![]() ![]() |
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
195.93.21.66
In Reply to: SLA cells are cheapie like borsch. posted by cheap-Jack on May 8, 2007 at 08:13:16:
SLA cells are indeed one way to go, but I didn't find they sounded better than a good current source filament supply when I tested a few different supply types. A supply consisting of two voltege regs followed by a common mode choke was pretty good. You'll need a heatsink with the 26 because the filaments are 1 amp, and the size will depend on the filament supply and the voltage dropped by the devices you use, but be prepared for something of a decent size.
![]()
Follow Ups:
This 26 is very similar to 27 I built which is outstanding w/ my 300B PP amp open baffle setup.Simple cap coupled jobbie that really sounds good.I am using all wood chassis w/ large copper ground buss as I did with the 27. I see my plate voltage is much lower than other schematic's I've seen,..? The simple filment supply shown is what I will try 1st w/ even more uf, about 300k per filament, might add a choke to filament supply also.?
![]()
![]()
Hi!nice simple design. What is the input level you expect the 26's grid will see ? Nominal CD level would drive the grid positive already.
If you stick with this op point, may I suggest to give 'filament bias' a try ? Connect the filament supply to one filament pin and the other filament pin to ground through a 1.5 Ohm resistor. This will give you the same grid bias without the need to resort to a battery. However in this arrangement the filament supply must be very quiet. I'd recommend a cutrrent source, or better yet a choke. Make it LCL, if you have the iron available.
Best regars
Hi.an effective decoupling filter is added downstream of the regulated PS, which is crucial for phonostage PS.
Nope they don't sound at all good this way decoupling or not ! About the last thing you need is decoupling , what is required is isolation . Chip regs (adjustable types) are best used as current regs but everytime I find it better using LCL . Also for critical filament (not heater) applications , current reg connected IC's can be noisy . Filament supplies are critical to good sound of DHT's , much more so than with heaters but practices for filaments can be implemented to good effect . A typical cap input supply filament followed by an LT1085 type chip can much be improved by changing the supply to choke input and improved once again by removing the chip and replacing it with a choke . Not all chip regs are equal either . Chips made using legacy technology such as LM317 sound awful and are noisy as CCS , LM108x and LT108x series are better but not as good as a choke . For phono stage use which you mention ,(which normally use indirectly heated tubes) balanced supplies provide optimum performance rather than the type of regulator chip
Hi.Or it just your own way of describing it?
c-J
it's LCL and I just have ! CLC does not isolate the filament from the last cap in the filament supply . The final choke isolates the smoothing cap from the signal which passes through the filament . For reasons which I do not understand , a solid state CCS , with it's huge impedence sounds inferior to a passive LCL supplied filament which relies on the reactance of the final choke . This is a well known practice of DHT afficionados . You obviously have never tried or even considered this
For reasons which I do not understand , a solid state CCS , with it's huge impedence sounds inferior to a passive LCL supplied filament
DN2540 and LM1084 for 250mA fils running from a battery supply , had problems with the drop across the DN2540 . LM1084 and LT1085 for currents above that . I'm yet to try mosfets . So far I've found that the CCS reg followed by a cap to 0V and a choke feeding the filament provides the best sound when a chip reg is in the chain . LM317 sound horrible , are noisy and the filament needs to be bypassed with a cap . The heavy , bulky LCL approach wins hands down every time ;)
![]()
Thanks for the info. The LM317 is indeed horror. Can I ask what kind of choke you use and where do you get them from? Thanks.
LM317 is fine for heaters ;) Chokes are mainly from radio rallies , where else could you find a 10mH 10A choke for example ? So far I've been lucky and been able to pair up many of the singles I've found . Ebay is also a good place to find stuff , although postage price can make some deals unattractive . I'm beginning to think that rewinding and gapping mains transformers may be a worthwhile alternative . Chokes in my linestage are 280mH 2.7A (more like 100mH at the 1A I'm using these at) from Display Electronics in the UK
![]()
So what would be a good LCL filament supply for a 26..? Just wanting to learn.
![]()
what filament chokes do you have in stock or access to ? DCR of a hot 26 filament is 1.5 ohms , finding an isolation choke with sufficient reactance should be fairly easy . Just treat the LC part as you would a choke input supply . My main advice regarding implementing this is to use an outboard supply , keep all the iron well spaced and well away from the valves , use schottkeys for rectification . You could always go the silly route and use a pair of CV5 MV rectifiers instead of the schottkeys if you want to avoid solid state altogether :)
![]()
The MV route has me intrigued, I am a silly type guy anyway. So how ould one implemiment the cv5 MV..? & what is the equivalent # of a CV5, I really like running the vr75's for reg, so why not go MV..even if a hassle, I care not.. Chokes I have on hand are 1- Hammond 193L-5H @ 300ma, 2- 17H with unknown ma @ 300ohm DCR made by National Radio Company & are potted. I have access to any of the new production Hammonds & can source what ever might be needed.I also have the 2- 9H 40ma chokes that are in my 27's PS that I could use also.
![]()
You need low inductance , low DCR , high current chokes rather than the other way round . CV5 is a beast , 4 volt 10 amp filament , mogul base and are about the size of an 845 . I have a pile of CV5 going spare if you're keen ;) I have never used an MV for filaments , although a pair should easily light up a 26 . Definitely silly 'silly' , 80 watts + CV5 volt drop all to light up 1.5V @ 1A :) Maybe a pair of 866A would be a more conservative option ? (hehehe)
![]()
| ||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: