Smokin' hot rectifier plate resistors on Deluxe Reverb clone

[Sansteeth]

Hello everyone,
a customer brought me an old Shade amp (French brand from the 60's / 70's) which is a straight up clone of a Deluxe Reverb except for two 220R/5W plate resistors on the GZ34 rectifier.
He was having a problem of intermittent sound, so he opened it and witness for himself what I had mentioned to him last time I worked on this amp: these waxy cardboard eyelet board get hella conductive in the right/wrong environment. But as he was chopsticking around he realized that the mains wire that was touching the rectifier plate resistor was starting to melt.
He brought it to me and sure enough, as soon as I hit the standby switch, the temperature of those resistors rises to more than 212 degrees F in about 30 seconds.
- General current draw doesn't change and stays at around .9A at 230Vac on the mains throughout the process
- No power tubes = plate resistor stay cold
- new power tubes don't solve the issue
- new rectifier doesn't solve the issue
- the voltage drop on those 220R plate resistors is at around 38Vdc at power up and steadily go down as temperature rises (by the time they are around 212 degF, the voltage drop is around 30Vdc)

If the rise in temperature was accompanied by a surge in current draw I would assume something is wrong downstream, but except for the rectifier plate resistors getting smokin'hot, there's no symptoms. what do y'all think?

[SEL49]

30V across 220Ω will dissipate 4.1W. Yes, the resistor will get hot. For a margin of safety, you should double the actual power dissipation and use the next common wattage resistor above that. IOW, use 10W resistors.
 

[AlNewman]

30V over 220R is also 136mA.  172mA before it gets hot.  That seems high, no?   .9A over a 1A fuse.... 

Are you sure there's no shorts in your power supply?  What are the voltages along the B+ rail?

[mresistor]

What is your filament voltage ?

[tubeswell]

Those resistors are eating up energy from a short circuit somewhere in the power rail

[Merlin]

- No power tubes = plate resistor stay cold
- new power tubes don't solve the issue

Kinda sounds like the power tubes are biased too hot then. Problem in the bias supply?

[Sansteeth]

- No power tubes = plate resistor stay cold
- new power tubes don't solve the issue

Kinda sounds like the power tubes are biased too hot then. Problem in the bias supply?

Bingo! The rectifying diode was not connected anymore, giving a steady 0Vdc bias voltage  hence the 0.9A current the whole amp was drawing at idle...
I don't know why I threw the towel so fast on this one haha, I guess what threw me off is that the temperature of those GZ34 plate resistors kept rising but current draw did not. I was also not thinking of "no negative bias voltage" because I was assuming in those conditions a tube would last seconds and do something real spectacular and dramatic with 440V on the plates.
It goes to show that assumptions while troubleshooting can be your worst enemy...
Thanks to everyone for chipping in!
Now back to getting this amp working as it should....

[Lectroid]

Okay, I'm a noob so I'll ask the noob question.  I've only read 40 - 50 schematics in all but I've never seen a rectifier tube with plate resistors.  Certainly they aren't on the Fender schematic for the AB763 DR.

So the resistors are between the diode plates and the incoming AC?  Then they ... limit... the incoming current, no?   And hence the resulting max B+.   But it seems like a real hairy way to go about it.  What am I missing?   Why are those resistors used at all? 

[trobbins]

Why are those resistors used at all?

Valve rectifiers have peak current limits, and the actual peaks (to stay under the limits) need a minimum series resistance which is usually provided by the power transformer windings.  But some power transformers have too little resistance, so it is not uncommon, and certainly for more vintage amps, to add extra resistors in the anode paths.

[pdf64]

Why are those resistors used at all?

Valve rectifiers have peak current limits, and the actual peaks (to stay under the limits) need a minimum series resistance which is usually provided by the power transformer windings.  But some power transformers have too little resistance, so it is not uncommon, and certainly for more vintage amps, to add extra resistors in the anode paths.

Why use 2 anode resistors in this scenario, rather than 1 cathode resistor?

[HotBluePlates]

Why use 2 anode resistors in this scenario, rather than 1 cathode resistor?

You already know that I don't know the answer to this question.

Wondering aloud:  The concern is not "DC Volts Output" but damaging the rectifier due to excessive Peak Plate Current.  (Yes, this "plate current" really comes from the cathode/filament, but...)  And the power transformer high-voltage winding is the potentially dangerous voltage that could have unrestrained current-flow.  So the "obvious" answer is place resistance (to limit current) between the fragile rectifier and the dangerous transformer-winding.

Maybe that's just a psychological trick, because adding resistance anywhere along the PT-Rectifier-Cap charging circuit is still "added resistance."
Unless there's some kind of edge-case where the resistance added to the cathode-side of the Rectifier is shunted momentarily, placing the Rectifier at risk again of excessive peak plate current.

[pdf64]

I just noticed this last sentence on Merlin's site 

The data sheet will present table or graphs showing the minimum limiting resistance needed for a given application. If the transformer alone doesn't have enough resistance to meet this requirement then you need to make up the deficit by adding resistors in series with each anode. These resistors need to have a power rating that comfortably exceeds:

P = (1.1 × Idc)^2 × R

Alternatively, you could use one resistor (with twice the power rating) in series with the cathode.

http://www.valvewizard.co.uk/bridge.html

[Merlin]

Why use 2 anode resistors in this scenario, rather than 1 cathode resistor?

I have wondered this before too. The best explanation I can think of is that two resistors also slightly reduce the peak voltage differential between the anodes, which will reduce the chance of anode-to-anode sparking inside the tube, which is a 'free bonus'. But the difference is so slight I am not really conviced this was the rationale. Maybe they just figured it would be easier for users to find two resistors with sufficient power rating, than one with twice the power rating.