Gosh I love Bottlehead!!! And a question (of course).

[Paully]

After doing something really stupid shorting the filament choke because I wasn't paying attention to grounding wires (and also didn't bother to cover said grounding wires during the initial build) I just wanted to say thank you to the Queen for shipping parts out so incredibly quickly and to both Paul Joppa and Grainger for helping me trouble shoot.  You both went to a lot of trouble and it is appreciated.  The amp that was having problems has been playing for a bit now so thankfully I did not damage the power transformer.  So thank you everyone for going to so much trouble!

My NOS 1930s 45s are precious so I wanted to check my voltages.  I raised a3 up to 171 trying to get a4-a1 below the rated 275v because just dividing initial a2 by 3 left a2-a4 well above 275v.  Amp A is the first number, amp B is in parentheses.

a1 220 (224)
a2 487 (488)
a3 171 (171)
a4 218 (221)

a2-a4=269v (267v)

So a4-a2 is around 269 volts for each.  Within rated maximum of 275.  But if I want to get it a little lower for tube life, say around 260v, should I just pop a3 all the way to 175?  Of course I don't really understand if that would be running it more conservatively or not.  Or should I do something with resistors?  As always, too much fun and thanks for the help!

Paul

[Paul Joppa]

Alas, the 45 is so old that it does not have specified maximums for anything except plate to cathode voltage. Most people take the specified operating conditions as a guide to maximum dissipation and plate current, which would suggest 36mA as the maximum current. You are running 220 volts with a 6K resistor, which is a hair over 36mA. Reducing the plate voltage voltage further would increase the plate current.

The specifications are conservative, allowing for some variation in voltages and conditions - usually 10 percent.

Your best bet would be to reduce the power line voltage a bit. Second choice would be to replace the 270 ohm resistor in the power supply with a larger resistance, which will reduce the PSU voltage. But that resistor would then dissipate more power, and would have to be larger, so it could get mechanically difficult.

Replacing the two 3K cathode resistors with a larger resistance would reduce the current, allowing a higher cathode voltage without exceeding the current limit. This is another candidate approach.

[Paully]

All right, I got the parts in to fix my PS Audio P500 which allows me to control wall voltages.  I went and measured the wall voltages and was surprised they were 117v.  I checked a different socket closer to the stereo and they were 123 there.  Go figure.  Anyway, I lowered the voltage on the Paramounts to 117V and adjusted the driver voltage.  This is what I get (which is what I used to have with the old board when my PS Audio was working) when I set the driver to 151v:

a1 199
a2 458
a3 151
a4 197

Plate to Cathode= 261v and 199v/6000ohms=33ma.  Seems good to me.

[Paul Joppa]

In the middle of the night, it occurred to me that the easiest way really is to add resistance to the two 3K cathode resistors and adjust the driver plate voltage as needed. No other circuit changes, and no other soldering needed. There is a red wire that connects the top of one resistor to the bottom of the other - just replace that wire with a small resistor. A 5-watt resistor will work for any resistance up to 1200 ohms.

Keep the current at 36mA, and use this resistor to subtract voltage from the 45 cathode. I'd shoot for 250v - a 17v drop would be 470 ohms. The lower voltage will reduce both plate dissipation and internal voltage, two of the three main factors in tube lifetime. If you go lower than 250v, reduce the current as well, ideally in proportion to the voltage to the 3/2 power. This will maintain the balance of current and voltage to match the output transformer impedance. For example, 240v and 34mA cuts dissipation to 8 watts and adds 1200 ohms to the cathode resistor.

[Paully]

All right, I will raise the line voltage back up to 120v which is where things are supposed to be operating and take measurements.  I will figure out from there what the optimal addition to the cathode resistor is and play around.  I think 45s are supposed to be around 7 watts dissipation but of course that is one opinion and there are others.  I will keep reading and calculating!  Regardless, this is quite the learning experience.  Thanks for the help!

Paul

[Paul Joppa]

Cool!

I'd suggest you raise the cathode resistor to 7000 ohms (add 1000 ohms as discussed before) and run the 45 at 260v/30mA. The driver plate.45 grid should be about 157v since you'll have around 470v at the 45 plate. Keep that ratio: 157/470 =~ 1/3. Parafeed cap is 4uF, so the original 3.3 would work fine. (The higher load impedance reduces the cap size by the square of the ratio.)

[Paully]

I am assuming you saw my other post about getting the new 5K output transformers (but I get confused so I like to make sure we are on the same page so to speak).  But anyway if getting the new output transformer allows me to keep the old Auricap parafeed capacitor like you say, that is really nice.  Fantastic actually, buying the output transformers actually knocked $100 off of what I was thinking about spending on caps so it was a really, really good deal going ahead and buying the transformers.  Perhaps I will play around with the parafeed cap values later when I have paid my wife back for the money I spent on all this iron!  But again, thanks for the help!  We are going to have a great time putting the new iron in and I will pay close attention to the ratio.

[Paul Joppa]

Yes, sorry, that note "run the 45 at 260v/30mA" was optimized for conservative operation of a 45 into a 5K load, and the 7K cathode total resistance was based on your measured 467v at the plate - the current is reduced a bit so I'm guessing you will go up to 470 volts on you power line voltage. I did it in a hurry because I'm going to be out of town for a while and didn't want to leave you hanging - or risk you spending $$$ on unnecessary capacitors!

[Paully]

Just wanted to post a follow-up.  I paralleled a 1uf cap with the 3.3 to see if I could hear any difference in bass or otherwise getting it closer to the calculated 4.0.  I couldn't though it did sound worse if anything (darn hard to tell for me).  Doubt that has to do with the uf difference, more to do with the cap I chose as it was what I had on hand.  But bass seemed to be the same.  So I guess I don't have to spend a C-note on a new set of 4uf caps as the 3.3 appears to be close enough to optimum.  Maybe someone with better ears could hear something more than I could but I am satisfied.  Probably will bypass with some .22 teflons later just to see what that does to the overall sound but obviously I don't think the uf difference does much as Paul was saying.

Paul