Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

March 28, 2024, 05:00:22 AM

Login with username, password and session length

Recent Posts

 

PCB question

Started by ilyaa, February 26, 2015, 03:51:49 PM

Previous topic - Next topic

Roly

Quote from: ilyaaexcept these are 6L6s

Yeah, but you will notice that the anode and heater are again on adjacent pins, 2 and 3.

{The reason that this is more common with EL34/6CA7's is that designers tend to flog them with more HT voltage, but in any case a momentary open in the load under drive will produce enough kV to flash over the gap; or just peaks twice the HT supply will test any insulation weakness.}
If you say theory and practice don't agree you haven't applied enough theory.

ilyaa

hmm yaya that makes a lot of sense

well im replacing the socket and 'hum dinger' resistors =

we'll see what happens!

ilyaa

#17
k replaced the tube sockets - hum dinger resistors - got some new tubes -

things seem OK but im not happy with the bias situation and i cant quite figure it out

first of all, ive attached the actual schematic (its a different version than the first schematic i posted)

at maximum negative bias, about -57V, im still getting 80mV at the test point - too hot! says i should have 60mV there - and doesnt make sense that these 6L6s would be passing so much current with a bias voltage that low....these are new tubes so i dont think its them. opened the PI coupling caps - no difference - checked the diode in parallel with the bias test point cathode resistor - its all good. no ripple or anything weird in the bias supply - other voltages look good - (477 on the plates)

(one thing that might be funny is that the flyback diodes are measuring 1.6V on a diode meter test - they are R3000's - shouldnt i be getting a 3V forward voltage drop? or should i not trust my meter to be precise? im thinking if i had a bad shorted tube  it could have damaged the diodes (this amp had a blown fuse when i got it in) - at least enough to break down the semiconductive junction and now making it difficult for the tubes to conduct properly?)

any thoughts?

Roly

Ilyaa!  Circuit Theory old mate - those diodes are effectively in parallel with the output valves.

The most likely reason that you are getting 1.6V forward drop across these protection diodes is that to get their rating they are actually using two diode dice in series inside.  They do that sort of thing.


At -57V these 6L6's should be shut off hard, so that obviously isn't right.  But unless you are getting a lot of "crinkling" followed maybe by red-plating it's doubtful that they are actually drawing the current implied by the cathode resistor voltage.

Class    Va    Vg2    Vg1    Ia    Ig2    Zout    Pout    THD    Notes
AB1 P/P    450    400    -37.0    116 - 210    5.6 - 22.0             5,600    55.0    1.8

With 450Va and -37Vg1 bias these should be idling at 116mA for the pair (for Class-AB1, but this looks like Class-B at only 30mA each, idle).

Good thought about the possible leakage from the PI, but I think something in the actual OP stage isn't what it seems.

First up, have you checked that fusible resistors R66 is intact and actually 1 ohm?

Have you checked that the parallel diode CR3 tests as okay (with one end disconnected)?

My first guess would be that the resistor has blown open and the diode has "shorted" to about a 60mV forward drop and it's not actually telling you anything about the current.

I'd also check the screen resistors (just because they are also fusible).
If you say theory and practice don't agree you haven't applied enough theory.

ilyaa

#19
i tested R66 in circuit and its reading 1.1 ohms so 1 ohm.

i checked CR3 totally out of circuit and im getting a 0.548 forward voltage drop.

screen resistors are OK.

ill dbl check R66 and the diode -

as far as the flybacks, i was thinking if werent as they should be theyd be showing up as a high-resistance in paralell with the OPT and therefore making current draw funky....

update: dbl checked R66 and diode - both seem good. opened them and used my own current detectors (same idea with a 1 ohm on the cathodes) and still reading upwards of 45 mA per tube when bias is turned all the way down to -57V.

not sure what to try next....

okay update update:

frustrated by the situation, i thought id try some old tubes that were pulled from another amp. popped them in and the amp biased up totally right! and i got it to give me ~40-45 watts! only issue is, one of the 'old' tubes started to kind of glow a whitish bright as i gave it power - not red - but a white glow from within the tube plates. bad, right? thats probably why i had it laying around in my 'used' tube pile. anyway i tried a bunch of swap combinations and determined that if i put in either of the 'new' tubes, i couldnt bias it right. this pair of 'old' tubes did the trick but im wary of the magic glowing one.

wtf?

should i ask tubedepot to send me another pair? sounds like maybe i got some duds (this has happened before with preamp tubes, but now with pwer tubes). i got JJ 6L6GCs - any thoughts?

Roly

#20
oops ... I retract what I said about them being hard cut off (mustav been thinking Class-AB).  :loco


Checking with the 6L6-GC anode characteristic (p3);
http://www.mif.pg.gda.pl/homepages/frank/sheets/093/6/6L6GC.pdf

...while it is off the plotted values your conditions are not as far off the chart as I thought.  Extrapolating and interpolating 477Va with -57Vg should give about 30mAIa (per valve), so that looks not unreasonable.

I think the clue here might be that this amp is Class-B, meaning it idles with the OP pair at reduced current, if not entirely cut off (in contrast to Class-AB where the OP pair normally idle close to maximum allowable dissipation).

So the intended situation is 483 * 0.03 = 14.49 watts anode dissipation, well under the 30W max limit (per tube).

What you seem to actually have is 477 * .04 = 19.08W Pa.  While this is a bit hotter (2/3rds vs. 1/2 Pamax) than on the circuit, it is certainly not over the top, still well within valve spec.  The question then becomes, what happens at full drive?


I'm not happy about it, but if there is no fault around the OP stage then I'd be inclined to the view that this is due to component (valve) spec spread at the extremes of operating conditions.  I don't find this satisfactory because a 6L6-GC should be a 6L6-GC, and whichever brand you plug in you should get the same results, but I think we all share some misgivings about the quality of modern valve manufacture generally.  It's not helped by circuits that push the envelope right out to the limits.

What I would do is give it some soak runs at increasing power levels into a dummy load, keeping a close eye, and if it can deliver 30 minutes at the onset of clip without any signs of distress I'd be inclined call it fixed (with some reservations).  It should certainly heat up because it is working hard, but anything more than the very slightest red tinge on the plates is not good.


As an aside, do you still have the bottles that were in it when it started blowing fuses?  I hate mysteries so I'd be inclined to do a little research, check them for internal shorts and if clear try them back in the amp.  What I would watch carefully for is any upward creep or even runaway of the idle cathode currents.

Yeah, you have given me a bit of a head-scratch this fine morning ilyaa, but I feel sure that between us will will sort it out.  Maybe one of the Brains Trust with more experience of Hot Rod DeVille's will have some observations?


{one of the reasons I favor 6L6's over EL34's is that screen supply blowups are quite common with EL34's when the (actual) screen grid overheats and collapses.  The 6L6 is actually a beam-tetrode, not a literal pentode, and it has beam-forming plates instead of G3 (as does the 6DQ6) and this makes it almost immune to screen resistor cook-ups.}

Quote from: ilyaafrustrated by the situation, i thought id try some old tubes that were pulled from another amp. popped them in and the amp biased up totally right! and i got it to give me ~40-45 watts!

Well that is really good, gives us a very definite point - it can do it. (so we therefore don't appear to have any faults in our circuit  :tu: )


Quote from: ilyaaonly issue is, one of the 'old' tubes started to kind of glow a whitish bright as i gave it power - not red - but a white glow from within the tube plates. bad, right?

Which is of course bad.  This is normally a pale bluish fog inside the structure, but "white"?  That is the spectrum for air, but I've never seen one really white rather than an intense pale blue.  Anyway, gas in there, lots of it it seems, it's gassy and a dead loss except mounted on a bit of wood as an exhibit.

With mild gassing the getter can mop it up if the valve is run hot, and being gassy allows reverse current flows and lots of gas ion bombardment, so such a valve can end up with its electrodes as brightly lit as the heater.  It depends which way it falls, a little gas and it recovers, a lot of gas and it goes into meltdown.  Mild gassing can occur when a valve is overheated and previously occluded gas in the structure is driven out.  Gross gassing might be due to a glass failure, often a tiny crack along one of the wire lead-out seals (have to remove octal base to observe).


{ed:typo}
If you say theory and practice don't agree you haven't applied enough theory.

ilyaa

#21
i figured that much when i did the dissipation calculations myself; ~40mA a tube didnt seem that crazy as it was still within 70% dissipation.

what puzzled me was just that the bias range didnt seem right - what's the point of the bias pot if we're already at 70% dissipation at max negative bias? i guess to account for crappy valve tolerances....

i guess ill give the 'new' tubes a go at fullish power for a while and see what happens. (and i'll check the original bad guys)

QuoteAnyway, gas in there, lots of it it seems, it's gassy and a dead loss except mounted on a bit of wood as an exhibit.

so thats a for sure bad tube? i guess its not white-white, but its certainly bright, as bright as the heaters, and neither healthy electron blue nor heater orange....

and on a similar note - if neither of the original tubes shows any internal signs of a short and seems okay in the amp, are they usable? remember one of them was red-plating pretty bad. once a tube red-plates it is gone-zo? or not necessarily?

Roly

Quote from: ilyaaif neither of the original tubes shows any internal signs of a short and seems okay in the amp, are they usable? remember one of them was red-plating pretty bad. once a tube red-plates it is gone-zo? or not necessarily?

Well the seriously gassy one is out for sure - history.

How a valve survives abuse is a very variable length of string.

During pump down in manufacture the entire structure is heated using an induction heater to drive any occluded gasses out so they can be scavenged, so the structure should tolerate getting red hot.  A risk is that this will also release more occluded gas than the getter can mop up.

In operation the heating isn't uniform and the electron beam will melt a hole in the anode, then the glass, before anything else gives up.



As long as it doesn't light up the wrong colour in the wrong place, and it makes watts, it's okay by me.
If you say theory and practice don't agree you haven't applied enough theory.

ilyaa

well i popped in the red-plater and one of the 'old' tubes (the good one) and it biased up fine and gave me about 40 watts....

not a model sinewave, but they are different tubes (the 'old' tube is a 6L6-GC and the red-plater is a 5881).

man why didnt i just try a tube swap to begin with instead of this wild-goose chase??

now im just wondering if i should try and get tubedepot to send me a pair of tubes that are actually properly in spec!!!

Roly

Quote from: ilyaanot a model sinewave,

The primary life line of all of these bottles is that cathode emission.  This is what falls off as the bottle is worked, and failing accident or illness the bottle will eventually die of old age by simply not having enough cathode emission to pass any anode current worth talking about - and I mean that literally.

I will never forget a Strauss 100 watt head with a pair of KT88's, brought in because it was sounding "a bit weak".  A "bit" mind you, in a thrash band playing basketball court dances; not "a lot", just "a bit".

On the wattmeter it made exactly one watt  :o at the onset of clip, and little more driven square.

dB = 10 * log10(P1/P2)

log10(1/100) = -2 * 10 = -20dB   And he only just noticed.

<3)  my dummy load/wattmeter.

It is possible to screw a bit more life out of a weak cathode by increasing the heater voltage.  In Days of Yore when the cathode (b/w) or cathodes (colour) in a CRT TV set started to get a bit weak your could buy in-line transformers, "CRT Rejuvinator", that boosted the heater voltage from 12.6 to 13-something (standard) or 14-something-volts (deluxe).  This worked about as well as you would expect such a "fix" to work - more emission but shorter heater/cathode life, an "emergency" stopgap at best.

"Flashing" a cathode by overheating it with say 8V (for 6.3V) for a short period can also produce some temporary increase in cathode emission.

It's fun and instructive to experiment with old valves in a test rig, but it won't make you (directly) rich. 




Quote from: ilyaaman why didnt i just try a tube swap to begin with instead of this wild-goose chase??

You made a set of assumptions on the basis of the information you were given, and what you then found inside the amp.

We all do it.  It's the monkey on the service tech's back.  "Oh *s!!t*, I didn't look over there".  How could I spend 20 mins probing an amp and not realise then mains fuse was blown?

It's not so much the wins, as the bloopers and stinkers that teach us the most about the philosophy of fault-finding.

What looks like the slow way, the formal every time up the river from New Orleans by-the-numbers ... looks old hat and slow, but on examination is the quickest and most reliable way.


Many moons ago in a small electronics company I was upbraided by one of the other tech's who also did office work, about the time jobs took me, "you turn everything into a research project" he complained.

So I took about a years worth of jobsheets home for the whole workshop, and Before Visicalc, spent a couple of nights analysing.  When you accounted for rework, not getting it right the first time, I came out as the most dollar-productive tech in the workshop (the Boss was the worst, my critic near the middle; and no, I didn't fiddle it, just confirmed a loose observation).

The moral is that taking the time and trouble to get it right first time is the most profitable.  It reduces "bouncers" to a minimum which means good reputation and word-of-mouth advertising, and overall profitability.  It means that you hit the nail often enough (95-98%) that the odd ballsup can be honestly faced up to (with gritted teeth) and Free Return.

You do get the odd "chancer" who wants the Moon for sixpence, under warrantee everything, but they are rare and easily recognisable, and the vast majority of clients are reasonable, decent (if somewhat freaked), people.

They are worried about surviving without an amp (and for professionals and other situation you have a good loaner rig), and they are worried about the expense, and they are some times worried about loss of/losing "tonality".  As a music tech (or Stage Manager, or Soundie/Mixer, or stagehand/roadie) you also have to be a therapist, calming, reassuring, communicating.

Remember, some of your clients will be on drugs, some will be actually insane, some both, but all of them will be musicians.

https://www.youtube.com/watch?v=6GlO2ZDwosQ




If you say theory and practice don't agree you haven't applied enough theory.

ilyaa

good lesson in diagnostic-ing - thats for sure

anyway, tubedepot was kind enough to send me a new pair - ill pop those in when they get here and im 99% certain things will be just peachy -