DC Filament Supply questions

[jonasnoble]

Just to update everybody on my original idea, I simplified it, put 3 ohms in series with the 7806 and got it to deliver... kind of. It was super hot ~90 deg celcius with an IR gun when I measured it. and eventually cutoff.

I tested the transformer, unloaded and loaded with the tube filament.

each filament winding alone - unloaded 6.5V
loaded 6.37V
parallel - unloaded 6.5V
loaded 6.48V
series - unloaded 12.7V
loaded 12.4V

I am sure the transformer is doing it's job, and I just didn't understand how my regulator circuit was supposed to work. I bought some 5A zeners and 10W dropping resistors, and can probably make this work. But now I've moved on to the new ideas you guys have posted.

 

[RoadrunnerOZ]

haha, that's almost the same as what happened to me. I was shopping for spindle motors and saw a $10 ad on Google. I figured it was some introductory offer to get me to sign up. I created an AliExpress account and went to buy the spindle kit, and right before I hit the "complete purchase" button, I noticed they had switched it out for just the bracket.

They don’t switch it they just take you to the item that is that price - but the low price part of what’s available is the thing that grabs you, the price is not for the item pictured but a selection of one of the items available, one of which will be the pictured item with its own price - if you look at the list of photos as you click on them the price will change to the price of the pictured item. I’ve been using AliExpress for some time - you’ve just got to check as you go and not go blindly to checkout thinking you got a great deal. It’s just marketing, not false advertising just tricky manipulation.

 

[Khron]

I tested the transformer, unloaded and loaded with the tube filament.

each filament winding alone - unloaded 6.5V
loaded 6.37V
parallel - unloaded 6.5V
loaded 6.48V

Well, with those numbers (single or parallel), you could probably just get away with a low-value resistor in series with that heater.

Math says 0.5 ohms to drop 0.3V; 0.47 ohms is the nearest standard value. That should only burn off like 0.2 watts, so 1/2 watt resistor should do fine.

 

[jonasnoble]

If I was running AC, why would I need to step it down at all? 6.3V is perfect. But I'm running DC, and cannot get enough voltage out of a 6.3V secondary to feed a regulator circuit. That's why I stacked the secondaries for 12.6V.

Maybe I'm not understanding something, but no big deal. I'm about to use a buck converter to do the job.

 

[Khron]

6.37 is not quite the same as 6.3, at least not in my book. And that's with the mains being whatever they happened to be at the time - what happens at +10%?

Just figured it's a bit safer in the long run to shoot for 6.1-6.2v, but... you do you

 

[scott2000]

each filament winding alone - unloaded 6.5V

doesn't make sense... usually safe to assume 10% higher or more unloaded

 

[Khron]

doesn't make sense... usually safe to assume 10% higher or more unloaded

I tested the transformer, unloaded and loaded with the tube filament.

Good point, the actual test load sounds a bit "hazy" to me...

 

[jonasnoble]

Good point, the actual test load sounds a bit "hazy" to me...

I had my Fluke connected to the secondary windings and the tube. I watched the VAC as I connected the ground leg of the tube, turning on the heater. I watched it heat up, monitoring the voltage. Then I disconnected the ground to double check the unloaded voltage. Maybe I make a video tonight?

 

[Khron]

Not an invalid test, but borderline risky (for the health of the tube).

How did you do the test for the series'd up secondaries, though? Surely you didn't connect the tube's heater winding across the 12-13v...?

 

[ccaudle]

parallel - unloaded 6.5V
loaded 6.48V

Explain again why you aren't just running the heater on AC? Plenty of classic amps did that.

I'm running DC, and cannot get enough voltage out of a 6.3V secondary to feed a regulator circuit

6.48V loaded is what you measured, assuming nominal line voltage (did you measure primary voltage?) then good practice is to design for operation at -10%, which would be 5.8V. That is 8.25V peak, using a full wave rectifier configuration instead of bridge you could get just over 7.5V filtered DC.

I am looking at the TI web page in another tab, and on the LDO selection page I am down to "only" 11 options with output current capability from 1A to 4A, a maximum Vin above 10V (in case you want to use it elsewhere with higher voltage), and dropout voltage below 900mV.

To give a little more margin I changed to 1.5A max output current, that goes down to just TPS7A45 and TL1963A as likely candidates.
TPS7A45 has less than 0.45V drop out voltage at 1.5A across the full temperature range, so would be fine in this application.

The packages are a little more difficult to deal with than a TO220, you might want to buy a pre-made board with the regulator and required minimum capacitors mounted already.
At +10% line voltage you might have up to 9.5V DC at the regulator input, so 3.2V across the regulator. That would be just under 2W dissipated in the LDO, so you would want it mounted on a PCB bigger than you might think is needed to give some extra heatsink capability.

 

[Gus]

This should be simple
600mA and if you add the 12a_7 +300mA (6.3V wiring) about .9 amps
So I would guess the regulator was not on a proper size heatsink and/or not properly mounted with a 3 ohm power resistor in series before the 3 term reg.

Did you look up the NS books or app notes?

Do you need DC heaters?

 

[SoundBlaster]

Simple! And pretty cheap. You'll still need to create the bulk DC supply to feed the regulator. These work at better than 90% efficiency so there's no thermal issue. https://www.amazon.com/Weewooday-Re...ds=5v+voltage+regulator&qid=1695824565&sr=8-3

 

[jonasnoble]

Hey everybody, I don't always just do things for utility. I want to learn more, so I try things just to have fun. I have built ~1500 guitar amps with ac filament supplies, but I've never done a DC filament supply (that worked).

I certainly don't "need" DC, I just want to prototype it to see what I can do.

 

[jonasnoble]

Not an invalid test, but borderline risky (for the health of the tube).

How did you do the test for the series'd up secondaries, though? Surely you didn't connect the tube's heater winding across the 12-13v...?

No, thanks for catching that, as it wouldn't make sense. The best I could do with that setup was (10) 250ohm sandblock resistors in parallel. Yes it looked silly. This gave me 25ohms instead of ~10ohms. Not perfect, but at least it showed something.

 

[user 137017]

keep it simple

 

[jonasnoble]

Buck converters from Amazon worked great. I got both zener and regulator topologies to work, but both had thermal issues. This buck converter module stays cool all night, and had enough juice to power my second tube when it's time.

It took me a long time to get there, but I understand why I need an 8-9v secondary to really run a regulator circuit like I was attempting. And I could certainly keep this AC, but now I know 3 ways to get a DC supply if I ever need one. I appreciate all your guys time, energy, and input.

 

[jonasnoble]

My output transformers came from Cinemag. They are much smaller than I was expecting, I doubt they'll do what I need them to. But I'm experimenting, so it's fine. At least I will learn what doesn't work, and why.

I'm going to buy some bobbins and bells, and proceed my first attempt at a REDDI style OT.

 

[Wavelength]

Sorry late to the party. I do this all day long and use a NJM2396F63, which is a low drop out 6.3V 1.5A regulator with enable (i.e. 4 pin). I bought the last 400 Digikey had last year and another 120 from Mouser. You can usually find them on eBay. I use them in guitar amps for 6072A and I am working on a 300B amp right now with 6SL7/6188 drivers and do the same. Great part, zeners in this kind of role are not so good. You would be better off with dropping resistors or CMC (common mode chokes) to get a good low noise supply.

 

[Simon B]

Very simple option, well known to my Dad, now aged 92; a lead acid accumulator (A) battery. One of his tasks, from the age of 8, was to carry this lump, every couple of weeks or so, down to the local radio shop, where it would exchanged for a charged one.

Times change.

 

[RoadrunnerOZ]

Very simple option, well known to my Dad, now aged 92; a lead acid accumulator (A) battery. One of his tasks, from the age of 8, was to carry this lump, every couple of weeks or so, down to the local radio shop, where it would exchanged for a charged one.

Times change.

I remember buying tubes from a radio parts store in the late ‘70s - they had everything you could want, modern (then) and vintage. Perfect for repairing studio outboard gear.
They also had WWII portable tube backpack transceivers for sale, US & UK - these operated on an external battery if I recall correctly and used battery DC for the filament and a “wobbler” or vibrator relay switcher which created low voltage AC, the output of which, when coupled to a transformer created high voltage AC, then rectified to DC. The vibrator circuit was a relay with coil supply coming via normally closed contacts, supply would shut off the moment the contacts opened, then the contacts would close and re-energise the coil, this continued in a buzzer like fashion. Battery was fed through the then oscillating relay contacts to the primary of a transformer to get HT. These ran on 90-150V HT or thereabouts. The larger field units were 100W using a 4 pin output triode. I think they started introducing transistors into field portables during WWII.
It’s a shame places like that store and the owner, who was a WWII radio op, have gone.