Bibster's 5687 Linestage, the return

[bibster]

Hi all,

After getting some usefull feedback via this thread, I finished my linestage last sunday.

Sounds like a fantastic improvement over my old'n'crappy Yamaha AX-470 integrated sandamp, so one may say 'Bibster is happy'.
(ANYONE a schematic of this yamaha thinggy? I'd like to bypass the preamp, just use it as a poweramp)

Took schiller's advice by heart, replaced the NiCd's by 500R multiturn pot's, bypassed by (ordinary) 2200uF's: One for each (of the 4) cathodes.
Inserted a 22K/5W R for each (of the 4) plates.
(Replaced the output couplings for 4 2.2uF's as well, to improve the lower end)
Managed to adjust the pots, so that over each of the 4 plate resistors there's 220 V (10mA * 22K, right?). (First, rough biasing)
Listened (Heart the tales of mystery and imagination like I never heard in before!) and settled for a while.

Yesterday, I wanted to finetune the bias, and what? Couldn't get more tha about 150V over each of the plate resistors on ONE of the two tubes... while the other tube just settles at 80-something (311V minus 220V)
Having a B+ of 311 volts, that's 150 at the plate....

What could be the cause of this? Broken, this little tubey of mine?
(It glows less than the 'working'one as well...)

Kind regard, Paul

[kathodyne]

(It glows less than the 'working'one as well...)

hmmm, check soldering joints from and to filament....

it could be broken also....

[bibster]

Kathodyne,

If I swap the tubes, the problem swaps as well...
So the tube IS defective.. Bummer...

Anyhow, it settles at about 220V over Rplate, but a few minutes after power-on, it sort of stops working correctly (Cathode stops emitting??)
I managed to settle one of the Rplates on this tube to 220V, before it went back to 150v.
If I put the cathode to ground, 162 V is all (7.36mA)

Does this sound familiar to someone?

[bibster]

So, here I am, answering my own questions...

That one tube is indeed at the end of his life... So I orderd a pair of NOS 5687's (Phillips, a little patriotism can't do no harm ) on eBay: I'll have them by the weekend I guess, so we'll see.

Fiddled a bit around in the amp yesterday, 'cause I think it hummed way to much (50, 100 Hz). Quickly put in a switch for the heaters of the 5687's.
But turning them of did NOT reduce the hum.. pity..
I than took out (after waiting a while ) the 'defective' tube, ending up having a mono preamp: Hum seemed gone.

My (first) question is now: Can a 'at-the-end-of-life' tube induce (or pick-up or so) the 50Hz AC?

Second: I presume my PT is a bit short on the 6.3V heater's tap: I end up (SS bridge, 4700uF/0.3R/4700uF) having 5.04V on the heaters...
With quite a lot of ripple: 1Vtt. This WILL lengthen the life of my tubes, but isn't it a bit TOO low? (Or should I go for AC heating, lust to get rid of the voltage drop, thus to end up near(er) to the desired 6.3V?)

Thanks, Paul

PS: Burnedfingers: I did redo my soldering I redid everything as a matter of fact...

[kathodyne]

i have no idea...haven't had any strange behaving tubes yet....did have tubes which did not glow anymore on startup

you can only get 6,3vdc from a 6,3v winding if you use something ike a low-drop regulator...
5,04 is too low (you have to be within 5%), low heater voltage might be able to lenghthen the filament's life, but surely the tube will not live longer as the cathode will get poisoned and isn't able to emit anymore.....

6,3vac is mostly sufficient....build a decent dc filament supply if going for DC heating

greets!

[arnoldc]

all my preamps use AC for filaments, with one side grounded. Works for me, no noise from that.

[kevinkr]

Actually heating the cathodes of 5687 at much lower than the rated voltage is likely to greatly shorten the life of the tube, particularly since they run at relatively high cathode currents. For good life they need to run within their nominal rating of 6.3V +/- 5% otherwise cathode poisoning is a possibility.

It is also quite likely that your bad 5687 has low emission at ~5V, as this over 20% below the correct filament voltage and may not be bad at all. (These are not specified for emission at 5V!!)

A very ripple laden DC filament supply is far worse for hum pickup than any AC supply I have encountered by virtue of the shape of the ripple waveform which is 2X your line frequency and loaded with mostly low order harmonics, this rich spectra of harmonics is injected into the vicinity of the cathode and grid both capacitively and magnetically as well.

If you still want to use dc filament supply you will need low loss schottky rectifiers and a lot of capacitance, a high current x-over inductor of several mH will help as well in a pi-filter configuration. Provided your transformer is not too lossy you might make it to the vicinity of 6.3V.

I recommend 6.3VAC filament supply biased up to maybe 20Vdc using a resistive voltage divider off of your B+. You can use high value resistors and a smallish electrolytic cap to provide a low impedance to ground. Assuming a 450V supply and 1mA this implies about 430K/20K divider and maybe 10uF across the 20K which should be 50V for safety.

I would replace those multi-turn pots with fixed resistors once you know what the right value is for them. Fixed resistors are much more reliable long term, and I would replace those large 2200uF bypass caps with a Black Gate in the range of 220uF - 470uF, with about a 16V rating, even the less expensive grades are still a huge improvement over most anything else I have tried.

Incidentally those Philips 5687 are most likely to have been made in one of the former Sylvania ECG plants acquired by Philips in the early 1980's. (I worked for GTE Labs at the time, GTE was Sylvania's parent company. I remember our lamentations about the sale, but it did give us access to Philips products at low prices in our company store where I bought my first Philips/Magnavox CD player in 1984.) I'm not aware of this tube ever having been made anywhere else except in the USA.

[bibster]

Kevin, Kathodyne,

Thanks for these encouraging words! Tonite, I'll heat my soldering rod, and get rid of that rectumfrier and filter. As Kevin states: the rectified and (pooly) filtered DC is way worse than 'clean' AC. Thanks for reminding me
I think my PT will make it to 6.3@2A, ' cause that's how I ordered hem (her?).

I'll try to 'float' them tomorry (No parts in stock here...): Having a B+ a about 305V, I'll use (Correct me if I'm wrong) 220K+20K (Both 5W cemented thingies?), and putting a 100uF/160V cap in // with the 20K: That'll make some 25V. Is this too much? Also: Can I pick this up from any of the two 'last' RC's of my B+?
Do I need some 300R from my thus created virtual CT to each leg of the heaters suppy, or can I just tie this 'ground' to one of the heater supply's legs?

Also, I had this thought flashing through my (lack of)brain that this 'doomed' tube *might* turn out not te be that bad! (Crossed fingers... quick wife hurry up with the food, so I can get to start!)

So these ' Philips' end up being Sylvania's? Don't know enough about tubes yet to tell me if that's good or bad My ears will tell me!

Thanks, I'll keep you informed about this li'l project!

Paul

[cogsncogs]

(ANYONE a schematic of this yamaha thinggy? I'd like to bypass the preamp, just use it as a poweramp)

I don't have one, but my guess is that the tone controls are wrapped around the power amplifier, in the NFB loop as a lot of the newer (80's and up) Japanese integrated amps are. If so there isn't a tone control amp proper that can be bypassed. It can be done by removing the associated tone control circuitry and replacing it 2-resistors and a cap or two of the correct values.
I've never liked the sound of sand amps that had their tone control circuit designed as such. Very boomy-muddy sounding little critters. It's CHEAPER not BETTER! This means the power amp has a little extra gain to give the extra 20dB plus what's left over to hammer it into submission.
If the tone controls are in a 'separate' circuit, just find the 1st (input) differential pair and the associated input cap, usually a 1uF - .47uF/50V electrolytic cap, possibly an Elna LR cap, orange or 'peach' in color. This assuming it's not an IC amp! Then just find the input cap.

Hope this helps a little.

Cheers
Wayne

[Original Burnedfingers]

PS: Burnedfingers: I did redo my soldering I redid everything as a matter of fact...

I'm glad its working better for you now