superlead with master volume redplates

[Roe]

my new 1959 sounds good and works well with one exception: two tubes redplates with 480v b+ (but not with 400v). usually the tubes that redplates are v4 and v5 but the redplating changes to v6 and v7 if I swap sides. the redplating is always on the push side (not the pull side), the side that normally is connected to the 82k in the pi.

these things dont work:
- changing tubes (v3-7)
- lowering 82k to 68k
- connecting a different preamp
- changing .022uf pi input cap
- swapping choke
- installing 10k swamp resistors and 47k on pi grid as well as upping 47pf to 120pf
- swapping plates or grid on pi
- swapping sides on the lar mar pots so that pi coupling caps go to the other end of the pots
- removing larry/aiken/vox presence
- moving presence/NFB wire

the only thing constant is that it is always the push side that redplates (if that makes sense), it need not even be the 82k side or only the two tubes connected to one side of the OT primaries

http://folk.ntnu.no/roef/IMG_3035
http://folk.ntnu.no/roef/IMG_3038
http://folk.ntnu.no/roef/IMG_3036.jpg

amp has following mods:
- two separate bias circuits, both are like the original circuit but the 2nd is squeezed into less space.
- dual larmar MV with tweaked presence circuit
- variable nfb
- changed grounding scheme (amp is quiet=
- switchable .68 on v2a
- switchable one-wire mod/straight 1959 bright channel.
but since it redplates just the same with a different preamp, I guess the PI/driver is to blame.

[Andy Le Blanc]

red-plating means that the tube is passing too much current.

lift the coupling caps from the inverter side and clip them to ground.

check the power side first for any erroneous voltage condition, and the caps, then check the inverter.

[Structo]

ROE,
Have you tried asking this at the Metro forum?

I seem to remember more than one guy there that experienced the same problem as you describe.
One guy suggested changing the corresponding PI plate load to lower the voltage to the two tubes that are redplating.

A couple other guys said that it was pretty common and you shouldn't worry about it.
I think one said his amp has always done that but his tubes last as long as any other amp.

[Roe]

red-plating means that the tube is passing too much current.

lift the coupling caps from the inverter side and clip them to ground.

check the power side first for any erroneous voltage condition, and the caps, then check the inverter.

thanks for you suggestions. the voltages seem to be perfectly stable, and the -48v bias voltage does not seem to drop under load at all.

so I should lift the two pi coupling caps so that they are connected to the bias resistors (220ks) on one side and earth on the other. I'm not sure this will work since the amp only redplates when running flat out for some minutes

[Andy Le Blanc]

yes..... leave the bias intact. lift the caps from the inverter side and clip them
temporarily to ground. Zero signal. Look for any imbalance in the power tubes voltage condition.
Plate voltage, G2, G1, any radical difference in tube dissipation between each
tube, between sides of the push pull circuit.
Anything that might explain the symptom.

then check the caps, then start exploring the inverter for imbalance

red plating alters the tube characteristic and shortens life expectancy in most cases.

Which style master volume?

[Roe]

yes..... leave the bias intact. lift the caps from the inverter side and clip them
temporarily to ground. Zero signal. Look for any imbalance in the power tubes voltage condition.
Plate voltage, G2, G1, any radical difference in tube dissipation between each
tube, between sides of the push pull circuit.
Anything that might explain the symptom.

then check the caps, then start exploring the inverter for imbalance

red plating alters the tube characteristic and shortens life expectancy in most cases.

Which style master volume?

thanks again! I'm using the so-called larmar MV, a refined version of the post phase inverter MV used by ken fischer and ken bran (in the late 70s parks). Basically the two 220k bias splitters are replaced with a dual 250k pot with 2m2 resistors connected to ground in case the wiper fails and so that the pot reads 220k. tube grids are connected to the wiper and bias voltage to one side and the pi coupling caps to the other side

ROE,
Have you tried asking this at the Metro forum?

I seem to remember more than one guy there that experienced the same problem as you describe.
One guy suggested changing the corresponding PI plate load to lower the voltage to the two tubes that are redplating.

A couple other guys said that it was pretty common and you shouldn't worry about it.
I think one said his amp has always done that but his tubes last as long as any other amp.

Structo, yes've asked at the metro forum but haven't received much response yet. the amp definitely redplates a too much, I'm afraid

lowering the 82k plate resistor to 68k doesn't seem to help at all

[Andy Le Blanc]

I don't trend towards a MV, if nothing out of the ordinary is going on post MV,
double check MV balance too, it might be compounding an imbalance post inverter.

I sure others will chime in if they have had trouble with that MV type.

[dynaman]

I second the notion of screwy PI caps. I just built an amp that would only see -10VDC on one side of the output tubes (push-pull). Swapped out the offending coupling cap and the problem disappeared.

I've used that MV extensively and never had a problem...

Um, you're not using a dual HV power transformer, are you?

[Roe]

hi, yes my PT (1203-80) delievers 475v and 395v to the plates. the 395v setting does not redplate at all but the 475v setting redplates with different tubes. of course, the 475v sounds much better than the 395v

[dynaman]

And have you installed the proper additional bias circuitry for the extra HV winding? The same bias setting won't work for both. You should have at least one extra resistor, but preferably a pot. Sounds like you set the idle current for the lower voltage windings. At the very least, you should set the tubes to run properly at the high voltage setting. That way, you'll at least be overbiased for the low voltage option.......

[Roe]

And have you installed the proper additional bias circuitry for the extra HV winding? The same bias setting won't work for both. You should have at least one extra resistor, but preferably a pot. Sounds like you set the idle current for the lower voltage windings. At the very least, you should set the tubes to run properly at the high voltage setting. That way, you'll at least be overbiased for the low voltage option.......

bias should work. I have separate adjustments for low and high voltage. right now high voltage has -48v on pin 5 whereas low voltage has a lower bias voltage (-40 or so, don't remember now).

I use a 3pdt standby switch so that I can switch between hi and low voltage and get separate bias voltages. I first adjust bias at high voltage and then have a extra pot that lowers the voltage even more but only works on the low voltage setting

[Roe]

..
Plate voltage, G2, G1, any radical difference in tube dissipation between each
tube, between sides of the push pull circuit.
Anything that might explain the symptom. ..

I measured the el34s under full load and at idle.

Idle:
plates: 480v
screens: 478v
grid: -48v

Full load:
plates: 450v at lowest
screens: 430v at lowest
grid: stable at -47-48v


voltages seem ok don't they measured at the bias pot the bias is dead stable but measured at pin 5 it varies slightly under full load (from -46.9 to -48.1)

[krash]

actually the screens look high to me, but I am no EL34 expert. However, with EL84s I common in my amps, I have had occasional redplating with modern tubes that have always been resolved by going to larger screen resistors with no detriment to tone. I use 2.2K or 2.7K screen resistors at all times on all tube types now and I have zero problems.

I haven't recorded the screen measurement but I am quite certain it is way more than 2V below the plate.

[Structo]

Those voltages don't seem bad at all.

And you said that you tried adjusting the PI plate resistors to balance it better?

Have you tried different PI tubes?

EL34 amps generally all use a 1K screen resistor.

Have you tried different EL34's?

Have you checked out the power tube sockets making sure there is adequate pin tension at the socket and pin 5 especially?

[raiken]

You may want to consider the fact that you are running the poor EL34's way over their max dissipation limits with 480V on the plates and a 3.4K load (for two tubes in a 50W, which is the same as 1.7K for four tubes in a 100W). You do gain a bit from the lower full-power sag voltage of 450V, but not much.

Note that some "real" Marshalls can typically sag up to 100V (!), so they don't tend to redplate as bad, although, in a class AB or B amp, max plate dissipation doesn't occur at full output, rather at somewhere around midpoint, so they may still be running high on the plate curves at less than full-clipped, sagged output.

The following image shows the 25W max dissipation curve in green, a 50W dissipation curve in red, and a 1.7K (850 ohm per tube) load line in yellow, all superimposed over the plate curves in blue:

[img:925:561]http://www.aikenamps.com/EL34_480V_3.4K_26mA.jpg[/img]

As you can see, if you are running on this load line, biased up at 26mA (50% of 25W at 480V), the poor EL34 is spending most of it's time way above the max dissipation curve, in fact, above the 2x max dissipation curve (the part where the yellow line is above the red curve).

In a push-pull amp running close to class B, you can get away with running up to the 2x max dissipation limit because the duty cycle is near 50%, so the average dissipation never exceeds the 25W max limit, and you also spend some time at the edges of the load line where you are below the dissipation curve. However, you can't expect to spend very much time above the 2x limit and not exceed the dissipation.

If you want to visualize what happens when the plate voltage sags, take the yellow load line and move it down vertically, keeping the same slope, but intersecting the x-axis at the new, sagged plate voltage. You can see that this will push the amp into safer territory. In addition to the load line shifting down, the blue curves will also shift down vertically, because the screen voltage is also sagging by a proportional amount. Effectively, everything shifts downward except the max dissipation curves, which remain constant.

The other problem you run into in a typical high-voltage Marshall is that you are not only running above max ratings, you also have to contend with the duty-cycle modulation caused by the long-tail pair phase inverter design and the relative matching of the phase inverter tube sections under full clip. This causes the output square wave to not be equal on top and bottom - one side is wider and the other is narrower, so the wider side dissipates more power. This is why you redplate on only one side. The bad part is that there is no bias setting, even 0mA, that will keep you out of the danger zone above the 2x dissipation curve.

What can be done to fix this? Well, if you must run a 3.4K (1.7K) output transformer, you must reduce your plate voltage to no greater than around 400-420V to stay within the ratings. You can get by with a bit more, since the edges of the load line fall under the 2x curve, which decreases the average dissipation. Alternately, if you want to run 480V, you must change your output transformer to around 4.8K - 6K (2.4K - 3K for 100W) to stay within the ratings, or make sure you have enough voltage sag to keep you out of the 2x dissipation area at all times as the signal increases. Any time you exceed these rules, you run the risk of redplating.

Randall Aiken