• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

5687 Linestage behaving odd...

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,

My first post in this forum, which I've been following for quite a while now: What a superb source of info this is! And amazing how some guru's'round here manage to actually help others, without even being able to see or touch or measure the stuff, because they're not living on the same continent or so!

For far for the good thing, now for the bad...

I've decided some time ago to build my very first tube amp. It being my first, I've settled for a linestage, as my integrated (ss) amp has some problems with the source selector (some don't work anymore, rattle, etc..)
I've searched around a bit, and found this linestage.

Looked like something for me, and fun to, Tube rectified: nice :)

Found a local iron winder in Toulouse, ordered the tranny and the self, found some tubes (I've replaced the rectifier by a 5UG4), bought some 'off the shelf' C's and R's, some wood etc. etc. etc.
(Ow yeah: I've chosen to use 2 22K in // for the plate, instead of 7k)
Note that I've used both 'sides' of each tube (and not both tubes!!!) in //... the plates and cathodes are both on the same B+ (The plates share their R, the cathodes are on the same batteries, one pair of NiCd's per channel: The + on the cath, the - on the Gnd)

Yesterday I finished the top of the chassis (0.5mm Aluminium, glued UNDER 6 mm. MDF) drilled the holes, put the iron and sockets in, and started laying out, soldering, some more soldering, food, soldering, and at about 17h00 I decided that 'it was done'...

Fired up the beast, no explosions whatsoever, so I'start measureing:
Cathode: +2.7V
B+: 291v (hmm... bit to high)
Voltagedrop over Plate R (11K): 210V for channel 1, 184V for the second
(that's a plate at 81V and 107V) a bit far apart, not ?
(Tubes are not matched, I *think* they're not even the same brand.....:( )

okay, let's go for it: Plugged my airport Express in, (Not audio gear in my garage... this is a verrrry handy thing for such situation, this Airtunes thinggy!), and on the out, I plugged some 'line level accepting wireless headphones'... (It's for testing, not for enjoying!!!)

Hey! It worked! I'heard some music, wow!

But then: It sounds *real* odd.... I mean: Very odd, rather like faint or so.... but then again very colse for some intruments...

itunes said: Shuffle, so it shuffled, and I ended up hearing the album 'check your head' by 'the beastie boys', and I was stunned, like really stunned: The 'music' was again far away, but: And now we're getting somewhere, the SAMPLES and the DJ's scratching, the recorded voices etc, WERE VERY AUDIBLE, like 5 times as loud or so...
Knowing that the album was recorded on LoFi 8 tracks at home, I thought: To save tracks, I'd put this kind of 'additions' to the music on a single (thus mono) track... hmmm.....

So when the 2 signals (L+R) are EXACLTY THE SAME (As in: mono signal played over L+R) the amp works 'okay', but when they're NOT, the channel kind a like cancel each other... :confused:

Has ANYONE ever experienced such behaviour? And where shoud I start searching the cause? I just can't think of anything for the moment. I've tried:
useing 1 'side' of a tube,
just one tube
invert the phase of ONE input channel
but it didn't work out....

Maybe there's some feedback through the PS?

Anyhow, if someone's got sone hint or tracks to investigate: I'd be rather thankfull!!!!!

Merci, thank you, danke, bedankt, gracias, tige dank,

Paul
 
pictures...

Hi,

Here the (verrry hi-res) pictures:
the linestage, and
the close up of one 5687.

Hope someone can help me now...

Just put both channels on the SAME signal (// cable) and than: No sound AT ALL.
So my IN phase thinggy in the 1st post should read 180 deg. OUT OF PHASE or so...

And: When I turn the power off, near the end, when there's allmost no power left, the 'effect' disappears...
Removing the B+ from one of the tubes, makes the sound okay as well: But then again I hear it in BOTH channels.....
Some feedback from output (out of phase!) to input, via the PS??

Cheers, Paul
 
Hi Paul, before you buy new tubes, try the following:

Seperate the 2 cathodes. Use 2 220-470 ohm potis for each,1/2watt is enough, and 2 2200uF caps. Try to establish 10 mA per tube. It will not be easy, because of the common plate resistor. Every time you adjust the one, you will need to
re-adjust the other too. After some iterations it must be work - if they are not that different. Use multi-turn pot's e.g. the small blue trimmers von burns (about 1,5 euro each).

This works perfectly for me, but i have a choke load, which doesn't affect much the plate voltage, as i change the current.

If you don't manage to balance them, seperate the plates too, use 22k per tube and 2 output caps, which are tied together at the output jack. Again 10mA per tube, this time it will be very easy to achive.

The typical 5687 will end up with 80v at the plate and about 3v at the cathode for 10 mA .
Don't get crazy if the cathodes are not at the same potential, every value between 2-4 volt is acceptable.

From this setup you can expect about (Uk*0,7)*15 Vrms output, where Uk ist the smaller cathode voltage of the 2.

If, for examble, yor measure 2v at the cathode, expect 20vrms at the onset of klipping and a very clean output under 10vrms. This is more than enough for every amp.

It will sound "big" and rich, not the last word in refinement (at least with the tung-sol and raytheon tubes that i use), but very pleasant und much better than every sub 1500 $ commercial preamp that i have heard - and i have heard many.

Greetings
Konstantinos
 
SOLVED!!!

Hi,

Well... heated my soldering rod, got my DMM out...
Fiddled a lot everywhere, measured again a lot of things... still nothing!

Got fed up: there was a lot of hum as well (sounded Like grounding problem)
Hmmm... made a new ground to chassis thinggy, measured resistance between some grounds, and then:
Input connector's ground to ground: a-lot-of-ohms.. Other input: Dito, outputs: Same

Odd.... well soldered (For my (lack off) skill :) ), so what can it be?

Well, these grounds are in fact rings around the connector, and they just don't make good contact! :smash: took one of these red plasticy things out between the nuts, and: BINGO!

Did all four of them, and we're having some good sounding sound here!
(Saw, router and glue tomorrow to finish the rest of the chassis)

Bedtime now, so I'll listen tomorrow!
(First idea: a lack of bass, but detailed)

(Happy) Paul
 
Administrator
Joined 2004
Paid Member
I noticed several of the resistors in your project are severely overheating and need to be replaced with ones rated at higher wattage.

Also I think you need to get some standoffs and mount your components to those instead. Flywiring as you have done is ok for prototyping but won't result in good long term reliability, and can also be quite hazardous if something shorts to your chassis or to each other..
 
This is not even to be considered at all... :)

I'm prototyping this thing, and for now it seems to work, so I'll start making it work *properly*!
Putting components in the right spot, and connecting them properly is one of the main things in that.
I'll replace the 22K resistors, as they're like burned by now.... I've put some 6k8 before, and ended up having 88.3 and 92.3 volt on the plates... that's allready a lot better.
I'll try schiller's hint: These R's and c's are supposed to be between the NiCd's and the cathode, right?
There's too much hum from the PS, so I'll fiddle with that a bit, and once it works *properly*, I'll just redo the whole thing !

Thanks,
Paul
 
Administrator
Joined 2004
Paid Member
You use either the nicads or the resistor capacitor combination, not both.

I was a big advocate of bias battery in the cathode circuit, but have come to the conclusion that it doesn't always sound even as good as conventional cathode bias with a cap and resistor. To some extent this seems to depend on the quality of the battery involved and whether or not the battery voltage allows you to achieve the desired operating point.

In my experience fixed bias applied to the grid often ends up sounding better, but requires a coupling cap between the grid and whatever is driving it. (This seems particularly true in stages required to deliver significant signal current such as the output stage in an otl headphone amplifier.)
 
Well,

They're new, these NiCd's, samsung, and deliver 2.78V.
I don't know if this is 'enough' to get me to the desired operating point, simply because I don't know what this desired, or optimal, operating point is for a typical 5687.

The soundstage is, at he moment, very flat: My wife said it lacks 'depth' ('Ca manque du profondeur'). As I've said before, I suspect a lack of bass, so, (thank you Andy for your hint) I'll try:

1st: 1 or 1.5 uF output caps,
2nd: R+C bias on the cathode.

I could then try to use the 0.47uF exit form the output on the input, and make a fixed bias for the grid (Could you give me a hint how to achive this, Kevinkr?)

Won't be doing that this weekend, but I'll try to measure the linestage a bit with my Edirol UA-25 + rightmark software (Any other hints for measuring? osX or WinXP)

Thanks!

Paul
 
hummmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm

Hi,

I've been fiddling with this thing for a while now, and I suppose my house is now filled with Bees.... Or at least: that's what is sounds like!!

I've taken out the NiCd's, replaced them with a 500Ohm multitour pot. // 2200uF Electrolytic. (one for each pair of cathodes, one per tube thus) as Schiller suggested.
Replaced the plate resistors by one 22K/5W per tube.
Adjusted the pot. to achive 220V over the 22k (10mA) per tube. (Something like 105 V on the plates)
Is this the right way to go?

Anyhow, as I said: humming all over the place this thing: Measured with RMAA, I've found 50Hz (Europe!) at -60dB, 100Hz at -64dB AND SO ON untill the end of the spectrum.

Hmm.. I think: Some pb. in the PS: Electrolytic broken or so thus not filtering very well.. (How can I figure this out? Haven't got access to a scope...)

But then, when I turn the amp off, by means of a SPST switch in the primary circuit of the tranny, this hum STAYS....

When the PS is completely discharged, and I keep watching the spectrometer, the hum is still present...

But when I unplug the powercord... you get it.

TEMPORARILY diconnecting the Earth from the mains doesn't affect the problem... Even though I've found 0.3v between N and EARTH in the outlet....

Any hints, as to help me finding the source?

Kind regrads, Paul
 
Okay,

Implemented some new (at least for me... dummy :smash: ) grounding scheme, and got the hum some 20dB's down.. Good
Now for the bad (Or again: odd) Inputs are shielded wires (Shield unused, not connected) to a 6x2 switch, then 2 inches of unshielded to the Alps bluepot, then again shielded (Shield still unused, not connected) to the 100R grid resistors.
Now: When I get my finger NEAR the LAST bit of wire (Pot to grid resistor) I get some heavy hum, like when you touch an unconnected input.
When I get near the other bits, no hum at all.....
And: When the pot in at NILL or at FULL, the effect in zero. in the the effect is largest...

How come? (Izzit the low input impedance of the grid or so???) and, far more interesting: What to do to eliminate this (so far) last bit of hum in my linestage?
(I think it picks up the 50Hz+harmonics VIA this bit)


Goodnight, Paul
 
kathodyne said:
something to do with miller.....
or: don't put your finger inside the lady....:rolleyes:
100Ohm/15 (at about) is indeed quite low, thus eager to pick up hum (??)
I see... I'll try some higher grid resistors...

The thing is however (Refering to your 'putting fingers in things') is that the 'last bit of wire' picks up ANY hum :mad:
(Tige dank foar dyn hint, however!)

Paul
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.