Looking for some PSU advice ...

[galac]

Hi @ll,

hopefully you give me a favour and take a quick look on the psu schematic for the 4 KT88/KT120 output tubes. No need to think about driver tubes - they'll get a seperate power supply.

I modelled the behaviour and voltages with the help of PSU II - a really great tool! So the filter section should be ok.

Beyond that i have some questions. maybe you are able to help me out.

1. Due to the fact i only have one secondary i use one (hybrid) bridge for the two channels and have to parallel the rectifier (cause tube current will be up to 250mA per channel). another possibility would be two bridges each with one 5AR4. I checked a lot of schematics and i've only seen one with two bridges from one secondary. Most use parallel tube rectifiers. What are the pro's and con's?

2. The spec sheet of 5AR4 says that with cap input (although 10µF is not that hefty) i have to use a series plate resistor (R1, R2 in the schematic). Will this be influenced by the use of ss diodes in a hybrid bridge?

3. The filament of both 5AR4 will be fed by only one winding of the seperate filament transformer (it has enough current capability). Any problems with that?

Thanks in Advance & cheers
galac

[ryuji]

I use these rectifiers myself, two in paralell if you can handle the filament draw should give you 800 mA max. If you can choose, buy a late lot number. They seem to last longer
http://ryujiwarui.com/data/5c8s.pdf
I got a psud2 model if you want

[DF96]

General rule is one bridge per secondary. Almost all attempts to get round this rule end in problems. So fine as it is.

Your resistors R1 and R2 are in series, so it is their combined resistance (with the secondary and reflected primary resistance) that counts. You could use just one resistor of twice the value, no need for two separate ones. However, when doing calculations, bear in mind that the valve datasheet will assume a centre-tapped winding so all the current goes alternately to one anode at a time. The SS diodes merely switch the other end of the secondary so have no effect on resistor value.

[Vinylsavor]

Hi Galac,

if you want to parallel 2 5AR4s I would recommend to parallel them the other way, half of one rectifier in parallel to half of the other tube, so each of the tubes has it's plates connected to separate sides of the secondary. As you drew it, the cathodes have to supply high peak currents. The other way, the peak current of the cathodes will be half but twice as many pulses, should be easier on the tubes.

Alternatively consider TV damper tubes, 6AU4 will work, 6CJ3 or 6CG3 can even supply more current. If you don't mind a top cap EY500 or PY500 will work nicely. These are also much cheaper options.

Best regards

Thomas

[DF96]

Interesting thought, but won't the area of each cathode facing each anode do exactly the same amount of work either way? In most rectifiers each anode faces half the cathode, so the only saving will be in peak current in the cathode lead (which is unlikely to be the limiting factor?).

[Vinylsavor]

Hi!

yes you are right! I was thinking too fast. Still I would think it might be better to connect them this way to ease things a bit, but as you mentioned the effect will not be dramatic.

Thomas

[galac]

Hi @ll,

thanks for your fast answers. It's a good feeling to know to be in right direction.

About the mentioned alternatives:
i've had the same idea on 6CH3 & 6CJ3 but i didn't find a ressource in Europe. Including shipping from US they aren't that cheap and it's not easy to get 3 (+1 for safety) of the same brand. Often they sell 6DW4 with to low peak and current rating.

6CG3 didn't work cause peak current is to low and i didn't like to use 4 of them ^^

5C8S is a very nice looking one but voltage drop & filament demand is to high.

galac

[Vinylsavor]

Hi!

Originally Posted by galac

i've had the same idea on 6CH3 & 6CJ3 but i didn't find a ressource in Europe.

Ask me
I have hundreds and can send you some at low cost.

Originally Posted by galac

6CG3 didn't work cause peak current is to low and i didn't like to use 4 of them ^^

You need more than 1A peak current ??
Even if paralleled 5AR4 have more according to the spec sheet, I'd redesign the filter network to reduce peak currents. I doubt current production 5AR4 can handle such peak currents reliably over a long period of time.

If you want to go extreme, check out the 866A
I wrote about it on my blog a few days ago

Best regards

Thomas

[galac]

Hi Thomas,

yes - you are right. I didn't consider the difference between the NOS specs and current production.

I'll send you a PM about the 6CJ3 ;-)

Cheers