Smokin' PT

[bluestone]

Ever had a bunch of good ideas but when put together they are not so good? The other night during practice we played and took our normal break. I put my "Blues-Tone" into stand-by and relaxed. I remarked about the quality of the sound of my amp. It was awesome. Upon returning to the stage, I flipped the stand-by switch and played a few notes, talked with our drummer and asked if he smelled anything burning. Turning to my amp, smoke was pouring out of the PT end.....

Build info:

Hoffman AB763 board, Allen TP40D PT, Allen TO26OT, All NOS pre-amp tubes, NOS Tung sol 5881 power tubes, NOS 5U4GB Rec. Tube. I beefed up the first stage filter to 110UF using 2-300 volt 220 UF caps in series with the proper resitors. I made the normal channel like an AB165 so I could use it for bass which was my reason for beefing up the PS filtration. Wired the 1st stage filters to the hot side of the rectifier so as to not stress it.
I have used differnt speaker configurations: Bassman cabinet for bass and 2-10" Kendrick Blackframes for guitar. It sounded awesome!

Dumb...Why would I beef up the power supply and then use a 5U4GB rectifier? I believe that after about a year of playing this was the reason for the PT failure.

Questions:
1) Should I reduce the filtration to 2- 350v 100 UF  in series (50UF total) and use one of my Mullard 5AR4's? I really don't want to go with a solid state rectifier...
2) David Allen suggested that I use a 1/4 amp fast blow
fuse in series with the CT to protect the PT from any HV side problems. Is there any reason that I shouldn't put that fuse holder in the unused ground switch hole?

Thanks for the help!!!

Tom

[HotBluePlates]

If you unhook the PT secondary wires to the rectifier socket, does the PT still output voltage? I just wonder about the exact cause of failure in the PT. I guess I'm asking if the PT is toast. Is the rectifier tube still okay? It would seem that too much capacitance would pop a rectifier sooner than it would smoke the PT, which then makes me wonder about some other short-circuit being present.

Is there any chance of a short-circuit or partial short-circuit that was a result of tinkering the filter caps? Maybe a stray bit of wire or solder blob?

All else equal, the 5U4 can output more current than the GZ34. 110uF seems high though, and the 50uF you are leaning towards sounds good with either rectifier.

[bigdaddy]

Talk with David, I think the problem was probably the over filtering and the 5U4GB.

The first stage should have been like no more than 40MFD. You could have used a 22mfd then a 10 watt 100R or less resistor and then the higher amount. I think you brunt a PT rather then a tube. Tube rectified amps do not need as much filtering as far as I know. And rectifier tubes have maximum ratings for the caps right after them for a reason.

I think you broke some basic amp design rules and paid the price eventually. I know, I have done it too. I racked up a bunch of 6L6GCs and other big bottle tubes until I found the right screen and dropping resistors for my SE amps. Had to cost me well over $100 worth of tubes, easily.

[bluestone]

Thanks for the replies. The PT main ins as well as the HV out are smoked.  The yellow filiment wires are ok. So the PT is gone. I think that I am on the right track with the filtering. I got the idea through one of G. Webers books about beefing up the PS. I have emailed David and he has responded. He is the one that gave me the idea about the CT fuse. Is it a good idead to put it in the unused ground switch hole?

As far as learning lessons, you are absolutely right. Yeah it sometims cost a few dollars but I love the tinkering aspect and some of the best "Mods" come from stretching the envelope. I think that I just stretched a little too far!  :shocked:

[RicharD]

I am a believer that B+ fusing should be internal.  My reasoning is that if it blows, you've got a problem that should be examined via a lamp limiter + if someone puts a 2A sloblo in there, it could really get ugly.  This is no rule that says the B+ fuse can't be external.  Orange and Hiwatt both did it that way.

[HotBluePlates]

Are you certain you didn't do something simple and silly, like installing the new filter caps backwards?

The PT main ins as well as the HV out are smoked.

Not the best terminology... Do you mean the insulation was burned off the wiring? I've done that before with a bone-headed mistake, but the PT itself was still fine. Slipped new heatshrink over the bare wires, corrected the blunder and used the same transformer.

I still feel like the rectifier would have been the item to fail if the cap value was the only issue. Peak charging currents of 1A or more are not unlikely or unknown, even with the proper filter cap values. It's a byproduct of how the input-cap type of power supply works. So although the PT cannot deliver 1A sustained without severe voltage sag, it can likely deliver it in brief intervals as seen in a normal power supply.

A good-quality short-circuit would be needed to really kill a PT. Backwards electrolytics can form a good quality short circuit.

[bluestone]

You are right, I should have said that the wires melted. The whole build has been untouched with a soldering iron for over a year and I have used the amp weekly.  So the components have been the same. The only thing that I changed this week was the baffle board and speakers. When it origanally happened, I thought that maybe a piece tore off the sheilding and fell into the chassis, but such was not the case.
 

[HotBluePlates]

I beefed up the first stage filter to 110UF using 2-300 volt 220 UF caps in series with the proper resitors.
...
The whole build has been untouched with a soldering iron for over a year and I have used the amp weekly.  So the components have been the same. The only thing that I changed this week was the baffle board and speakers.

So the 110uF filter has been in place the past year? Only the baffleboard and speakers?

I can shoot a moving target, but pop-up targets at 300 meters are significantly harder than a paper target at 25 meters.

So now we're at everything was kosher, only the speaker situation changed, and the PT becomes a ghost? There's a huge missing ingredient here somewhere...

[alerich]

I believe I got this graphic from Merlin' site - not entirely sure.

[bluestone]

I am sorry about any confusion...I do not believe that the speaker change had anything to do with the situation.  Yes, the chassis has been unchanged for a year. I was just trying to give you all a build description. I just now pulled the PT. The overheating is without a doubt centered around the CT ground wire. I did order a new PT and will put in some smaller filter caps. I will use a current limiter when I power it up. I have looked and looked but fail to see any obvious problem. I don't have a tube checker so I cannot check the rectifier tube with out putting it in the circuit. As I have several, I may just as well toss it. Sorry again HBP...

[darryl]

...I put my "Blues-Tone" into stand-by and relaxed. I remarked about the quality of the sound of my amp. It was awesome...

Why does an amp always sound its best in the minutes before a total meltdown?   

Check the output valves also, in case the problem began there. Redplating output valves may draw enough current to overload and damage power supply components.

[HotBluePlates]

Not a problem.

David Allen suggested that I use a 1/4 amp fast blow fuse in series with the CT ...

Try it and see if it holds.

Due to cap charging currents, which are very short, severe spikes of current, I don't think a 1/4A fast-blow fuse will survive normal operation (depending on how "fast" it is). But Dave may have some experiemtnal knowledge on this that I don't. Due to the current spikes (which might be near or over 1A), a fast-blow fuse rated to withstand the spikes will fail to protect against a steady partial short-circuit that could damage a transformer. A slow-blow might be the right choice. You may want to use the Duncan power supply calculator to estimate the needed fuse value.

[PRR]

Put 100 ohms 5 Watts between rectifier and first cap. Very preferably flame-proof... this resistor WILL blow someday.

Most non-guitar 5U4 Commercial designs either have skimpy cap, modest PT, or a series resistor to control 5U4 peak current. Here you have a souped-up (low resistance) PT and a very large cap. You either need to do some heavy computation, or add some resistance.

[bluestone]

Thanks for all of the tips. I am just going to go back to stock parts with 50UF first stage filtering and a good Mullard 5AR4. I have the parts on hand. I will also try fusing the CT as David suggested. I will be interested to see if the one of the power tubes went south. I may check that today in my Bassman build.....

[FYL]

Put 100 ohms 5 Watts between rectifier and first cap. Very preferably flame-proof...

I'd rather add resistors between the PT and the valve plates, so the valve would also be protected.

Finding the right value is quite easy:

Get the minimal source R from the valve datasheet. For example a GZ34 @ 400V and a capacitor-input filter with 60µF at C1 needs to see a minimum of 75R per leg.

Measure PT primary and secondary DCR and calculate the reflected R (Cf. Duncan's PSUDII for a handy calculator). Say you get 35R.

Subtract reflected R from minimal source R, this will give the minimum resistor value. Here 75 - 35 = 40R.

Get the next E12/E24 value up. 47R here.

Calculate maximum dissipation in the series R's - say the amp will draw 250 mA max @ 400V for app. 12V voltage dropped and a 3W dissipation. R's should be at least 5W and preferably 7W or 10W. Flameproof and suitably mounted of course.

[HotBluePlates]

Get the minimal source R from the valve datasheet. For example a GZ34 @ 400V and a capacitor-input filter with 60µF at C1 needs to see a minimum of 75R per leg.

This information is not always available though. At least, not in that form on the data sheet.

I promise I'll evenutally get around to posting the whole procedure. It is in RDH4 in the section on rectification. The charts presented there avoid the use of a long formula.

[FYL]

This information is not always available though. At least, not in that form on the data sheet.

You've either got data for a given set of params, or curves allowing you to do the same, or both.
Cf.  5AR4.gif below, from the GE datasheet.

I promise I'll evenutally get around to posting the whole procedure. It is in RDH4 in the section on rectification. The charts presented there avoid the use of a long formula.

The charts in RDH4 chapter 30 are just examples using a 5Y3GT or a 5T4, copied from the RCA datasheet... Page 99 gives the basic calculation procedure.

Cf. RDH4 Recto.gif below.

[HotBluePlates]

Except that peak plate current for the rectifier is determined based on a chart, which relies on a result from another chart, which is there to avoid having to do some involved math.

You wouldn't necessarily know that the peak plate current is too much, because it is based on filter cap value and the current draw of the amp, to know whether a resistor (or how much resistor) is needed.

[bluestone]

OK, New PT installed Same Allen PT.
I decreases the 1st stage filtering to 50uf and am using a Mullard 5AR4 and NOS GE 6V6GT power tubes biased @ 30ma. I installed an in line fuse (Red and Yellow wire as directed). Powered up using bulb limiter, no problems but, if I put a 1/4 amp fast blow fuse, it blows upon flipping the stand-by switch. Using a 1 amp Slo-Blo no problems and it sounds good. This amp is dead quiet.

I posted my voltages:
bluestone
Everything seems ok except V5 values are unsteady. Pin 6 is the only steady one @437. I do NOT have the reverb pan hooked up. Perhaps this is why, or is it a bad tube? I haven't substited either yet.

Questions:
1) Will the voltages on V5 be unsteady w/o a reverb pan connected?

2) Is a 1/4 amp fuse too light for this circuit?


Thanks for all of your help.

[bluestone]

Subbed V5 same thing so it's not the tube...I don't know if the voltages posted so I will list them:

Pin        V1     V2    V3    V4    V5    V6
1          265   256   434   260    ?    219
2           0       0      0       0     ?      60
3           1.9     2      0      1.9   ?     64
4
5
6         258    249    435    253 437   229
7            0      0        0       0    ?     64
8          1.9      2        0      20   ?     95


Pin          V7         V8
1              0           0
2   
3           435           435
4            435           435
5            -35           -35
6            -35            -35
7
8              0              0

Heater voltage is steady@ 6.8 and 5.3.
Wall voltage 123

[sluckey]

1) Will the voltages on V5 be unsteady w/o a reverb pan connected?

V5 is the tremolo oscillator/CF buffer for a Hoffman AB763. Kill the oscillator with the footswitch or a shorting plug in the trem footswitch jack and the voltages will settle down.

2) Is a 1/4 amp fuse too light for this circuit?

Probably, especially since it is popping when switching STBY on and off. Try a 1/4 (or 1/2) amp slow blow.

Are you sure V4-8 is 20 volts? 2 volts would be much more likely.

[bluestone]

The voltage tamed out after shorting the tremolo jack...Thanks
Yes, it was 2 volts not 20...Old age!!!
I will try the 1/4 and 1/2 amp fuses went I get them tomorrow.

Thanks for all the help!!!

[bluestone]

Got a 1/2 amp slo-blo. Played for about a half hour. No problems so I think that I am all set for now....

Many thanks to all.....

Now onto the next project!!!!!