What's Up With My Plate Voltage?

[dankinzelman]

Hi everyone,

I'm working on an FBT500R, and the plate voltages for all tubes fed from the last node (D) of the power supply is significantly lower than spec.  Any ideas of where to look for the culprit (see attached schematic, notes in pencil are values in my amp).

Also, why does the schematic show different plate voltages for channels using equivalent plate and cathode resistors? I have only a very basic understanding of tube circuits, but would have expected that with all stages using similar value resistors, I would expect similar current (and thus similar voltage drop) across the load resistor on each plate in the absence of signal...

[EDIT] Forgot to mention that my amp has two 5ohm resistors between the diodes and the first reservoir cap on the high voltage circuit. They aren't shown on the schematic, so I penciled them in.

[shooter]

since there is no original spec on the PS rails, I'm thinking drafting error, since the #'s look reasonable

[dankinzelman]

So you think they made a mistake drawing the schematic, and my readings are representative of a functioning circuit? I feel like I should have more headroom from this amp, and was hoping that resolving the power supply issue would take care of it.

[shooter]

both ch equally without that which you seek?
put recorded music in, set dials to "where the music sounds best" and post #'s
yellow highlight's, are they, parts you swapped? verified?

[dankinzelman]

I used a yellow highlighter as I traced the circuit, simply a sign to myself in case I got distracted, and my initial plan was to replace all the capacitors because it seemed easier than learning to troubleshoot. I've since decided to study the circuits, and have thus abandoned that idea, but I'm still pretty ignorant as to the finer points of how everything functions.

I should mention I'm using this amp for a Fender Rhodes Mk I, which I presume presents a much more dynamic signal than most guitar pickups. Both channels actually sound great if I don't push the volume too much, but like I say, I'd expect more headroom, ie the amp responds to dynamics a little too readily, and not in a very pleasant way. So if I get it sounding really nice at normal playing level, then I play the instrument hard or accent some notes, the distortion sound can be kind of fuzzy. This is what I imagine cutoff or blocking distortion to sound like (although I don't have the experience or knowledge to know for sure). I can make a recording of it if it would be useful.

With two brand new Jensen C12K12 speakers and 40W, I was told this thing should be thunderous. It is if I don't dig in too much, but I'd like to get a bit more headroom, and I expected fixing the plate voltage to be the first step.

I'm new at guitar amps because mostly I play saxophone, but eager to learn.

When you say 'where the music sounds best', you mean I should send a fullrange music signal out of my audio interface into the amp and set the knobs (volume and EQ) where I like it, then turn off the music and measure quiescent voltages at the plates with those settings? Or do I measure them while the amp is playing? What volume should the input signal be? I presume very low, to simulate the weak output of single coil pickups? I also have a transformer-based DI box which I can use in reverse if that would be better for driving it.

[shooter]

being a Tech, musical explanations are somewhat subjective 
so one "test" i use with non-tech is the music in, any working amp should play "clean".  that gives me a sense of the amp without having it to scope it out.
that amp uses a cathodyne PI which really hates to be pushed into distortion.


have one of your guitarist friends plug in and whale on it and give you their impressions.


If you just have to get in and tinker;
lift (unsolder) one left leg of the 20uF cap, CH1 1st tube, then evaluate.
IF that is the "right" direction;
disconnect the NFB wire (CR on schematic), evaluate

[raza]

Nice, no pops! Amazing!

well now... what is better to switch  rele or opto? All amp run with optos. but I dont now if in this point is better rele.


thanks!!!!!

[thetragichero]

if it's got a cathodyne have you tried the "paulc mod"?
just replacing components for no good reason usually results in less than ideal results
if it has a "low" input try that for your rhodes, or try turning the volume down on the rhodes

[dankinzelman]

Thanks everyone, tragichero do you have a good link to the paulc mod? I have found a lot of mentions of it, but no good explanations of what it actually is.
I am going to try removing the cathode bypass cap and let you know what I find. FWIW, I have installed a variable NFB pot and like it better with more NFB, but I didn't try it with no NFB at all
Can anyone here help me figure out a good bias point from the datasheet? It's not too easy for me to understand what I should be aiming for (most beginner biasing articles just tell you "el34s need x watts", but none of them is talking about EL503s :-p
Also, I ungrounded the HT filter caps and tested them for capacitance. Each appeared to be nearly spot on in terms of capacitance, but I don't have an ESR tester and didn't test for continuity. Is a basic multimeter capacitance test sufficient, or would I be safer replacing them? They appear to be original (ie, late 1960s or early 1970s).

[thetragichero]

https://sites.google.com/site/yourtubeamp/mods-and-maintenance/paul-c-mod-for-princetons

removing cathode bypass caps might certainly be helpful if you've got a hot input/ power tubes that don't need a ton to be driven (I've got an ecl86 build that that certainly helped)

if you can find an el503 datasheet that gives max plate dissipation (brief Google search didn't turn one up) you can plug into an online calculator to determine the necessary bias point (just choose whatever the closest available tube is in the calculator)

[shooter]

I find the simplest way to bias is putting 1 ohm resistors from cathode to ground of the PA tubes.  voltage read = tube current.
you still need to calculate;
dirty quick;
(reading from 1 ohm X plate voltage DC = tube power dissipated in Watt's) * .7 ~~ = to "set point" for bias

[shooter]

forgot this link,  real good read;


The Valve Wizard

[dankinzelman]

@thetragichero here's the EL503 datasheet. I'm guessing max is 29W? I don't know what 'design center rating system' means. And most articles suggest 60-70% of max dissipation so ~17.5W? I'd like these tubes to last a long time (they are really hard to find).
According to my tech, the tubes looked strong, and they appear to be really well matched from my poking around, although I think there is a bit of heater hum.

[thetragichero]

looks like 29w if the screen current is kept in check. I'd probably pretend like they're el34 in bias calculations (25w). and shoot for close to 70%. as you're looking to amplify a rhodes and not play metal guitar erring on the side of caution is a good idea
the punk rocker in me believes that if nothing's smoking and it sounds good, it is good

[sluckey]

http://tdsl.duncanamps.com/show.php?des=EL503

[dankinzelman]

So I just spent a couple hours tinkering. Removing cathode bypass cap certainly reduced distortion but also gain and it felt less full to me. I did find a loose pin on the PI which I fixed. I prefer the circuit with a lot of negative feedback (significantly more than the stock circuit), and it looks like turning down the volume on my Rhodes gets me the best result for now.

Anyone want to help me calculate appropriate values for the Paul C mod? It doesn't look that hard to implement. The stock values in my amp are pretty different from the ones I've seen in the articles, so I would need some help adapting it to fit.
Thanks for all your help so far!

[sluckey]

Just use the same values as shown in the 5E3 mod, 1M, 2.2M, 56K, 56K.

[Pietro]

lose the 5 ohm resistors and make direct connection to first filtercap.

[dankinzelman]

OK, I'm leaving for a few days for work, but will get back to this next week. Thanks for all your help so far!
What is the function of the 5ohm resistors before the reservoir cap, and would it be worth removing them?

Any advice on good places in Italy (or Europe) to order components?

[mresistor]

There's Tube Amp Doctor GMBh in Germany. Seems to have good products but I can't speak to the details of ordering from a European location.  Some USA companies carry their products.

[dankinzelman]

looks like 29w if the screen current is kept in check. I'd probably pretend like they're el34 in bias calculations (25w). and shoot for close to 70%. as you're looking to amplify a rhodes and not play metal guitar erring on the side of caution is a good idea
the punk rocker in me believes that if nothing's smoking and it sounds good, it is good

So you suggest 70% of 29w = 20.3w?

[pdf64]

I can highly recommend Tubetown in Germany. Typically much better prices than TAD.
https://www.tube-town.net/ttstore/
Unfortunately the new trading arrangements look likely to make it prohibitively expensive for UK customers to use EU suppliers any more

[tubeswell]

looks like 29w if the screen current is kept in check. I'd probably pretend like they're el34 in bias calculations (25w). and shoot for close to 70%. as you're looking to amplify a rhodes and not play metal guitar erring on the side of caution is a good idea
the punk rocker in me believes that if nothing's smoking and it sounds good, it is good

So you suggest 70% of 29w = 20.3w?

Those datasheets rate EL503 Pmax = 27W. Duncan's page suggests 25W.

[thetragichero]

use 70% ofthe smallest number found, which was 25w. obviously can be adjusted based on sound but that's a good starting point

[dankinzelman]

I just got my oscilloscope running and it's SO MUCH FUN! Now I'm digging around trying to get rid of some hum.

How much rectifier ripple voltage is acceptable on the B+ rail?

I'm seeing approx 2 volts 100hz sawtooth (with B+ around 400V at 17W cathode dissipation) at the first node. I'm also seeing the B+ oscillating through the entire power supply in unison with the tremolo oscillator, regardless of whether tremolo is activated, and regardless of the volume of the tremolo channel.

I haven't tried the Paul C mod yet, as soon as I get a chance to pick up the resistors and plan out the layout I will try it, but I'd like to figure out if I have power supply issues first. I tried temporarily jumpering in some new capacitors and things didn't seem to change much...

[shooter]

How much rectifier ripple voltage is acceptable on the B+ rail?

how loud you play? 


I like 2vac at 1st tap, <20mV last tap

[dankinzelman]

I'd like to be able to play pretty loud, but I also need good clean sound at lower volumes, and SNR is not so good as-is. I haven't had too much chance to fool around with measuring sources yet, but I feel like the Fender Rhodes tends to push the amp harder than a guitar (a lot more bass energy and sustain).

Should I be concerned about the tremolo oscillator working its way back up the rails to all nodes of the power supply, or is this pretty normal?

@shooter, I was able to clearly see the distortion mentioned in the Valvewizard article you cited above (the cathode 'spiking').

[shooter]

I was able to clearly see the distortion mentioned

He talks of "fixes" that are worth the effort

[tubeswell]

Should I be concerned about the tremolo oscillator working its way back up the rails to all nodes of the power supply, or is this pretty normal?

1) bad filter cap ground- Use your R-meter to Check that the filter cap on the PS node (which is supplying the LFO stage) is grounded. there should be DC continuity between the filter cap ground lead and the chassis


2) poor lead dress (check wire runs from the LFO are not close to signal grid wires) or


3) crosstalk, if the LFO stage is sharing a dual triode envelope with anything other than an LFO buffer stage or the cathodyne stage)

I was able to clearly see the distortion mentioned in the Valvewizard article you cited above (the cathode 'spiking').

Gain spike distortion? Try a 1M grid stopper directly on the cathodyne’s grid pin and sizeable grid stoppers (upto 47k each) on the output tube grid pins

[dankinzelman]

I tried a 470k grid stopper on the PI, but didn't hear much difference. Is 470k not enough, or is this useless without installing grid stoppers on the output tubes?
Intuitively, I'd have expected a resistor here to lower volume, but this did not occur. On second thought, I presume the negligible current flow means voltage doesn't really change, is this correct?

[dankinzelman]

I finally had time to do some more experimenting yesterday. A 1M grid stopper on the PI removed the gain spiking completely, but made everything sound super distorted so I took it out. Any idea about why this made everything sound so bad?

I did compare waveforms at the PI output with the speaker outputs with and without the 1M stopper. Without the stopper, gain spikes were clearly visible on the cathode, but they were not making it to the speaker jacks (Blencowe says the respective output tube is in cutoff at this point). On the plate, I couldn't see the frequency doubling mentioned in Blencowe's article, just heavy clipping. Putting the stopper in made the PI output look nicer, but the waveform at the speaker jack was severely rounded off on the bottom half.

I tried the Paul C mod next, and it seems to be what the doctor ordered - more headroom, less distortion, and a much nicer color to the distortion.

I do have two questions:
1. Can I  drive the PI from a different tap than the triode stage that precedes it (which is in the same envelope)?
2. How obsessive should I be about matching the two 56k load resistors? When I drive the amp hard, the speaker output shows a fair amount of crossover distortion. Is this is to be expected and not worth worrying about, or should I try to get a very precise match on the PI resistors to improve it?

[PRR]

Resistor match does not improve crossover.

The cathodyne is good up to a point but when you overdrive the finals it gets a bit wonky. Some players like it, some don't. Do remember that Oscilloscopes do NOT buy tickets to shows, so you do not care what they think.

As Tubeswell says, 47k series grid resistors change overdrive action and reduce abrupt crossover. The LTP may be somewhat more polite in heavy overdrive. But there is always a limit. And something 'bad' at that point. Switching to grid transformer drive changes some of our usual flaws into other flaws.