12AX7 heaters in paralell with bias resistors

[Tom Bavis]

The grids are biased with an adjustable POSITIVE voltage with respect to ground - there is the usual 15V or so from grid to cathode. H-K used this biasing scheme a lot - some early Fisher integrated amps did as well, as did a Calrad amp I happened across... though that one did without an adjustment.

As far as using a 20W resistor - I would since there's about 6W there - a 10W resistor will run pretty hot.

[coman61]

O.K a 20 watter it is. I need to order some supplies anyway.

[EWBrown]

A good rule of thumb for using power resistors is to use a power rating of at least three times the actual dissipated power , so a 20W resistor is good value for 6 watts actual dissipation.

It will still get pretty warm / hot, but at least it won't burn out.

I remember seeing in a Decware "Zen" amp, there was a 1K, 5W WW "Xicon" resistor in the B+ string. :o

Trouble was, that the current drawn through it was over 90 mA, and by running the math, that was close to 9 watts being dissipated through it. Yes, it got VERY hot, and the markings quickly got discolored, and I'm surprised that it didn't melt the resistor... .

I ended up replacing it with two 1K, 10W resistors in parallel, as the 90V voltage drop was excessive, not to mention cooking the heck out of the innards of the amplifier, and after the fact, the amp sounded a lot better.

Yes, it was running the SE SV83s very hard, around 45 mA, and 365V on the plates, but the tubes are reasonably cheap, and sound much better when forced into "hard" Class A operation. I later "morphed" it into a 7189 / 6P14P-EV UL SEP design, with around 4.5 WPC and it's pretty decent sounding, and a lot "ballzier" than the original design concept.

/ed B in NH

[erichayes]

Thanks for the schematic, Mark. If you draw the heater string on the same page as the bias control, it's much easier to understand what's going on.

The 68K fixed, 5K variable, and 10K fixed resistors are wired in parallel with the preamp heaters. The heaters are going, for the most part, to determine the voltage on the cathodes of the output tubes; that voltage, however, is going to be determined by the voltage on the control grids. It's a tail chaser.

A nominal 33 VDC is established across the heaters by the 83K voltage divider, with the wiper of the 5K pot determining the exact point. Fortunately, in this case, the engineers kept the breadth of the pot's influence fairly small, so as long as all four output tubes are operating properly, there will be around 35~45 mA cathode current on each tube.

The problem with this scheme, as is with all common cathode resistor schemes, is that, if one or more tubes or sockets should fail, the current draw on the remaining tubes becomes excessive and meltdown can occur. In this case, it can also cause heater failure in one or more of the preamp tubes, as well.

I'm stickin' with the 240~250Ω 20W cathode resistor, Mark. Just check the voltage on pin 2 of the output tubes; it should be around -11~12V with respect to pin 3--NOT to ground.

If you need more advice, you have my email address.

[coman61]

Thanks again Eric. The only thing that really sucks about this, is that 20W resistors are not that common. I was just at mousers web site and they have 2 that will work in this application. 1 a nice Vishay Chassis mount mount at $3.43 each, but the min order is 10. The other choice is an Ohmite solder lug type and they are $8.98 each but there no min order.
Yellow_Light_Colorz_PDT_13

[TomMcNally]

Handmade has the 25 watt aluminum ones for $ 5.99 ... he
has lots of cool stuff in stock, and fast shipping too.

http://www.hndme.com/productcart/pc/viewCategories.asp?idCategory=37

[coman61]

Thanks Tom, problem here is that I would still need 2 of them. 2, 510 ohm in parallel = 255 ohm. They Don't have any 240~250 ohm resistors.
DARN-IT. Yellow_Light_Colorz_PDT_21

[EWBrown]

I've re-read this post from the start, and I'm wondering if all the adjustable pots, resistors, etc are simply for "floating" the filaments at some DC voltage above ground, in order to reduce hum.

Sometimes it helps to raise the filaments' "bias" to a value near or even above the cathode voltage, in order to reduce noise, hum, etc...

Not to be confused with using a 12AX7 or similar 12.6V, 150 mA tube as a "cathode resistor" for a quad of EL84s.

just wondering...

/ed B in NH

[coman61]

Ed, I think you will need Eric to anwser that question. If you like I could send the schematic to you. I hope this amp sounds good after all this effort to get it working.
After I finish this project I'm going to do a scratch build. I have an old heath kit AA100 (I think) the one that uses 7591s. I want to do something like Tom McNally Did with his .. (from his web site) but built on a differant chassis and differant tube style.
I also have an old GE amp that was set up to run 6973s that I could use as a donor for Iron. I like the idea of useing 6973s, they sound a bit like EL84s with a bit more kick.

[EWBrown]

That schematic would be interesting to see...

I had an old AA100 on which the pink phenolic PC boards were seriously damaged (cracked and just plain broken) so I salvaged the iron. It was dirt cheap, so it wasn't a total loss...

6973s are a neat tube, I have a quad of NOS RCAs. FWIW, 12AB5s have the same pinout but are rated more closely to 6V6s or 6AQ5s.

They were intended for use in 12V tube car radios, in PP for 10 watts out.

Once I get over this "martian flu" (it keeps coming back) then I can get back to any tube projects that have been in limbo for the past few weeks...

/ed B in NH

[EWBrown]

OK, I got it now! Yellow_Light_Colorz_PDT_01

Mark sent me the Fisher 480 schematic PDF file and I worked it out:

This circuit is using an odd mixture of cathode and fixed biasing. :o

On the schematic, the EL84 / 6BQ5 cathodes are indeed at 33VDC above ground, but the G1 grids are set to +21VDC (adjustable) above ground, so that the cathodes are approx 12VDC higher than the grids, which is about the right K to G1 voltage relationship, for the 6BQ5 / EL84s to each run at approx 35 mA cathode (140 mA total current for the quad) current.

Very weird approach, but it should work fine, as long as the pin 11 in the pre-amp power connector is making a good clean, low-resistance contact. This also requires a well-matched quad of EL84s.

The series string of 6.8K, 5K pot and 10 K resistors, connected across the three series filaments, (connected to the 4 EL84 cathodes at the + end and to ground at the - end, allows for the grid voltage setting, adjust for approx 21VDC or 140 mA through the three preamp DC filaments.

The resistors' total 21.8K, so the current through them is just a tad over 1.5 mA. 1.513 mA to be "anally" precise. So the 6.8K resistor drops 10.3V, the 5K pot has 7.565 V across it and the 10K resistor has 15.13V.
(I know, Mister Spock, the numbers don't exactly add up) Yellow_Light_Colorz_PDT_02

So the grid bias voltage can vary between 15.1 and 22.7VDC (probably a bit wider range, as the 33V from the EL84 cathodes will also vary accordingly, as the current (and thusly the voltage) is adjusted. It is a very simple voltage and current regulator, using the power amp tubes as series pass elements, as well as for audio power ampliiers.

Double your pleasure, double your fun....

What this all amounts to, is a very creative way to save a buck or two, and get DC filament voltage for the high gain preamp stages.

This allows the quad of 6BQ5s to act as a current source for the three 12V 12AX7 filaments, which are wired in series in the preamp. It works out to 33V @ 140 mA approx, or 11VDC for each tube. It is OK to run the filaments a bit lower, and allows for longer tube life.

Think of it as the electrical equivalent a snake swallowing its own tail, but it also knows just when to stop... Yellow_Light_Colorz_PDT_05


/ed B in NH

[coman61]

Thanks Ed. I think you and Eric have come to the same conclusion. The only difference that I think that you noticed that we did not is that one of those resistors is auctually a 6.8K not a 68K resistor. I think that makes it a bit easier for us all to swallow on those bias resistors in series.
Now back to the question at hand Yellow_Light_Colorz_PDT_09 Do you agree that in order to use this amp,... (without)...the Fisher Pre\tuner... I will need to install a 240~250 ohm 20 watt resistor where the 12ax7 heaters were... (pin 11 to ground)... Yellow_Light_Colorz_PDT_11 Yellow_Light_Colorz_PDT_11 I think we all agree. Yellow_Light_Colorz_PDT_02

[EWBrown]

That (220 ohms to 240 ohms, 10 or 20 Watts) would work just fine, but would be basically "wasting" about five watts' worth of electrically generated heat. This does, however, address another potential situation, addressed in the last two paragraphs..

The simplest fix, if you don't use the preamp, is to remove the 6.8K resistor, pot and R24 5K resistor (or at least just disconnect the EL84 G1s from the pot's wiper) connect the former R24 pot wiper connection to ground, so that the G1 grids are now properly referenced to zero volts ground theough their 330K resistors, and then connect a 100 ohm, 5 or 10 W resistor from the four conjoined EL84 cathodes to ground, this would then simulate the approach used in the ST35 and SCA-35.

The combined (approx) 140 mA cathode currents would then gnerate (approx) 14VDC across the 100 ohm resistor. Which would approximate the correnct combined cathode / plate current.

The elegant approach would be individual 390-430 ohm 2-3W resistors and s 100-470 uF, 35VDC bypass cap, for each EL84 cathode.

But... Nothing is ever quite as simple as it initially appears...

Either method will leave you with about 19-20 VDC higher plate-to-cathode voltage than with the original circuit configuration, so perhaps the 35 mA can be tuned down closer to 30 mA, by using 430-470 ohms for individual cathode resistors, or 110-120 ohms for the combined resistor. Or just go for broke and run those tubes "hot"... :o

Another approach, in order to get around 380VDC at the OPT CT connection is to insert a 100 to 150 ohm, 10 Watt WW resistor between the CR1 cathode and the C15 100 uF first PSU cap. That would lower the 400V B+ to around 380-385VDC, which is perfectly fine. voltage

Or simply go back to the first approach, skip the rest of the "fixes" and just use the 220-240 ohm power resisstor.

Another snake is swallowing its own tail Yellow_Light_Colorz_PDT_08



HTH

/ed B in NH

[coman61]

The elegant approach would be individual 390-430 ohm 2-3W resistors and s 100-470 uF, 35VDC bypass cap, for each EL84 cathode.

But... Nothing is ever quite as simple as it initially appears...

This may be the solution that I have been looking for. This tube matching is becomeing a real PITA. Yellow_Light_Colorz_PDT_11