a little psu problem

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

e.oelberg

Well-known member
Joined
Aug 14, 2004
Messages
791
Location
Berlin
I do have a nice very low noise psu on my bench. It puts out clean 7.3 Volt and it is not adjustable and nothing to fiddle with. I'd like to use this psu for a tube filament of say 6.1 V .... I thought about putting a lm317 after it, but the voltage drop is to low for a lm317.
I thought than doing simply a voltage divider, but maybe there is a much better solution.

anybody has an idea ?  Nicholas
 
Two of them, actually :) But since the voltage is so low, any member of the 1N400x family would work.
(1N4001 = 50v reverse voltage, 1N4007 = 1000v)
 
Doesn't the 78xx series have a minimum drop of (over) 2v or so? And you'd need a preeeeetty beefy zener, for a 1-300mA fillament draw, wouldn't you?
 
+1 for a 7806 followed by a 1uf 35v tantalum and the same before (close to the regulator)
put a 4700uf 35v electrolytic next the diodes bridge (4 diodes schottky)
don't forget 100nf ceramic for diodes decoupling and a heat sink for the regulator
 
Yep, all variants have a (typical) voltage drop of 2v - https://www.fairchildsemi.com/datasheets/LM/LM7815.pdf

So if you go the 7806 route, you'll be using it outside of specs, with non-guaranteed results.

Can't beat the simplicity of two series diodes though, imho ;D
 
if you have a transformer with a least 6,3v AC  (linear psu) and a suitable power, then go to the 7806 and rework the design with 4 schottky diodes, it will work...higher AC voltage is better as you can even put some RC before and/or after the regulator...
 
Silent arts, if you have a secret to produce DC without AC (except with battery) then PM me  :)
Nicholas wants a "very low noise" heater psu, he needs to redesign his psu (if it's not a switching psu) as i wouldn't recommend him to put two diodes...
Or, a simple voltage divider at the output will produce less noise...
 
To get 7.3V DC (that's unloaded, i guess?), you'd need about 5V AC.

And what exactly is wrong with 1-2 series diodes, in order to drop around 1v? :) Are you going to argue it introduces noise?
Fine, just one diode with an RC filter after it, big deal :D

PS: Have there been any studies / tests made for the "heater ripple rejection"? I'd be quite curious about if and how sensitive the output of tubes is, with respect to imperfections in the heater voltage...
 
that's called commutation noise of the diodes...
nothing can prove that he has less than 6,3v AC in his psu...
noises in the heater are critical with microphones (or other high Z preamps), especially if the bias is derived from the heater voltage
 
Communication noise is rather high in this thread, isn't it.  ;D

Might be good to read some basic information given in the first post again:
e.oelberg said:
I do have a nice very low noice psu on my bench. It puts out clean 7.3 Volt and it is not adjustable and nothing to fiddle with.

Nicholas is looking for the best way to drop 7,3v DC to 6,3V DC, and he can't or doesn't want to fiddle with his given 7,3V PSU.
 
[silent:arts] said:
Nicholas is looking for the best way to drop 7,3v DC to 6,3V DC, and he can't or doesn't want to fiddle with his given 7,3V PSU.
olala , this is a lot going on here !  Yes you are right, that's what I want. So if I understand you right. a 1n4007 will give me 0.7 Voltage drop, two of the 1,4V .? ..my load is 6.3V at 200mA  (ef86).
so I will get slightly below 6 , just perfect ahh ?


 
granger.frederic said:
that's called commutation noise of the diodes...

We're (still) talking about DC here, so there won't be anything "commutating" through that added series diode ;) Not unless there's some magic evil voodoo going on...

e.oelberg said:
olala , this is a lot going on here !  Yes you are right, that's what I want. So if I understand you right. a 1n4007 will give me 0.7 Voltage drop, two of the 1,4V .? ..my load is 6.3V at 200mA  (ef86).
so I will get slightly below 6 , just perfect ahh ?

More or less correct, yes. Or, at the very least, "close enough" ;D
 
if you put some diodes (semi conductors) in series, you will decrease the  SNR , especially if the heater current is high.(it is).
if they're in series at the output , you don't have commutation noise at this place (DC) but some other noises for sure :

http://www.eng.auburn.edu/~wilambm/pap/2011/K10147_C011.pdf

check if the bias is derived  from the heater and if it's the case then you'd better redesign an low noise psu !
if it's not, noise in the heater supply is still your enemy in ultra high impedance sections (in microphones)...it depends on your tolerance level.
when i see how it's easy with new regulators and a couple of usual low cost components (one regulator and two capacitors! ) , i can't see why you wouldn't choose that option.

ps: LM317 or LT1083 have a much better noise ratio than 7806 but it's a bit more difficult to work with for beginners...
consider also the ripple rejection factor in the equation...check the datasheets...

cheers
 
you got a point frederic, I suspected the diodes to introduce noise. I still would like to stay with this design because it is so damn quiet. I would make no sense if I have a noise generator afterwards. On the other hand, if I feed super clean DC to a diode, will it still add noise ? Changing the design itself is no option, it is a complex design and I have no schematics. I know how to build a PSU, I could do that anytime, but that's not the point of this thread.
 

Attachments

  • snr.png
    snr.png
    841.2 KB · Views: 10
The simplest solution is to use RC network (4r7, 1000uF) because you already have a good stable power source and your load is constant.
 
Back
Top