Minor debugging needed for 1987 Plexi

[mort]

I've sorted a few preliminary bugs and have only minor issues left. I've posted a photo of my notes.

Items to discuss:

Bias and dissipation check, how to.

Presence control hums considerably as it's turned up.

Little random pops and sizzles in the sound at higher volumes. (edit: actually sounds more like minor fluttering)

The overall sound is good, bold, and plenty of gain. Controls all work normally.

It's pretty much this circuit with 1K 1W grid resistors and I didn't put those 22uF caps across the secondary of the PT(should I add those??). http://www.drtube.com/schematics/marshall/1987u.gif

[g1]

For the bias dissipation, plate voltage times plate current equals plate dissipation in watts. The measured wattage divided by the rated maximum wattage for the tube type equals the % dissipation. (ex. a modern 6L6 is a 30W rated tube, if your measured dissipation is 21 watts, then you are at 70% dissipation, 21/30).
Now the question is how do you measure the plate current? The most common method is to install precision (1% 1W is common) 1 ohm resistors in series with the cathode of each power tube. The 1 ohm value makes it very easy, you measure the voltage across it with the amp idling, and due to ohms law (I=E/R), the current is the same number, but in amps, not volts. So, for example, if you measure 40mV (.04V), the formula I=E/R tells us there is 40mA (.04A).
The resistor is installed in place of the wire between the cathode and ground. Although screen current is also flowing through the cathode, it is small enough to be ignored, so basically measuring the cathode current tells us what the plate current is.
The main reason it is usually measured at the cathode (rather than measuring plate current directly) is for safety. The cathode is at low voltage, the plate is at high voltage, so it is safer to measure the current at the low voltage spot.

[mort]

I'm using a 22k bias pot and if I max it out I can get as much -42v [between the 15k and 47k]. I'm about to install the 1 ohm resistors. If I find I need more negative voltage can I swap in a 50k bias pot to get more swing?

[g1]

Yes.

[loudthud]

One big problem with that schematic is that the bias supply is connected on the wrong side of the standby switch. When you switch the standby to play mode, it takes several seconds for the bias supply to charge up and the tube will draw very big currents until it does. This could blow a fuse or damage the tubes.

Those are not 22uF caps across the power transformer secondary, they are .22uF. They should probably be special safety rated caps, better to leave them off.

There are no grid stoppers on that schematic, that might cause spurious oscillations that would be layout dependant.

[mort]

ok I have 5% 1 ohm 1W resistors on hand, not 1%. So I put them in and I'm reading 70 mA on v4 and 56 mA on V3. At 460v each plate voltage, that's working out to 25 and 32 watts.

Could the difference in measurement be the variance in resistors? My meter read them each as 1 ohm, but does not show decimal places at that low of resistance. At the socket the voltages were almost identical from tube to tube... 460/461 plate and -41.2 & -41.4v bias at the socket.

[mort]

One big problem with that schematic is that the bias supply is connected on the wrong side of the standby switch. When you switch the standby to play mode, it takes several seconds for the bias supply to charge up and the tube will draw very big currents until it does. This could blow a fuse or damage the tubes.

Those are not 22uF caps across the power transformer secondary, they are .22uF. They should probably be special safety rated caps, better to leave them off.

There are no grid stoppers on that schematic, that might cause spurious oscillations that would be layout dependant.

I put the standby after the rectifier in the actual build. Forgot to mention that. And I also put two diodes in series on each HT leg instead of just 1.

[mort]

A photo of the layout...

[loudthud]

I put the standby after the rectifier in the actual build. Forgot to mention that. And I also put two diodes in series on each HT leg instead of just 1.

That will work just fine. The bias voltage needs to be increased (more negative). An EL34 is only rated at 25W so the bias voltage needs to be more negative. A 50K pot should help.

Normally you would expect the resistors to be within their 5% tolerance so the voltage might be off by 5%. You could swap the tubes from side to side to compare the readings, see if the higher reading follows the tube or stays on one side.

[mort]

Yeah looks like it's following the tube. 56mv and 67mv I'm reading now. So what does that mean I should do, replace the tubes? they're basically new [and supposed to be matched] but I've had them for a while so I can't return them. I guess I could try another meter but this one is pretty decent one made by Fieldpiece.

Anyway, I've got a 50k pot and I'll install it tomorrow. For now it's miller time

[big_teee]

You may need to lower the 220k resistor before the diode, on the bias voltage.
You can vary the bias voltage with that resistor.
T

[Chuck H]

Second that. Modern EL34's seem happier with 150k bias feed resistors and 1k screen resistors. The extra loading is a bummer but it's not that bad. Don't sweat a modest mismatch too much, it can actually be a good thing. The drive from your PI won't be as balanced as all that either. Swap the power tubes from side to side and leave them where it sounds best. Do add some grid stoppers and screen grid resistors.

[mort]

Do add some grid stoppers and screen grid resistors.

I had taken these two terms to mean the same thing. Are they not?

In my build I added 1K 1W resistors in series with pin 4 on each power tube. Is there somewhere else I need to add resistors?

[Chuck H]

I had taken these two terms to mean the same thing. Are they not?

In my build I added 1K 1W resistors in series with pin 4 on each power tube. Is there somewhere else I need to add resistors?

Well there's the screen grid at pin 4 (you seem to have that covered with the 1k/1W) and then there's the control grid at pin 5. It's a good idea to have a resistor here. Usually between 1.5k and 10k at 1/2W. Solder them directly to the pin 5 of each tube with virtually no lead between the resistor body and the pin. People say that higher values can have an audible affect. I haven't noticed anything detrimental up to 10k.

[mort]

Ok I have some 6.8k resistors I can put in there. Is this just to help smooth changes in current flow? or does it act kinda like a fuse? What's the idea behind them?

I had understood that the resistors on pin 4 were to keep grid voltage lower than plate voltage for tube protection.