Yep, that's the one... And the 100k is connected to the negative side.
BTW, it's oscillating with or without feedback connected.
Here it is:
guts
As I said in the beginning of this thread the amp was working flawlessly for over a year...only when I started changing things on the first stage the tok-tok-tok began. But of course inadvertently I could have done something stupid though not obvious.
Thanks again!!
Wires too long...
OK. Your 0V and cathode wiring is, um, less than optimal - at RF a short bit of wire is an inductor. I suspect that fitting LEDs has exposed existing problems.
Your incoming mains earth goes straight to chassis. That's good. But your 0V bus bar shouldn't go to chassis at the point, it should go to chassis at the other end, at the input valve socket. You have great big loops in your input wiring. There is no reason why the LEDs should not go directly from their cathodes to an input star earth at the solder tag on one of the screws holding the socket in place. You do not want extra bits of wire; dress those LED leads so that the very shortest path is taken. 10mm of extra wire is 10mm too much. Connect R7 and R8 in series, with as short a path as possible from the input grid to the star earth, then take a longer lead from their junction to the output of C4. Use the same technique for R9 and R1. R9 should have such a short lead going to the grid of V1 that its body almost touches the grid pin.
Another thought is to do with those output coupling electrolytics. You've glued them down to the earthed metal chassis. I hope they didn't have exposed metal cans on the face that touches the chassis?
Do the heaters have any 10nF capacitors going down to chassis (I can't see any)? Is there a DC path from the heaters to a defined voltage (perhaps +50V), or are they floating? Ideally, each valve's heater pin should have a 10nF ceramic capacitor going down to chassis by as short a path as possible (an extra 1mm is 1mm too much). Don't forget to use serrated washers between solder tags and the chassis and do them up tight so that the washer cuts into the aluminium.
After all this, it still might not work, but when you do finally find what's causing the tock-tocking, all these mods will make it sound better.
OK. Your 0V and cathode wiring is, um, less than optimal - at RF a short bit of wire is an inductor. I suspect that fitting LEDs has exposed existing problems.
Your incoming mains earth goes straight to chassis. That's good. But your 0V bus bar shouldn't go to chassis at the point, it should go to chassis at the other end, at the input valve socket. You have great big loops in your input wiring. There is no reason why the LEDs should not go directly from their cathodes to an input star earth at the solder tag on one of the screws holding the socket in place. You do not want extra bits of wire; dress those LED leads so that the very shortest path is taken. 10mm of extra wire is 10mm too much. Connect R7 and R8 in series, with as short a path as possible from the input grid to the star earth, then take a longer lead from their junction to the output of C4. Use the same technique for R9 and R1. R9 should have such a short lead going to the grid of V1 that its body almost touches the grid pin.
Another thought is to do with those output coupling electrolytics. You've glued them down to the earthed metal chassis. I hope they didn't have exposed metal cans on the face that touches the chassis?
Do the heaters have any 10nF capacitors going down to chassis (I can't see any)? Is there a DC path from the heaters to a defined voltage (perhaps +50V), or are they floating? Ideally, each valve's heater pin should have a 10nF ceramic capacitor going down to chassis by as short a path as possible (an extra 1mm is 1mm too much). Don't forget to use serrated washers between solder tags and the chassis and do them up tight so that the washer cuts into the aluminium.
After all this, it still might not work, but when you do finally find what's causing the tock-tocking, all these mods will make it sound better.
EC8010,
thanks for taking the time...I really appreciate this!
I will go through your suggestions one by one and see what happens...
I already had similar issues when I built the C3G (grounding resulting in hum), and I've seen much worse wiring in working amplifiers...😉 But then, it's my amp that oscillates.
thanks for taking the time...I really appreciate this!
I will go through your suggestions one by one and see what happens...
I already had similar issues when I built the C3G (grounding resulting in hum), and I've seen much worse wiring in working amplifiers...😉 But then, it's my amp that oscillates.
That means I will have two connections to chassis...I thought ALL grounds are connected to chassis only at one point. I know this has been discussed many times but so far I followed the "only one" route.
In case you are recommending this I would form a local star ground at the input valve socket.
It does...crammed against the socket.
No, those Rubycons are insulated (isolated 😉 ) all around.
Heaters are AC, and they are lifted to a higher potential by a voltage divider (see my schematic).
Stixx said:
No, only one connection to chassis. Alright, I'm about to describe two, but bear with me...
One connection comes directly from mains earth to the chassis. The other connection directly to chassis is from your 0V bus bar and it goes to chassis at the input valve socket. You might think that's the same as you already have, but trust me, it's not. What is meant by "one connection to chassis" is that the 0V should only contact the chassis at one point.
Glad to hear the Rubycons are insulated all round (a bit surprised, mind), and to hear that your heaters are at a defined voltage.
Got it...
In fact I ripped the entire wiring except the heaters out and started from scratch. I'm sure it will look much neater...
Done...saved me about a meter of wire.😉
I also shortened the ground bus bar considerably and connected it as recommended. My question now is: since ground of the psu and of the headphone jack are in close proximity...can they go to the bus bar at the same point or isn't that advisable? In fact, I have psu ground, headphone jack and the voltage divider in about the same place (lower left in my pic).
In case I forgot somewhere else: Thank you!
Stixx said:
No, not a good idea. Your various HT filtering capacitors will be drawing all sorts of unpleasant currents through your ground bus, so your headphone socket's ground wants to connect o the bus after those currents. The idea place for it is at the bottom of the output stage's cathode resistor/capacitor combination. And if those resistors and capacitors can have shorter wires and move closer to the star earth at the input, so much the better.
It might be time for another photograph soon.
Since this evolves into a "How to achieve a clean layout in a tube amplifier to avoid oscillation"-thread...
let's continue...
Making progress (slowly, since it is New Year's Day and I do have a family) and time for the next photagraph:
I am not finished yet, but almost done with the signal wiring. Since the image was taken I have relocated the capacitors C2 and moved them much closer to the output valves.
I also took the cathode resistor and cap combos out and am kind of undecided where to put them. For some reason I can't remember those caps are rated at 25V and large...don't know at the moment whether I should get more compact ones or reinstall the ones I have.
As for the recommended 10nF caps from the heaters to chassis...
How important are they? I've read about the issue in Morgan Jones' book but never seen it done in other amps...
And when you say "chassis"... does that mean star ground or just the chassis nearby?
Hope I can't hear people slapping their forehead 😱
let's continue...
Making progress (slowly, since it is New Year's Day and I do have a family) and time for the next photagraph:
An externally hosted image should be here but it was not working when we last tested it.
I am not finished yet, but almost done with the signal wiring. Since the image was taken I have relocated the capacitors C2 and moved them much closer to the output valves.
I also took the cathode resistor and cap combos out and am kind of undecided where to put them. For some reason I can't remember those caps are rated at 25V and large...don't know at the moment whether I should get more compact ones or reinstall the ones I have.
As for the recommended 10nF caps from the heaters to chassis...
How important are they? I've read about the issue in Morgan Jones' book but never seen it done in other amps...
And when you say "chassis"... does that mean star ground or just the chassis nearby?
Hope I can't hear people slapping their forehead 😱
Those 10nF capacitors need to go to from each heater pin to the chassis by the shortest possible route. Usually, you need a solder tag for each valve base. They reduce muck from the mains.
The arrangement of components around the first valve base is looking much better. 😉
The arrangement of components around the first valve base is looking much better. 😉
Ok, hope i am not being a smartass here, but I seem to see coaxial wire in the signal path with the schield connected to nothing. That is useless and makes things worse.
And I see you made the signal wires shielded and the feedback wires not, while the feedback wires carry the signal just as well all being it at lower levels, making them even more susceptible to HF radiation.
Please continue this post, i am learning a ton!
And I see you made the signal wires shielded and the feedback wires not, while the feedback wires carry the signal just as well all being it at lower levels, making them even more susceptible to HF radiation.
Please continue this post, i am learning a ton!
beamnet:
I think you are absolutely right.
The more I work on this amplifier the more I see how sloppy my wiring was...
After I had successfully finished my first amp (C3G, can be seen somewhere here on diyaudio) I wanted to build another one...more compact in size and also less expensive (having sunk hundreds of €'s in the C3G). But of course less expensive doesn't mean less accurate in building and wiring...and that is the lesson I'm learning here 😉
Thread to be continued...
groet,
Oliver
I think you are absolutely right.
The more I work on this amplifier the more I see how sloppy my wiring was...
After I had successfully finished my first amp (C3G, can be seen somewhere here on diyaudio) I wanted to build another one...more compact in size and also less expensive (having sunk hundreds of €'s in the C3G). But of course less expensive doesn't mean less accurate in building and wiring...and that is the lesson I'm learning here 😉
Thread to be continued...
groet,
Oliver
what wire to use
Since beamnet brought it up...
Now that I have finished the signal wiring and even crammed two pairs R7 and R8 around the input tube (I'm glad I only burnt my fingers once 😉 ) I have to do the longer leads going to the output caps. With feedback applied I will have indeed four wires...two from the input tube and two from the output tubes.
What kind of wire would be best for the job? Just regular hookup wire (I greatly prefer silicon insulated wire), or shielded wire? Could I even put the two wires of one channel together (like... microphone cable) and just run each lead where it has to go... ?
regards
Since beamnet brought it up...
Now that I have finished the signal wiring and even crammed two pairs R7 and R8 around the input tube (I'm glad I only burnt my fingers once 😉 ) I have to do the longer leads going to the output caps. With feedback applied I will have indeed four wires...two from the input tube and two from the output tubes.
What kind of wire would be best for the job? Just regular hookup wire (I greatly prefer silicon insulated wire), or shielded wire? Could I even put the two wires of one channel together (like... microphone cable) and just run each lead where it has to go... ?
regards
Use shielded wire if you whish, and connect the shield at one side to a good ground connection.
Regular hookup wire will probably do too.
Two wires running paralleled will crosstalk, which in your case is a form of unintended feedback, which may cause stability issues.
Keep some distance...
Regular hookup wire will probably do too.
Two wires running paralleled will crosstalk, which in your case is a form of unintended feedback, which may cause stability issues.
Keep some distance...
Ah, I was going to mention/query the rather odd screened cable with disconnected screen...
As beamnet has pointed out, the feedback signal ought to be screened. If you were really posh, you'd add an RF filter at the 1/4" jack to prevent RF picked up on the aerial (headphone lead) from entering the amplifier. I've never tried it myself, but A N Thiele (of Thiele & Small fame) espoused it in a JAES paper.
And you're not trying hard enough if you only burnt your fingers once. Good soldering is all about a willingness to be burnt rather than produce a poor joint.
As beamnet has pointed out, the feedback signal ought to be screened. If you were really posh, you'd add an RF filter at the 1/4" jack to prevent RF picked up on the aerial (headphone lead) from entering the amplifier. I've never tried it myself, but A N Thiele (of Thiele & Small fame) espoused it in a JAES paper.
And you're not trying hard enough if you only burnt your fingers once. Good soldering is all about a willingness to be burnt rather than produce a poor joint.
After some soldering I usually handle molten solder with my bare hand and my fingertips have all kind of white patches that smell like fried chicken.
I never use the cleaning sponge in my station, i juse my fingers
But I might be a little strange in that
just might...
I never use the cleaning sponge in my station, i juse my fingers
But I might be a little strange in that
just might...
If I was really posh I would have started entirely different and built a bigger case...now there is absolutely no space to fit an RF filter.
Well, I am trying to balance pain and perfection...😉
Thanks for the input on what wire to use...going to be implemented today.
- Status
- Not open for further replies.