Adjusting 300B filament voltage

[Paul Birkeland]

Alright - so I'm guessing that the 1085 drops a minimum voltage doing so, and hence the need for a min difference between V-in and V-out. Increase V-out too high, and the regulator starves for voltage. 

Sure, though when the regulator drops out, you'll still get decent voltage out of the regulator, but the DC regulation will become poor and the noise rejection abilities of the regulator will follow.

 

[Deke609]

Adding 18.2R in series with the 931R gets me 9.97V-out on one side and 10.02V-out on the other. PosIn is about 13.2V on both sides. Approx. 4.7 V dropped across the fil. No hum at start-up.  Going to bring the fil voltage up to approx. 4.85V or so -- just within 3% of target.  If that works and no hum after 2 hrs, I will call it good.

 

[Deke609]

Hmm. The right channel is humming - not much, but it's there intermittently. I am going to measure the difference between Pos In/Out on both boards to see if the right channel has the smallest difference - as I expect.  And then tweak Rset2 to get the same difference as the other channel.


Failing that, I will simply reduce Rset2 on the right channel to where there's no hum.

 

[Paul Birkeland]

Where you get no regulator hum today may be different than where you get no regulator hum tomorrow unless you're on a regenerator.  The big challenge is that if you lower the 8 ohm filament bias resistors, then you will both increase the voltage across the 300B filament (a good thing) and increase the filament current (not a great thing for proper circuit operation).  The extra current will drop more voltage across the active noise filter at the output of the regulator, can increase the thermal stress on the 1085 regulator, and can draw down the raw DC voltage available to the regulator in the first place.  It's a pretty delicate balance!

The BeePre is just not a particularly flexible circuit when it comes to altering it to work with a tube that isn't a 300B, and unfortunately the EML 300B doesn't meet the factor specs. 

 

[Deke609]

It's a pretty delicate balance!

Yeah, you're not kidding there! 

I've swapped the 18.2R for a 10R in series with the 931R on the offending right channel, bringing Pos-Out down to about 9.93V from 9.97V which gives me almost exactly 4.8V across the filament (so just inside the 4% target, exactly where you recommended that I start over a weeks ago!)  with a 120R added in parallel with the stock 4R.  A tiny bit of hum of intermittent hum in the first 10 seconds after power-on, but no hum since then. I'm going to let it play for about 4 hrs and then check again.

If need be, I'm even willing to run the tubes with a little less than 4.8V across the filaments. This will undoubtedly hurt tube life, but for me it's worth it. To my ears the EML 300B has more energy and detail than the JJ 300Bs, and havign heard them in the Beepre, it would be hard to go back.

Well ... regardless of what ends up working, I can rest assured that I made some serious efforts to get the amp dialed in and that whatever "solution" I end up with is pretty much the best I can do. Plus, I learned a bit about Schottky diodes and the Beepre circuit - which is great.

The Kaiju fil reg circuit will be a little more forgiving, right?  Please say "yes".

cheers and thanks again, Derek

 

[Deke609]

Just to cap this off: it's day two and no hum. I ended up with only 10R in series with the 931R Rset2 on both boards. That brought PosOut down to 9.88V on one side and 9.93V on the other, @117V AC mains. That's still about 60 mV higher than where I started before installing the IXYS diodes and tweaking the Rset2 value. With a 100R in parallel with the 4R cathode resistance on one side, and 120R in parallel on the other, I get very slightly over 4.8V dropped across each filament.


My present mains voltage of 117V AC is the lowest I've ever seen it. We are having a hot and humid summer in Toronto this year and I suspect that everyone is running their air conditioning pretty hard and causing some voltage sag. I expect that by the end of September my mains voltage will be back up to 120V AC. I'm pretty sure I've seen it go as high as 124V AC in non-summer months.  So I expect the risk of shunt reg dropout to be limited to summer ... but so far it looks like I have a winning combination even for the low voltage months.  Very pleased. These tubes sound great to me, with more goodness to come as they are still breaking in.


cheers, Derek

 

[Deke609]

Well, I've only heard the regulator drop out once since finalizing the tweaks for the EML 300B. And I expect that with cooler fall temperatures, my mains AC V will bounce back up and it won't be a problem.


But ... I like playing with stuff, so I've been thinking. What about heating each filament with 2 x adjustable LM1085 reg boards, paralleled, each set for 5V out? I.e., split the work between them and stay well under any current and heat dissipation limits? Further, if the current used by two boards after voltage doubling and rectification, when added to the 1.3A filament draw, would exceed the current rating of the existing heater winding, the boards could be powered by a new separate heater PT.


AND ... if that's workable in principle, then would it not be possible, with the right heater PT, to use EML 300B mesh tubes with 1.5A 1.4A filament draw?


I have in mind a separate enclosure to house one large-ish heater PT with dual heater secondaries, rectification and regulation (four LM1085 boards).  I figure I could put together the LM1085 boards on protoboard.


Would this be workable in principle? 

[edited to correct misstated current draw of the EML 300B Mesh filament]

cheers and thanks, Derek

 

[Paul Birkeland]

Paralleling regulators doesn't do anything helpful in this scenario.  Regulators don't split work well.

Buying an AC regenerator would definitely be the way to go.

 

[Deke609]

Thanks PB.

Paralleling regulators doesn't do anything helpful in this scenario. Regulators don't split work well.

I was thinking this might help if the reg was hitting its thermal or current limits - not as a solution for my low summertime mains voltage (which just got worse! The city repplaced the line on my street today and now I only get approx 115V! Hoping it's just everyone's fridge and ac working overtime after being off for 8 hrs and that mains V will go back up to at least 117V tonight).

But I did some reading and see your point about regs not splitting current well. Plus, it looks like the current limit of the LM1085 isn't an issue anyway - since it's rated for min. 3A output. 

          Question: If the current limit of the LM1085 doesn't limit the output current available, what does (assuming, hypothetically, that one has more than enough VA from the
                          heater winding)?

Buying an AC regenerator would definitely be the way to go.

Yeah, it's looking that way. But they are crazy expensive (PS Audio, Monster Power, etc.). Tripplites with mains regulation are a lot less expensive, but they only kick in at 111V, and just pass 112V - 127V straight through -- so they're no help.  You'd think that a mains voltage regulator would be an ideal diy audio project, but so far my search for a schematic or build guide has turned up nothing.  But I will keep looking.  Maybe the fact that it involves mains power keeps people from posting about it? For liability reasons? But i find that hard to believe since there are lots of build threads/schematics for 211 and 845 amps that operate at really high B+.  Dunno.

In the interim, I'm using my variac to bump up the mains voltage - but it's only good for 6 hrs max continuous use and needs frequent monitoring. So I will need to find a better, permanent fix.

cheers, Derek

 

[Paul Birkeland]

You could also get a 115:120 autoformer or power transformer to sit ahead of your BeePre.

Dropout occurs when input voltage and output voltage get too close together.  That's why low line voltage is an issue.  The regulator will not overheat with some other major problems going on.

 

[Deke609]

You could also get a 115:120 autoformer or power transformer to sit ahead of your BeePre.

Excellent idea! Thanks PB. And I can install a cheap analog panel mount voltmeter to monitor voltage.

Dropout occurs when input voltage and output voltage get too close together.  That's why low line voltage is an issue.  The regulator will not overheat with some other major problems going on.

OK, so I think I understand this part, b/c you've explained it to me before and I've read up a bit about regulators.  But then why/how does a draw of 1.3A cause dropout?  And why couldn't the heater circuit handle the 1.4A draw of an EML 300B Mesh? The LM1085 is good for at least 3A output - so I don't think the 1085 is the limiting factor. Just throwing out guesses here: Are the transistors current limited? Or does voltage output of the secondaries sag as the Volt-Amps demand approaches the VA rating of the secondaries? Is the power demand of 13.5V (after rectification) @1.3A just too much for the PT to handle? Or does something else explain this?


If the issue is limited VA output of the secondaries, then a relatively simple solution would be to get a separate heater transformer with higher VA rating. Does that make sense?


cheers and many thanks, Derek

 

[Paul Birkeland]

But then why/how does a draw of 1.3A cause dropout?

The voltage available from the raw DC supply drops, not necessarily only because of the 1.3A, but also because of low line voltage. 

  And why couldn't the heater circuit handle the 1.4A draw of an EML 300B Mesh?

The noise elimination circuit also has some series resistance, and drawing extra current through this will reduce regulated voltage.

If the issue is limited VA output of the secondaries, then a relatively simple solution would be to get a separate heater transformer with higher VA rating. Does that make sense?

The 6.3V secondaries have some extra current to give.  If you bought a power transformer that had two 7V/4A windings, that would bring up the raw voltage a bit as well.

 

[Deke609]

Awesome. Thanks PB. 7V/4A windings maybe hard to find. But I figure I can go with 10V/4A and use dropping resistors to dial in the voltage.


cheers and thanks, Derek

 

[Paul Birkeland]

10V/4A with a bridge instead of a doubler might also work.

 

[Deke609]

Cool. Thanks PB.

Along these lines, I might also be able to go with 12.6V/3A with full wave and skip the doubler. I can play around with PSUD to see what works.

Is there a rectification advantage to using a 4-diode bridge instead of two-diode full wave other than not needing a center tapped secondary? I.e., does a bridge produce less ripple?

For modelling the heater circuit in PSUD, do you have a recommended load resistor value? Or can I just use PSUD's default value (I forget what it is)? Guessing that the answer is: add up all the resistance in the heater circuit and plug that in - but just want to be sure.

Edit: one more question - I plan to install switches on the transformer to allow for nominal +5V, 0V, -5V changes to the mains voltage. Shorting or non-shorting? I have no idea which is best for changing between windings.

Apologies for all the questions - I'll try to stay quiet for a while after this ;D

cheers and thanks, Derek

 

[Paul Birkeland]

In PSUD, you can change the resistive load to constant current. Most of the questions you're asking can actually be pretty well answered by doing some modeling in PSUD and looking at the power transformer current, ripple, output voltage, etc. 

If you're asking about +5, 0, and -5V switches for a step-up before the BeePre, I would just leave it at +5, there isn't much penalty for having a little extra AC voltage.  Otherwise I would recommend non-shorting switches.

 

[Deke609]

Cool. Many thanks PB.  that constant current feature will sure make things easier.


cheers, Derek

 

[Deke609]

I'm going to try using an autotransformer to adjust mains volatge going to my BH amps. Since it's an autotransformer, it offers no isioation from mains power - but I figure this is ok since all BH amps have their own isolating transformers.


I'm going with the Hammond 168J - rated 2000VA and weighing 23lbs!  I plan to use a 10A fuse on the incoming mains power - so run it at a max of 1200-1250VA (depending on mains voltage variations).


And so here's my question: Am I correct that running the 168J well below its max VA rating gives me some latitude on how to wire it? Specifically, the Hammond schematic shows incoming hot wired to 115V with options to take 85, 95, 105, 110, 115, and 125V out. I want to take power from the 115 tap and use switches to connect mains incoming power to either 105, 110 or 115 for nominal +0V, +5V and +10V output options.  I figure I can  use  2 X SPDT switches (high voltage/current rated) - please see attached wiring scheme.


Does anyone see a problem with this?


many thanks, Derek

 

[Paul Birkeland]

I'm going with the Hammond 168J - rated 2000VA and weighing 23lbs!  I plan to use a 10A fuse on the incoming mains power - so run it at a max of 1200-1250VA (depending on mains voltage variations).

Is this just for the BeePre?

And so here's my question: Am I correct that running the 168J well below its max VA rating gives me some latitude on how to wire it?

You may find that the listed input/output voltages are not at all accurate with that transformer lightly loaded.

I figure I can  use  2 X SPDT switches (high voltage/current rated) - please see attached wiring scheme.

What you're doing would work, but something like this will be a lot simpler:
https://www.mouser.com/ProductDetail/CK/R10307RN02Q?qs=sGAEpiMZZMvNbjZ2WlReYgfl1vEqXQaLTI2c32B8PEc%3D


The Hammond 168B is a far better choice to use with the BeePre.  This is an instance where spending 5x the money will net you far worse performance.

 

[Deke609]

It's for all BH amps - so, at minimum the Beepre and either Kaiju or an SII. And down the road, it might be running the Beepre, Kaiju and both SII's off of it at the same time.