Yet another 4P1L PSE amp - review request

[andyjevans]

It depends if you connect each 4P1L in series or in parallel. Since your input is 4,2v I'm assuming series. So current is 325mA plus 325mA = 650mA. Could vary a little around that from my experience. Series connection is pins 1 and 7.

 

[rplate]

I have noticed that there is unequal voltage drop between tubes filaments connected in series (1, 7 pin) voltage drops are 4.2V and 3.7V respectively, hence my total current is less than ~650mA, should I be concerned?

 

[andyjevans]

I'm not sure what you are doing here. Each pair of tubes has its own filament supply, 2 supplies? That's necessary. So each tube connected in series, then 2 tubes paralleled together?

Do you have a few more 4P1L to test? Test them individually for filament specifications. I usually match the voltage and current in a pair of tubes.

 

[rplate]

I'm not sure what you are doing here. Each pair of tubes has its own filament supply, 2 supplies? That's necessary. So each tube connected in series, then 2 tubes paralleled together?

Wait, do I have to use 2 separate fillament supplies for 2 paralled tubes in the same channel?

 

[andyjevans]

Four tubes in 2 pairs, one pair per channel. Each pair has its own filament supply, so 2 filament supplies in total.

 

[rplate]

Ok, thats what I am doing. I hope it's more clear now.

 

[maton00]

Ok, thats what I am doing. I hope it's more clear now.

Your wiring has a apreciable high DCR or the lower filament voltage tube is a bit gassy, you can take turns and test each tube with the supply/CCS prior powering the HV to discard any issue a good 4P1L will eat ~320mA with filament in series (~4V) and the double of that in parallel at ~2V

 

[andyjevans]

Are you saying that measured separately one tube is 4.2v for 325mA and the other tube is 3.7v for 325mA?

I don't see how you can get different voltages for two tubes when they are wired in parallel - won't they be at the same potential?

 

[rplate]

I don't see how you can get different voltages for two tubes when they are wired in parallel - won't they be at the same potential?

That's why I'm asking, I don't understand how can it be either.

Your wiring has a apreciable high DCR

This could be a problem.

 

[mondogenerator]

My suggestion would be to break the parallel links, and measure each tube to check they're vaguely close.

They cant be very close, given the difference you're seeing.

Are you sure you havent over-voltaged the filament?

During testing or accidental over zealous adjustment of the PSU?

I say this as I have done it with the smaller 1J29B/1J24B (I forget which), when changing between series and parallel heater connections

After that, the the operated almost as well as before...less gain, but, still usable.

I guess I aged the filament pretty well!

 

[mondogenerator]

Also one tube could be conducting far more than another, then this added Quiescent DC current will generate a voltage difference at that node that will try and move the parallel device filament voltage about.

Circulating current.

 

[rplate]

Apparently my crocodile clip connections made high DCR path.
Its all good now, equal voltage drop and 650mA of current. Thanks!

Any recomendation for good fillament bias resistors which are not obsolete?

 

[Rod Coleman]

Please use multiple parallel wirewound resistors for filament bias. Parallel multiples are able to reduce the effect of bad-quality end-cap mounting in ordinary wirewound resistors.

current production recommended types:

Welwyn W24 (good, but end-caps are steel)
Ohmite AG10 (non-magnetic, probably best value)
Mills MRA10 (non-inductive)

 

[andyjevans]

Filament bias resistors....
- something porcelain or vitreous enamel, like Welwyn WW series
- russian resistors like this, various values
2.2 kOhm 10W Russian Precision Wire Wound Resistors 18 pcs | eBay

 

[mondogenerator]

I'd plump for the Welwyn myself...steel end caps or not,I have found them to be both reliable and good value, in comparison to Mills or thick film Caddocks.

 

[Rod Coleman]

Yes, thick film is about the worst resistor element material for filament bias.

 

[mondogenerator]

I was talking about cost/value, I dont know why Thick film would be worse, just less economical (I'll take your word for it)

If buying Russian, please watch out you dont buy these:

PTMN-1 PERMANENT WIREWOUND RESISTOR 1 Watt 27kOhm 1% NOS Lot of 20pcs | eBay

I'm sure they have a use...just not where hifi is concerned.

 

[andyjevans]

Hmmm.. interesting. The Russian resistors I listed are the best sounding I've yet found. I prefer them to the Welwyn's, in fact.

Never liked thick film resistors for this kind of thing.

 

[Rod Coleman]

Thick Film has high excess noise. (Excess meaning beyond Johnson noise)

 

[mondogenerator]

The ones I linked are nichrome WW types, only good to about 1000Hz