Valve amps are dangerous

[TimFox]

The biasing arrangement you described:
Was this approximately class-B biasing of the control grids (near cut-off), or was there an active circuit that changed the bias towards cut-off as a function of signal level?

[Audiorepair]

Check out the bias pot. VR12

The schematic is drawn a bit wonky, what a mess.  Pot VR12 is dual 200k and it's drawn near the bias section but is MASTER VOLUME.

Yes, at first I thought it was some kind bias balance arrangement, and looked all over the PCB for some kind of trimmers to adjust that aren't there.

Took me a while to discover VR12 is the Master Volume pot,  never seen this arrangement before, using a stereo pot.

[floobydust]

I thought it was an AC balance pot lol.
Master Volume is as far down the chain as you can get, for best signal/noise ratio. At zero volume, you can only hear noise from the EL84's and not all the pre-amp chain, effects etc. It might be for getting more crunch where you can overdrive everything upstream yet still play at lower levels.

Let us know how the bias setup goes, I say ~0.33V at the test point giving 16.5mA per tube is ballpark. It's heavily into Class A so they are going to get hot maybe 8W each? Marshall biasing to ~12W the max. rating and they would likely glow dull red. They really do need a cooling fan pulling that off on a PCB nowadays, the sockets and board just cook.

[Audiorepair]

The OPT is faulty on this one, have ordered another.

Bit of Googling shows OPT failure on these is alarmingly common.

[floobydust]

How did it fail? The amp has no loop feedback so an open loudspeaker jack would make a Tesla coil. The secondary has a pseudo-ground CH2 which seems silly, 70mA ground loop before D5, D6 turn on. The schematic tags the OPT a "safety critical component".

[Audiorepair]

I'm not sure how it failed, DC resistance measurements on each half of the primary are similar, so no clues there.

I took a photo of the output into a dummy load, then, because as you say there is no feedback, I just swapped the primary windings around, and the output waveform became inverted.

So I guess that proves the OPT faulty.




Oh, and some of you may notice my Rigol scope is not covered in mud, like my Fluke.
The Fluke is used all day, 5 days a week, 50 weeks per year, and is about 6 or 7 years old.
I get grubby amps and PA speakers and the like in every day, so yes, it does become grubby too.

If I thought it would improve its performance by constantly cleaning it, I might consider doing that.  As it is, it has a job to do and it does it very well.

Clearly the Rigol is newer, and has some catching up to do.

[Audiorepair]

Ah, actually, I'm not sure that proves the OPT is ok.....


I scoped the phase splitter pins 1/6 and they were fine big sine waves.  Need to go back and check tomorrow.

[trobbins]

OPT's can also be checked by applying a low mains frequency AC voltage to the secondary (perhaps start low but aim to increase to near equivalent full power output voltage) and confirm the unloaded/unconnected primary half-windings each measure the same (getting to quite high) ACV - that aims to confirm no leakage/breakdown within the primary windings, although the other important test to make is an insulation resistance test from primary to core at 1kVdc (if you are doing electrical testing then an insulation resistance meter is an important tool to have).

[vk6zgo]

How did it fail? The amp has no loop feedback so an open loudspeaker jack would make a Tesla coil. The secondary has a pseudo-ground CH2 which seems silly, 70mA ground loop before D5, D6 turn on. The schematic tags the OPT a "safety critical component".

The "Q" of some output transformers is quite high.
Many years ago, I built a Mullard "3-3" mono amplifier.(3 valve, 3 watts).

It made quite a lot of noise for three watts, used with an efficient 12" speaker.
One day, the speaker connection fell off, & I was treated to the sight & sound of zapping across the valve socket.

Going purely by eye & ear, it was similar to what you get by holding a sparkplug lead a short distance from the car engine .
It didn't damage the old Ferguson output transformer------they were made of sterner stuff in those days!

[vk6zgo]

DSL201 full schematic pg 1&2.
I see a single filter cap C32 100uF 450V and R133 470k 1/2W as a bleeder resistor... which is a couple minutes to discharge the cap to something safe. 1/2W is low cost but takes forever if there are no O/P tubes helping out. 500V electrolytics are special construction and extra cost. With high line the 450V part is getting pushed in standby but this is rock and roll.

Look at a 5U4 on power up, there's always an odd blue flash when emission starts that I assume is not good, it doesn't matter as it is the sacrificial lamb as far as cathode stripping. Directly-heated filaments I have seen advice to wait until they are heated up before pushing current through them.

That schematic is the most complicated heap of crap I've seen for a long while, & all for a miserable couple of EL84s!

The old Mullard designers would "turn over in their graves!"

Even if you ignore the various "audio effects" circuits, the remaining audio amp stuff is way more complex than necessary.
Marshall may have been a legitimate manufacturer at one time, but it seems they have succumbed to the "make something that looks like black magic" school of thought..

 The suckers will be thrilled, & go "Oooo! Shiny!"

Marshall customer:-

[CaptDon]

I think they had an active 'slide-bias' arrangement where when you played it hard the finals
shifted toward AB1 or AB2, but when there was a no signal condition it drifted toward 'B' at
what must have been as close to cutoff as you could get. In the active mode with no input
signal you could actually place your hand on the 6L6's without getting burned, that was one
of the first things I noticed while working on the amp. It had bleeders and the standby switch
lifted the center tap of the H.V. winding through a double pole 'double break' single throw
switch. You would still do the 60HZ or would it be 120HZ shuffle if you got a hand on the
Stand-By switch (Terminals were not insulated). This amp seemed to be perhaps a limited
production or one of a few Beta releases of something they were toying with. Most of the
Earth amps were exact knock-offs of Fender or Peavey, This thing was different in many ways
and did not resemble any picture of Earth amps I can find. Oh, and the bias thing, when it
slid to near cutoff or even real cutoff it was quiet as a mouse, no hum no buzz (unless you
had one of the shitty guitar coil cords so popular back then!!) Another indication of 'slide-bias'
was the blue halo glow in the 6L6's going fully away after about 10 seconds of no input signal.
Wish I still owned it!!! I have a video of me playing a Gibson Grabber Bass through it at an
outdoor concert.

 

[rsjsouza]

DSL201 full schematic pg 1&2.
(...)

That schematic is the most complicated heap of crap I've seen for a long while, & all for a miserable couple of EL84s!

The old Mullard designers would "turn over in their graves!"
(...)

Oof... Indeed. This is a brutal schematics.

Someone was tasked to draw the engineers' original designs and didn't realize that paper sizes larger than A4 existed. That or the schematics distributed to the public are mangled on purpose. 

[Audiorepair]

Well it seems I diagnosed the OPT faulty when it wasn't.

When I swapped the primary over and the asymmetrical secondary waveform also swapped over, for some reason I blamed the OPT.
In reality, the OPT is shown to just pass through the primary waveform, it is THAT that is asymmetrical.

Funny how simple logic can get so confused.


Anyway, turns out the stereo Master Volume pot was bad.  One side way lower than the other.
I had ordered one along with the OPT (which fortunately was very cheap from Marshall).

So now, with the new Master Volume Pot and the original OPT, the amp is happy again.



It doesn't actually have a bleed resistor, R133 doesn't go to ground, but to some point marked Ch2? 
I dunno, this schematic is indeed very crap and confusing.

The main PSU cap and thus Standby Switch still had 260v on it after being off for 2 days.


Bias wise, as I think Floobydust suggested, the Marshall readings are WAY too high, what were they thinking?
I adjusted to onset of crossover distortion before clipping, and it was 500mV at the test point.

I haven't measured actual currents/power as yet.

[floobydust]

I thought it wasn't the OPT but didn't know the work and readings you'd taken. It's no fun scoping the plates, I avoid that measurement it's dangerous.
You have to trigger the scope from the signal generator to keep absolute phase correct when looking at one side or the other.
Lop-sided output is usually a weak tube or high value screen resistor. The potentiometers out of china are very low quality, I have replaced many. If you can get Alps or Panasonic 9mm or 12mm that fit they are better.

500mV is 25mA/EL84 but don't know what plate voltage that is, for heat.
I say it's a squawky Marshall design having no loop feedback makes the damping factor poor and loudspeaker/cabinet/room resonances will cause peaky freq. response. So they have to run bias very high to help tame that. Even the simplest Fender/Vox use loop feedback and lower distortion output stage.

[Audiorepair]

I think I got fooled into blaming the OPT because it came in with one dead output valve and its burnt out screen grid resistor.

New valves and resistor resulted in asymmetric output, yet the phase splitter outputs were good.
(I would not be scoping the power tube Anodes either. I scoped phase splitter and OPT outputs only)


Interesting what you say about lack of feedback and bias.

Should I be biassing this much hotter than normal then?

[floobydust]

The Marshall DSL201 bias spec reads "0.65-0.7V" which IMHO is near Chernobyl level for the EL84's. You get a fatter sound of course but a constant need for new tubes, cooked sockets and repairs. You can work out the voltage/current at rest and see how close to max plate dissipation ratings it is. Your 0.5V seems better if B+ didn't come up too much with the lighter load.

[Audiorepair]

Ok, I'll measure tomorrow, and maybe set it halfway between sensible and Chernobyl or something.

I do tend to set the bias so the clean is just totally clean, and the tubes have longest life.


Thanks.

[trobbins]

I'd recommend checking the main cap bleed path - R133 goes to a gnd lift circuit which should have a 10R resistor to gnd - if that path is not working as a bleed then you have a fault that should be repaired wrt safety.  Imho this is an issue with people repairing equipment that don't have the skill to appreciate schematics and safety and getting zapped.

[floobydust]

It's 470k and around 2 minutes to discharge (if the O/P tubes don't drain B+) but it should eventually discharge the filter cap. At a 1/2W rating, not really suitable for near 500V either.
I have no idea why Marshall put that 1mA into the ground lift CH2 (10R) it seems wrong. The schematic doesn't show PE ground etc. What a dog's breakfast.

[trobbins]

CH2 on page 1 appears to be associated with a 'metal washer' safety critical marking, and CH could mean a chassis connection, as CH1 and CH3 are also identified for circuit sections that make sense for local chassis grounding.  Not being the service agent for Marshall with access to training and other documentation, and just relying on a schematic, is a risk to any unauthorised repair work.  Companies do just prepare schematics for in-house use, so being a critical outside observer is a bit rich.  It's lucky that schematics have made it in to the public domain, as that is not the case for some manufacturers (eg. TEK and Picotest are two instruments I have that I would love to get schematics of) and a schematic may only become available via the effort of a diyer doing some reverse-engineering.

[vk6zgo]

CH2 on page 1 appears to be associated with a 'metal washer' safety critical marking, and CH could mean a chassis connection, as CH1 and CH3 are also identified for circuit sections that make sense for local chassis grounding.  Not being the service agent for Marshall with access to training and other documentation, and just relying on a schematic, is a risk to any unauthorised repair work.  Companies do just prepare schematics for in-house use, so being a critical outside observer is a bit rich.  It's lucky that schematics have made it in to the public domain, as that is not the case for some manufacturers (eg. TEK and Picotest are two instruments I have that I would love to get schematics of) and a schematic may only become available via the effort of a diyer doing some reverse-engineering.

It's not rocket science---it's just a valve type audio amplifier, the like of which were manufactured in their 100s of thousands back in the day.(mostly vastly superior ones, too!)

Comparing that to Tektronics equipment (all the older analog stuff not only provided schematics, but detailed repair manuals) is like comparing a duck poo to a diamond!

[Audiorepair]

Amp returned to customer.  Customer happy.


After the myriad of other stuff to sort, turned out 470k bleed was open circuit, replaced with 1 Watt.

Heatshrinked all the Standby Switch connectors, so you would really have to try hard to get zapped now.


I didn't feel the need to modify the bias further.

Clean is clean.

[coppercone2]

They need to identify hazards with stickers, its the best way, because eventually you will run into some hidden electrolytic behind a wiring harness or some shit like that.

You need to dig through it real slow like an archeologist and use nonconductive hooks to look past movable obstacles. Lifting a panel to have a board mated with something else without being able to see the other side is downright frustrating though. Ideally you would use a boroscope but no one is gonna do that. You can always bath it so long its not potted.

The worst I have seen I think is in a hughes amp, if the bleed resistor was not there, it would be very dangerous, since there was a cable harness infront of a hidden screw stud capacitor, if I recall correctly. So if you reach near the cables to maybe yank on one, attach a sensor (magnetic field), or reseat, you could brush your fingers on a hidden electrolytic that is fairly substantial. However, they did have a socket crimp leg power bleed resistor over the cap that drained it in a hurry.. I would have however appreciated a screw on shield or clip on shield between the cap and the cable harness, to increase the safety factor. Ever notice that with tek scopes too, the early 90's one, they have cable harnesses just running about possibly making direct contact with the chassis too? I managed to nick one when I was closing a box. God forbid they put some clamps for that kind of stuff.


Yeah this is how I felt with that capacitor, at two minutes, complete with the overkill breaker switch.



I just say this, hey dumbass, the raptors capacitors belong in a pen.


But, I know why it happens. Its not the electrical designers fault because you can easily lose your lungs and your job screaming at desperate mechanical integration engineers that make life so unpleasant.

[floobydust]

This is about putting profit over quality. DSL-201 did retail at $600 "Made in England" and here we have a R133  470k 1/2W sitting at ~450VDC... pretty much the max. ratings.
I think Jim Marshall would've preferred hiring designers that can do Ohm's Law and not cater to the executive trying to save $0.10 towards his new Jag lol.
Modern Vox amps no better, the PCB arcs and burns up because the designers bungled the HV spacings 

For SMPS primary caps, I did make an LED+resistor in a old ballpoint pen for checking HV and slow discharging. Something to make it easy to add this as a habit in your workflow.
It needs an update to using a PTC or larger 5W bleeder resistor, or maybe an SCR would be best.

[duckduck]

My question here would be why is the power switch between the DC smoothing caps and the tubes? If you switch the AC mains you don't get this problem as the tubes themselves will bleed it down to at the very most a few tens of volts, probably down to only a few volts if you give it enough time.

Exactly!
Thousands of tube radios & amplifiers were constructed over many years without a separate HT switch.
Such things were normally reserved for transmitters.

I believe the Standby switch was invented so that you could mute a  noisy amp when you weren't playing it, and quickly unmute it when you wanted to.

It otherwise serves little purpose.


Edit:
In fact, I have read on several occasions it is a popular myth that  the Standby Switch is there to prevent "cathode stripping" on warmup, but the reality is that only very high voltage transmitter tubes suffer from this.

Is this a thing?

I think this guy covers all of the pros and cons of standby switches on vacuum tube guitar amplifiers. The explanation that resonates most with me is "Fender was a cheap-ass and Marshall was a copy-cat".

https://debontamps.com/standby-switches/