Bias surge with Hiwatt DC coupled PI

[Gaz]

I have a Hiwatt build with the older style DC coupled PI as seen here:

I've noticed that when I switch off standby the bias current shoots up until the preamp filter caps charge up. The bias voltage has it's own winding on the PT, and is applied to the tubes during standby. It's solid until I measure on the coupling cap side of the bias feed resistors, then it surges when B+ is applied. I don't notice this behavior with a "normal" cathode bias LTP, so figured it has to do with the DC coupled or fixed bias arrangement. Can some explain what's going on here? Can it be fixed? Does it need to be?

Thanks!

[Helmholtz]

I've noticed that when I switch off standby the bias current shoots up until the preamp filter caps charge up.

Do you mean power tube idle current?

Could you post the second part of the schematic showing power stage and supply?

[nickb]

"Shoots up". By how much and for how long?

A worst case scenario would if the supply to the PI were to instantaneously reach it's final value and the PI tube was not conducting. Then the power tube grids will rise to zero volts for about 14mS as the coupling caps charge through the plate resistors, followed by a a slower charge as the grid resistor comes into play to cause the grids to approach the final bias voltage. The whole event would be over in about 100mS or so.

If the PI supply rises more slowly, then the event will take longer but the current surge will be less. Unless the current is very high for long time it's of little consequence.

[Gaz]

Do you mean power tube idle current?

Could you post the second part of the schematic showing power stage and supply?

Yes, my mistake, idle current surges when flipping standby. Here's the rest of the power amp:

[Gaz]

"Shoots up". By how much and for how long?

A worst case scenario would if the supply to the PI were to instantaneously reach it's final value and the PI tube was not conducting. Then the power tube grids will rise to zero volts for about 14mS as the coupling caps charge through the plate resistors, followed by a a slower charge as the grid resistor comes into play to cause the grids to approach the final bias voltage. The whole event would be over in about 100mS or so.

If the PI supply rises more slowly, then the event will take longer but the current surge will be less. Unless the current is very high for long time it's of little consequence.

Thanks, Nick. This sounds like what is happening, but it's hard to tell how long it's taking. I've tried to measure the bias voltage with the "Max" setting on my Fluke, and it went to 1V (i'm guessing close enough to 0V considering), I can see cathode current shootup to about 500mvdc, but it's hard to read on the meter. I'm just wondering why I do not see this in my Marshall or Fender based 100W amps when I switch off standby. The cathode current just goes to set bias without the jump. I thought maybe it had to do with the PI...

[mhuss]

This is normal and (as others have noted) it shouldn't be a problem. The PI bias supply tube is connected to the last PS cap to be charged, and always takes tens of mSec to stabilize.

Nice schematics.

[Helmholtz]

I don't think the effect is caused by the DC coupling of the LTPI. I rather suspect it is caused by the standby wiring which suddenly connects a relatively large and fully charged capacitance to the power tubes. And of course filter time constants matter.

BTW, the DC coupling of the LTPI doesn't mean fixed biasing as the cathode voltage follows the grid voltage. So it's a mixture between fixed and cathode biasing.

[Gaz]

This is normal and (as others have noted) it shouldn't be a problem. The PI bias supply tube is connected to the last PS cap to be charged, and always takes tens of mSec to stabilize.

Nice schematics.

Yes, I considered it may be a little more noticeable because of the way the B+ string is arranged - that makes sense. And don't think I don't know posting the word "Hiwatt" on this forum your Bat Signal

[Gaz]

I don't think the effect is caused by the DC coupling of the LTPI. I rather suspect it is caused by the standby wiring which suddenly connects a relatively large and fully charged capacitance to the power tubes. And of course filter time constants matter.

BTW, the DC coupling of the LTPI doesn't mean fixed biasing as the cathode voltage follows the grid voltage. So it's a mixture between fixed and cathode biasing.

One of my swings in the dark before posting was actually moving the standby to the AC side of the rectifier, but it didn't seem to make any noticeable difference. I tried a 220K/2W across the standby as well, but no real difference. Still a surge in idle current when flipping the switch.

Noted about the PI, I need to hit the books again because I've built it not really understanding the operation completely...

[Helmholtz]

What happens if you pull the PI tube, do you still get the effect?

[Gaz]

What happens if you pull the PI tube, do you still get the effect?

No, no change with PI pulled. Something weird is going on - my hunch is this is not normal, but I can't find miswiring (yet). I would say it take a second or more for the bias to stabilize. I've never see the bias voltage jump like this switching a standby switch. The cathode current momentarily shoots up to a volt or so. On other 100W amps it may go up 150mA for a few milliseconds, but this is more extreme.

[nickb]

Gaz, current is measured in amps, not volts. So what is meant by a cathode current of one volt above?

An EL34 with 460V on the plat eand 450V on the screen grid would have a cathode current of about 600mA ( very roughly extrapolating from the datasheet) for a grid one voltage of zero i.e 1200mA for a pair, 2.4A for all four. So, a transent high current is to be expected. Using a meter to measure transients is next to useless as you need something that samples faster than the event e.g. a scope. I realy don't think you can say there is real problem here or not without good data.

So, worst case where the voltage is applied to the PI suddenly you get this for one EL34. a max of 550mA, lasts 100ms.




For this amp the HT3 supply take somethng like 100mS to rise so you'd get this, amax fo 450mA and lasts 200mS



if you were to increase the filter cap on the HT3 suppy from 50uf to 500uF to ge a 1 second rise you'd get this.


I'm aappriximating the HT voltage rise using a linear ramp when it really should be an exponetial rise (typical RC charge) to this isn't 100% accurate, but you get the idea.

Or....

Just move the standby switch from the HT supply and rewire it into the EL34 cathodes. Then there will be no surge at all.

[Helmholtz]

No, no change with PI pulled.

Does that mean you get the surge current also with the PI pulled?
You actually should, if the surge is caused by the charging of the coupling caps, as the charging currents only need the PI plate resistors.


Another explanation of the surge effect could be a short instabilty/HF oscillation.

[Gaz]

I wanted to follow up on this because I just figured out the issue after deciding to live with it. Turns out the poly coupling caps to the PI were accidentally 1uF, not .1uF. So what I was observing was the increased charge time. Doh.

[Jazz P Bass]

Cool!