Need Help with First Tube Build - WE91 300B Parafeed Derivative

[Paul Joppa]

The cathode resistor will stabilize the bias, reducing it in rough proportion to the voltage change. In the old days it was called "automatic bias" or "self bias" for that reason. Nothing needs to be changed for less than +/- 30% voltage changes.

A way to visualize this (that works for me - YMMV!): If the plate voltage drops, the current will drop, reducing the bias voltage which increases the current.

It's a form of "negative feedback", effective at DC but not audio because of the bypass capacitor. You could call it "passive servo bias" as well. Don't you love it when something simple has several different names to confuse the unwary?

[Paul Birkeland]

You can draw the load line for the cathode resistor on a set of 300B curves knowing that you have a 1K resistor, you can go to the 100V grid bias line and put a dot at 100mA, then another dot at 50V grid bias and 50mA, then connect the two with a line.  Your amp will always operate on that line, so you can find the spot that's a little under 400V (you'll lose some voltage across your plate choke) and see what current should be drawn through the amp. 

The bucking transformer is indeed a good idea and a 6.3V 1A transformer would do that job nicely per amp, or a shared 6.3V/2A unit for both.

[EricS]

PJ: Thanks for the explanation.  I've read a little bit about self bias circuits and was wondering if this change would sort itself out. 

PB: Thanks also for the explanation using the load line graphs.  This is an area of understanding tube amps that I haven't invested the necessary time to understand, so I appreciate the primer you've provided.

Seems like I have a whole new round of experimentation ahead with a bucking transformer and re-tweaking my filament supplies.  I think the best place to start might be with my variac to see what happens when the mains voltage is dropped by 6 or 7 volts. 

Thought I was done with tweaking for a while 

[EricS]

I've been poking around looking for a suitable filament transformer for bucking use.

Found a Hammond P-T166M6 (link is here https://www.hammfg.com/electronics/transformers/power/166) that is rated at 117V primary, 6.3v center tapped secondary rated at 3A, with an 18.9VA overall rating.  At first glance, this seems appropriate to me (6.3v 3A secondary), but the overall rating is only 18.9VA which seems small to me.  Does the transformer itself need a rating that is closer to 200VA if it's going to power two amps that each draw close to 100w from the AC mains?

Or, should I choose a 12v or 12.6v secondary and only use half of the secondary winding instead of using both as I would with a 6.3v center tapped secondary?

Thanks for the guidance!
Eric

[Paul Joppa]

The transformer is not handling the full power; only the 6v winding carries any load current. So it does not nee to have a large VA rating. A 3A secondary can handle 3A of load current.

[Deke609]

Further to PJ's post above, Rod Elliott has a good article on the theory and wiring of buck/boost transformers: https://www.sound-au.com/articles/buck-xfmr.htm#s30

cheers, Derek

[EricS]

Excellent!  Thanks to both of you!  Derek, I found that article last night shortly after I posted my question to the forum.   Should have known this is a topic Rod would have an article for...

I made a new round of measurements last night using my variac to reduce AC mains by 5-7v and see what happened with the rest of the circuit.  I think I'll need to wait until I have a transformer in hand to see how much actual voltage drop I get, then see if I need any new resistors to trim heater voltages.  I see the Hammond transformers are also specified with 117v primaries, so I it's not clear to me exactly what kind of voltage drop I'll get from a 6.3v secondary.  One step at a time...

Eric

[EricS]

I picked up two Hammond transformers, one is a 10vCT and the other is 12.8vCT.  My logic is that which ever one I didn't use would have utility in some other project.  The 10v one is better suited to my target voltage drop and as I expected, using half of the 10v winding gives a bit more than 5.0v drop. 

I wired both of them up and connected 200w worth of incandescent light bulbs (same draw as both amps) to measure the resulting voltages.  The 5.0v winding provides 6.8v worth of voltage drop with a 200w load on it.   My AC mains tend to range from 122v to 125v and the corresponding range after voltage drop is 115.2v to 118.2 which seems just about perfect of the primary of the PGP8.1 transformer.  I can get rid of the dropping resistors on the rectifier filament, but still need to trim the 300B filament by about one volt and take a pinch off of the input tube filament.   Need to wait for more resistors to arrive now... 

Almost there 

[EricS]

Finally had some time to return to this project after spending 5 weeks cleaning up the mess in my basement left behind by a washing machine that overflowed for 45 mins on the floor above    I've been through a few gallons of bleach lately killing the smell of mold and mildew, but it's almost back together again...  Now I have a $5 water alarm on the washer.

Anyhow, I finally wired up the the 10Vct transformer for my amps.  Got one amp adjusted last night.  I was able to completely remove the filament voltage dropping resistors from the rectifier.  I still need dropping resistors for the 300B and the 6SJ7 filaments, just smaller ones (which means they run a little cooler as well).  Here are some of the final voltage measurements at this point:

AC Mains:  124.6v
Mains after bucking transformer: 117.8vAC
Rectifier Filament:   5.04vAC (target 5.0v)
300B Filament:        5.12vDC (target 5.1v)
6SJ7 Filament:        6.4vAC    (target 6.3v)
B+ to ground before plate choke:   418v
B+ to ground at 300B plate:           398v
Voltage at wiper of humpot:            66.5vDC
300B plate voltage:           332v
300B plate dissipation:       22w

This looks better to me.  Haven't let it run for an extended time yet to measure temp reduction at the power transformer.  Will likely get there this weekend after I tweak the other amp.  I'm wondering if this might help tame some of the sub-400Hz noise I was seeing on the FFT of the speaker output - maybe the PGP was being pushed too close to the edge with 125v mains?

My plan is to build a nice box to house the bucking transformer.  I've also found a nice vintage meter with a 100v-130v scale and an old analog hours meter that I'll build into the face of the box.  A switch will energize two "lower voltage" outlets for the tube amps as well as a few more "normal voltage" outlets for my turn table, phono stage, and preamp.   

[Paul Joppa]

Thanks for the update! Those voltages look just about perfect.

And sorry to hear about the laundry disaster; what a mess that makes. Glad you were able to recover.

[EricS]

One amplifier has been modified to operate with the bucking transformer, the other has not.  Here are the final PGP8.1 temps after about 4 hours of use today:

With bucking transformer:       69c
Without bucking transformer:  72c

I just realized I neglected to measure ambient temp, oops.   Somehow I was expecting a larger temp difference.   I understand with solid state transistors, lowering the operating temp by 10c will generally lead to double the service life of the device.  Is there a similar rule-of-thumb for transformer longevity?

[EricS]

I have now modified both amps to run on something closer to 115-117vAC.  Temps are down a bit, the PGP8.1 power transformer runs a bit cooler at the top than it does at the bottom of the laminate stack.  I also noticed that the temp is hotter on the side the faces the rectifer tube by a few degrees.  After 4 hours of runtime, temps are closer to 65-66c (non-rectifier side), exhibiting a 44-45c rise over ambient. 

I ended up building a new box to power my analog rig: it has a single switch that controls two outlets on the back.  The first is standard household AC mains pass through, the second one uses the bucking transformer for a 115-1117vAC supply to the 300B mono amps.  This outlet has a separate 2A slow blow fuse to prevent overloading the Hammond filament transformer.  Inside the box are also my outboard power supplies for my turn table and my turn table preamp, both connected by an umbilical cord. 

I made a nice front panel to match the amps and installed a vintage voltmeter and a vintage hours meter.  I liked it very much until I plugged it and found that the mechanical hours meter makes a bit more whirling noise than I want.  I'm thinking I'll look for a more modern replacement that will run more quietly.

[Paul Birkeland]

If you find a quiet hours meter, I would be very interested!

[Paul Joppa]

That looks really nice - I like it!

[Tubejack]

.... not cheap ....

https://www.rammeter.com/panel-meters-instrumentation/elapsed-time-hour-meters?display=842

https://www.mcmaster.com/elapsed-time-meters/display-type~digital/