:
for that peak current is responsible cap bank after choke;
what is Iq in your particular amp? |
|
|
| K-amps |
| It is a high bias class-AB amp. I need it to be 2 ohm capable. 37 v at 37 amps > 1kW. |
|
|
| lineup |
| quote: | Originally posted by K-amps
It is a high bias class-AB amp. I need it to be 2 ohm capable. 37 v at 37 amps > 1kW. |
Such high currents circuits are very tricky to build
- practically
Because at those levels rails that normally have 'no resistance and inductance'
will start to behave as resistors and inductors
PCB layout and wiring plays very important role!
Instead of making one very thick rail/wire,
you may need to parallel several tracks at PCB.
---------------
It was interesting to read the building instructions
to an Elektor Amplifier project, for use with 1-2 ohms loudspeakers.
They are called Ribbon speakers, or something like that.
The Amp Power supply was only +-15 Volt. But the currents were THE BIGGER.
And very special considerations for construction and wiring had to be used,
to keep performance quality at high level ... and avoid disasters ...
I think some other members may also recall and have this article.
I have it somewhere in my Elektor magazines collection.
This power amplifier was tested with data at 0.5, 1.0 and 2 ohms loads, if I remember.
lineup |
|
|
| Zen Mod |
| quote: | Originally posted by K-amps
It is a high bias class-AB amp. I need it to be 2 ohm capable. 37 v at 37 amps > 1kW. |
I'll give you tomorrow exact data for my "5A" chokes;
they are used in ancient movie projector machines in tone bulb circuit; if memory serves me well-they are pretty cold with 5A DC on them.......
you can make 4 of them if your route is two monoblocks
or even just 2 of them -as I do- for common PS
but-if you insist on higher current....that's easy-bigger pack of iron and thicker wire |
|
|
| K-amps |
| Granted this is an absolute worst case scenario. Under normal running I'd say 600 watts max. or 15-20 amps. |
|
|
| Zen Mod |
| quote: | Originally posted by K-amps
Granted this is an absolute worst case scenario. Under normal running I'd say 600 watts max. or 15-20 amps. |
take EI84 lams ,at least 40mm package height,fill entire bobin with 1.5 to 1.6 mm dia enamell wire,make 0,4 to 0,6 mm air gap.......
and you are there
two chokes ,common PSU for both channels....?
just one thing (just now I read again first post) : how Aleph 2 can be AB amp? |
|
|
| jacco vermeulen |
| quote: | Originally posted by Zen Mod
I'll use ~9 mH in my Babelfish,with some 0,15-0,2 ohms Dcr |
Ya tha nutza.
Show me !
For 37 Amps i'd go for phatty bubblin Jantzen aircoils, PartsXp. too. Or the 800Watts 14GA Erse air core, great prices. |
|
|
| Zen Mod |
| quote: | Originally posted by jacco vermeulen
Ya tha nutza.
Show me !
For 37 Amps i'd go for phatty bubblin Jantzen aircoils, PartsXp. too. |
from what I see there ,just
http://www.partsexpress.com/webpage...109540#14gaugec
line is usefull;
everything else (thinner wire) have muuuuuuuuuuucho Rdc
:devilr:
just one catch22 - these are toroidal........nono..........
so-conclusion- where da schmuck are fat 14 gauge coils? |
|
|
| Zen Mod |
first link: too much Rdc,unless is needed for lowering PS output
second link: used without saturation? at which current? |
|
|
| jacco vermeulen |
| Yep, William Wafwaf got the same good results. |
|
|
| Zen Mod |
| quote: | Originally posted by jacco vermeulen
Yep, William Wafwaf got the same good results. |
with Torobars ,right?
:bawling: :bawling: :bawling: :bawling: :bawling:
Yacco-whatza with MA24 sssssssssccannnnnnnn?
:hot: :bawling: :bawling: :bawling:
ps-I'm so curious- with which amp you can compare sound of MA?
TIA |
|
|
| K-amps |
| How does one calculate saturation... PE rates them in watts... which without a meaningful amperage number is meaningless... |
|
|
| Stuart Easson |
IIRC the saturation current of the erse parts is documented on their website...to calculate the number needs a bunch of parameters associated with the core material that are probably not known...
I've mentioned this before, but it bears repeating: At the point in the current curve where the inductor saturates, it then behaves like a resistor as far as increasing current goes, in other words, you have a CRC power supply. Still not a bad solution and for lower currents you get the much improved smoothing of the inductor...
Another thing, I haven't tried cross coupled inductors, where the windings are bifilar woudn so the AC ripple from each rail attempts to cancel the induced voltage from the other...I would think for a class A type amp where the upper and lower rail current demands are largish and mostly equal the extra reactance would be of benefit...I've seen it used on SMPS, but I have no direct knowledge or experience of this in a 'normal' PSU, too lazy to wind my own inductors...
Stuart |
|
|
| K-amps |
| I am drooling over the 60 amp trafo !! |
|
|
| K-amps |
Built the PSU... one of the more stable ones I have built.
Non-signal rail steady at 59.5vdc. Full power into 2 ohms** at clipping and the rails sag only to 56.5vdc... I have never seen this in an amp I owned before.
3x Panasonic 27,000uF paralleled. Bypassed with 1000uF, 120uF and 0.5uF per rail.
CL-40 on the primary.... no smoke yet :-)
** Four 8 ohm woofers in parallel... yes they were bottoming at times. I reckon I am getting 600w into 2 ohms RMS. Have not done more measurements. By the way, I have not yet installed any inductors. |
|
|
| K-amps |
Actual results: 8 ohm clipping 190.34 watts (600Hz sinewave). At 4 ohms 330.31 watts RMS. PSU droop at 8 ohms 57.2 vdc (from 59.5vdc).
PSU droop at 4 ohms 56.7vdc.
Now while (because I think I biased it 5 watts class-A plus the higher device temps i.e. 40C) the sound is softer than before, there is more midbass. I am not sure if this is good thing or not but the 100-200Hz is pronounced. Is this because of the PSU going from 12000uF per rail ro over 82000uF per rail? |
|
|
| jacco vermeulen |
| quote: | Originally posted by K-amps
Non-signal rail steady at 59.5vdc, into 2 ohms sag only to 56.5vdc Getting 600w into 2 ohms RMS. |
Very impressive, Arif.
A 2 Ohms power number between 2.5 and 3 times the 8 Ohm continuous power is already good.
This is with the 2KVA transformer ? |
|
|
| K-amps |
Yes with the mystery 2kva trafo that looks like 1kva :-).
The caps are Panasonics. 105deg: 63wv, 79v surge, 27000uF CE THA series. |
|
|
| K-amps |
Close-up.
The caps were not the screw mount types so I had to jury-rig some bare copper busses, Seem to work fine.
At full power the rail sag (measured from cap junction to the collector of the OP device is 0.39mV ). |
|
|
|
