Valve amps are dangerous

[SeanB]

Just remember one thing about high voltage, and high value bleeder resistors, is that they can often quietly go open circuit, so never trust them to discharge the power rail capacitors. I keep a few 10W 68k ceramic resistors around to discharge those capacitors, you see the small spark, but it will not damage the capacitor. Just check the resistor value every so often and see if it is drifting up, and replace with another one from the box.

Standby was normally to allow you to set up and get the amplifier ready, with hot heaters, so you could turn it on and be ready to go in under a second, without having any noise out of the speakers. Plus it also kept thermal shock down in the larger tubes, as they often also would reduce filament voltage in standby, so the filaments were warm, but not hot to fully emit, and thus did not suffer any failures from turn on surges. Warm to hot is a lot less stress, and a lower surge on a fragile filament.

[ogden]

It's not a design mistake since there is no need to discharge capacitors and constantly waste power. And it's not any different from many SMPS where capacitors remain charged for a long time too.

Be careful with blanket statements   Safety standards requires X-capacitor bleeder resistors. For instance IEC 62368-1 even requires *two* redundant bleeders.

[rsjsouza]

Just remember one thing about high voltage, and high value bleeder resistors, is that they can often quietly go open circuit, so never trust them to discharge the power rail capacitors. I keep a few 10W 68k ceramic resistors around to discharge those capacitors, you see the small spark, but it will not damage the capacitor. Just check the resistor value every so often and see if it is drifting up, and replace with another one from the box.

I adopt the same philosophy as well, although I am a bit more impatient and use a 2k2/10W - the U1273A ZLow 2k resistor is too low power for hundreds of volts, otherwise I would use it instead

[Gyro]

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.

There are plenty of ways of muting a valve amp without resorting to switching the DC HT - a muting switch on the input would take most noise out, as would shorting the output (something you can do on a valve amp of course, it's open circuit that they hate). I think it was a brainless design decision, probably done on the fly when a musician flagged it on an early model.

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

Is this a thing?

Cathode stripping is caused by pulling so many electrons out of the cathode that it loses its space charge and is exposed to ion bombardment. It's certainly a thing in high voltage tubes where full HT is applied while the heater is still warming up. It's debatable (and frequently debated!) whether this is a factor down as low as 450V - at a few kV it certainly is, but this is a 'danger zone' between where oxide coated cathodes are still useable and where they transition to thorated or pure tungsten cathodes (big industrial or transmitting station tubes, X-ray tubes, Kenotrons etc. working at 10s of kV.

At the same time, suddenly hitting a hot amplifier with (over)full HT via the standby switch might hit the tube cathodes with a momentary over-current  situation depending on the cathode / grid biassing arrangements, charging of coupling capacitors etc. That might count as stripping, caused by the standby switch (and poor design).

What is a thing is Cathode poisoning, where the tubes sit for extended periods with heaters running and no bias on them. This creates an insulating cathode interface layer under/over(?) the oxide coating that reduces its emission. That's the one I would be more concerned about with amplifiers accidentally sitting in standby for extended periods. [Edit: this used to be a real problem in so called 'quick start' valve TVs until they learned and started reducing the heater voltages too ('slightly slower start'?)]

[Audiorepair]

Interesting stuff.

It is possible to mute a typical push/pull valve amp by disabling the Screen grid voltage to the output tubes.


I am perplexed, after once seeing it done in an amp I can't remember now, why it was never widely adopted instead of the Standby Switch, which as you describe, is a pretty brutal approach.

[floobydust]

[...] Safety standards requires X-capacitor bleeder resistors. [...]

Anyone inside a power supply is a "skilled person", not the "user/operator". I've seen bleeders required only to prevent a user electric shock occuring at the power cord, as in an SMPS primary capacitor backfeeding through the bridge-rectifier.

The Marshall at B+ 500VDC is too much for a single electrolytic, you'd use two in series with balance resistors which are the bleeders as well. Did they do this?

Cathode-stripping is more of a happening with solid-state rectifiers, a 5U4 takes long enough to warm up and moderates it. I've seen stripping happen, you can see some pretty strange flashes or glows from tubes if you watch them on power up. Just watch a 5U4 someday. The Marshall can have 500VDC up long before anyone else is warmed up so I would leave Standby on until it's warmed up.

Cathode poisoning I don't think was ever as problem (outside of nixies). Old TV/radio station transmitters had the spare tubes in sockets with filaments lit, to shorten down time.

I did have many white LED's fail on HVDC like 220k+LED on 350VDC. They would go dim, or just fail. Drove me nuts and adding a reverse-diode seemed to fix it. Strange but the dropping resistor's parasitics must be a problem. Anyone else run into this?

[TimFox]

Actually, 5U4s (with directly-heated cathode, i.e. filament) warm up very quickly.  The 5AR4 and other tubes with indirectly-heated cathodes were preferred in audio equipment to delay application of  V_bb while the rest of the cathodes warmed up.  Did the Marshall unit in question actually have a single high-voltage electrolytic after the rectifier?  I believe the poster referred to measuring 550 V across a 450 V rated single part.  That is a good example of non-conservative design practice.

[Audiorepair]

Actually, 5U4s (with directly-heated cathode, i.e. filament) warm up very quickly.  The 5AR4 and other tubes with indirectly-heated cathodes were preferred in audio equipment to delay application of  V_bb while the rest of the cathodes warmed up.  Did the Marshall unit in question actually have a single high-voltage electrolytic after the rectifier?  I believe the poster referred to measuring 550 V across a 450 V rated single part.  That is a good example of non-conservative design practice.

Tim Fox, the OP claimed no such thing. You just made that up.

[TimFox]

Floobydust raised the 5U4 question.
Capt Don raised the 550 V on a 450 V capacitor question. 
My reference to "the poster" was sloppy--I did not mean "the original poster".

[trobbins]

The Marshall at B+ 500VDC is too much for a single electrolytic, you'd use two in series with balance resistors which are the bleeders as well. Did they do this?

500Vdc is not too much for a single electrolytic - it is at the spec limit for a 450Vdc rated cap during a power on event.  Without the amp's schematic and some actual measurements it is just conjecture about what voltage the first filter capacitor could be sitting at long term with the standby on.  The DSL201 schematic on Ampwares does not show the HT1 power supply circuitry and only shows 450V rated parts after the standby switch.  There are 500Vdc rated e-caps made for the guitar amp market.

There is also some doubt about whether the DSL201 did not have a bleed resistor, as the DSL401 schematic on Ampwares shows a bleed resistor is in circuit.

[TimFox]

In Reply #15 above, Capt Don stated that he had measured over 550 V on the 450 V capacitor during standby.

[Audiorepair]

Floobydust raised the 5U4 question.
Capt Don raised the 550 V on a 450 V capacitor question. 
My reference to "the poster" was sloppy--I did not mean "the original poster".

Ok, I take that back.

Sorry.

[TimFox]

Another issue with older vacuum-tube equipment (in the US) is that the transformers were originally specified for a 110 V primary voltage, but line voltages have increased since then and the present values are 120 V nominal, 125 V or so (depending on utility) maximum, for consumer service.  Thus, the rectifier outputs may be 13% higher than in the original design.  That doesn't affect the personal safety question (against shocks), but does affect capacitor voltage rating questions.

[trobbins]

The full schematic is on drtube and in a thread over at ssguitar
https://www.ssguitar.com/index.php?topic=3357.15

There is a bleed resistor across the first filter cap before the standby switch.

[floobydust]

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.

[Teledog]

Have an old Fender Bassman with an AC polarity switch.
After one too many shocks with bare feet touching ground, replaced the original non-polarized plug with a 3 prong grounded plug (N/C to the gnd pin)
Put a brass cover over the polarity switch -to keep the chassis in original shape, of course- without someone inadvertently flipping the polarity.

[Audiorepair]

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.

Check out the bias pot. VR12

What is it doing?

[NiHaoMike]

Perhaps mod it to switch off the HV at the transformer allowing the cap to discharge in standby?

[floobydust]

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.
SCREEN_A,B,C,D seems to be an AC shield, a capacitor to ground and the leg perhaps to the pot's housing, another to the front trim etc. It has metal shield to SCREEN_D, no taps.

Bias is set by PR1 as quiescent current across the common 10R cathode resistor R33 675mV at (Test CON5 pin 1 or 3 to pin 2 GND) giving 33.75mA (by my math) each, including screen current. Seems a little hot.

The Marshall manual seems to have a math typo for the bias specs for the DSL201? This is what the sheet says:
DSL410 (Two pair EL84, single 10R cathode resistor R33, so divided by four)
1.3V=32.5mA/EL84
1.4V=35mA/EL84

DSL201 (One pair EL84, single 10R cathode resistor R33, so divided by two... but not)
0.65V=16.25mA/EL84
0.7V=17.5mA/EL84

I think that's wrong, out by a factor of two? Like this:

DSL201
0.325V=16.25mA/EL84
0.35V =17.5mA/EL84

[vk6zgo]

[...] Safety standards requires X-capacitor bleeder resistors. [...]

Anyone inside a power supply is a "skilled person", not the "user/operator". I've seen bleeders required only to prevent a user electric shock occuring at the power cord, as in an SMPS primary capacitor backfeeding through the bridge-rectifier.

The Marshall at B+ 500VDC is too much for a single electrolytic, you'd use two in series with balance resistors which are the bleeders as well. Did they do this?

Back in the day, in Oz, 600v rated electrolytics were quite common--apparently not so much, now!

Cathode-stripping is more of a happening with solid-state rectifiers, a 5U4 takes long enough to warm up and moderates it. I've seen stripping happen, you can see some pretty strange flashes or glows from tubes if you watch them on power up. Just watch a 5U4 someday. The Marshall can have 500VDC up long before anyone else is warmed up so I would leave Standby on until it's warmed up.

Cathode poisoning I don't think was ever as problem (outside of nixies). Old TV/radio station transmitters had the spare tubes in sockets with filaments lit, to shorten down time.

I have worked at a lot of TV/Radio sites, & never seen that actually done.
From memory, one Tx had extra sockets, but they never had anything in them.

Thinking about it, any time saving is pretty much illusory, as the whole transmitter has to be powered down, the spare tubes shifted to a used socket, then the full heat up time has to happen as per the built in timer, before the application of HT.

Add to that necessary cooldown time before you can safely handle the tubes (they get bloody hot!), & you save very little time.

A tube change in a TV transmitter usually requires a check of the swept response of the Tx after such a change, stretching the time necessary out even further.

Most places have a standby transmitter, or if running parallel transmitters, can go to one Tx only, so there is not that much urgency.

I did have many white LED's fail on HVDC like 220k+LED on 350VDC. They would go dim, or just fail. Drove me nuts and adding a reverse-diode seemed to fix it. Strange but the dropping resistor's parasitics must be a problem. Anyone else run into this?

[trobbins]

Back in the day, in Oz, 600v rated eletrolytics were quite common--apparently not so much, now!

That was likely to be '600V surge' or 600V VP (with 500V VW) on chassis mounted can caps from Ducon.

[vk6zgo]

Back in the day, in Oz, 600v rated electrolytics were quite common--apparently not so much, now!

That was likely to be '600V surge' or 600V VP (with 500V VW) on chassis mounted can caps from Ducon.

And UCC!

They made "axial" ones in that voltage rating, too!
The 500v "0ver run" of the HT line  referred to in this thread would have been comfortably handled by such devices.

For higher VW, large rectangular oil-filled caps were often chosen.
These, in most applications would operate flawlessly for decades.

[m3vuv]

reminds me of working on a homebrew 1kw rf amp a few years ago,is had a 200uf cap bank with 2.5kv on it, the rf cage had an interlock microswitch that bled off the hv thru a 50 ohm welwyn 10 watt wirewound resistor,i forgot to defeat the switch once the resistor exploded in less than a second when i turned the amp on,had to change my underwear shortly afterwards,lesson learned!!

[m3vuv]

one hand in the pocket or one foot in the grave!!!!

[CaptDon]

I had an old Earth guitar amplifier years ago. As an indication of how solidly built
it was, the amplifier head survived a bar fire (Not the cabinet however). I built
a new cabinet for it and cleaned up the chassis and it played beautifully. I did
finally sell it. Here is the interesting thing, it had two pairs of 6L6's in push pull
and when the input was absent or very very low they would bias the outputs
to near cut off!!! Screen voltage was derived from plate voltage and the screens
were also run in a sort of ultra-linear fashion. I think it had one of the best tube-amp
sounds I ever heard and with 4 X 6L6 it got really loud if you needed it to.