Bias supply question

[W8ACR]

Does the bleeder resistor in a bias supply also act as a grid leak? If so, how does one compensate for this? I am close to completing a homebrew transmitter that has two taylor T-20's in the final and a T-20 also as the driver. My bias supply is a simple full wave rectifier circuit, and provides about -80V and has a 20K variable WW resistor as a bleeder. I was hoping to simply tap the bleeder at the proper points to obtain needed voltages. But now I'm not sure this will work. I'd appreciate any advice.

Thanks, Ron W8ACR

[KM1H]

What you are describing is a straight fixed bias supply and it could also be a battery. This will give you maximum protection in the event you lose excitation.

You should provide enough bias in standby to completely cut the tubes off so you dont get diode noise in the receiver. Use the full voltage for standby and the resistor tap for operating bias.

Carl
KM1H

[WD5JKO]

Does the bleeder resistor in a bias supply also act as a grid leak? If so, how does one compensate for this?

Ron,

     If the bleeder resistor current is greater than the tube grid current, then the bias will be controlled and only rise a bit with increasing grid current. If the bleeder resistor current is smaller than the tube grid current, the bias will rise immediately and directly when the bleeder current is canceled out by grid current. Then Yes this will behave somewhat like grid leak but a little more complex to calculate.

     Remember that your FW rectified bias supply can only source current and it cannot sink current. The bleeder will sink current to a point as I said above. You can also get fancy with an active shunt regulator circuit. This approach allows you a stable bias without the need for a lot of bleeder current. This can be as simple as a series resistor and a zener diode.

Jim
WD5JKO

     

[WD5JKO]

This can be as simple as a series resistor and a zener diode.

Here is an example of a transistor amplified -100v bias supply:
http://pages.prodigy.net/jcandela/CE20AQRO/PS.JPG
The transistor was a TIP-50, and zener was 100v 1/2 watt. This simple circuit works extremely well saving the cost of an expensive 5W zener diode. The zener knee is amplified by the gain of the transistor which improves regulation.

Here is where I use a 5w 100v zener diode to make a bias supply, and from there I use an adjustable zener diode (TL431) to make bias for a class AB2 RF amplifier (some grid current). This adjustable zener is set for a range of -9 to -33 volts with the twist of a pot:
http://pages.prodigy.net/jcandela/Central%20Electronics%2020a%20QRO%20%232/CE_20A_QRO_Power_Supply.jpg

If you need -80v for your T20 two stage RF amp, first figure if that is to be fixed bias or a combination of fixed bias and grid leak bias. If you are going to plate modulate the final you DO want some grid leak bias. Maybe have enough fixed bias to cut off the tubes when RF excitation is absent. One thing for sure though is that variation in the final grid current should not effect the driver tube grid bias. A simple way to eliminate interaction is that amplified zener using a power NPN transistor such as the TIP-50. The transistor will have to pass the grid current, so bolting the collector to the chassis will be required. This is shown in the first link above.

Jim
WD5JKO

[N2DTS]

I think its good practice to have protective bias and grid leak bias, decoupled by a diode.
That allows protective bias, but also allows the grid drive to build up bias voltage well above the fixed bias volotage, so you go more into class C which is better under plate modulation I have been told.

On my 3x4d32 rig I do it that way, with a big wirewound pot as grid leak, and can get much larger bias voltages then recomended, at the expense of slightly more drive power.
Grid current can be the same, just much more bias...

On the 2x813 rig, I dont have the diode so I can adjust things as an ab1 amp, but it works the same way, a little more drive power gives way more bias, at the same grid current.

I think the bias bleed resistor does act like a grid leak resistor a bit, but most of the grid leak is in the grid leak resistor.
Make sure the bias supply cap is rated for much more voltage than the output of the bias supply, and the bleeder is on the low resistance side, and higher wattage than what is needed for just the bias supply voltage.

Going much bigger here is a good thing...

Brett

 

[Bill, KD0HG]

Ron: The trick for Class C is to use fixed bias up to the tube's cutoff point, then provide the rest of the operating bias with a fixed resistor.

When I get a chance I'll post a schematic.

[KM1H]

5W zeners are around $0.25

Carl
KM1H

[WD5JKO]

5W zeners are around $0.25

I bought mine local at a Frye's store, cost about 5 bucks for one of those NTE jobs. Looking at Digikey, I see 46 cents! 

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=1N5378BRLGOSCT-ND

[KM1H]

I think I paid .28 at Mouser for low quantity of like 2-3 a month ago. For more common values I buy 10 at a time and they get used fast.

Carl
KM1H

[WD5JKO]

I think I paid .28 at Mouser for low quantity of like 2-3 a month ago. For more common values I buy 10 at a time and they get used fast.

  Yea, but take that 25 cents and add shipping and handling and your well over 5 bucks. Your smart to buy 10 at a time.

Also consider that if you had say, a 100v zener pulling 10ma from the bias supply, and then you use that for the fixed bias portion to a class C final where the grid current is 100ma, we get 10-100 or 90 ma through the zener. Take .09 X 100 and you have 9 watts dumped by the zener. So a 100v 5W zener won't do, nor will two 50v 5W zeners in series. Why? Because the heat of a zener is dumped into the leads, and the spec sheet assumes the leads are short and soldered to big fat copper traces which act as a heatsink. To avoid all this thermal mess I like to use a little zener connected to a big transistor like I linked to earlier. Then the transistor can dump the heat to the chassis. Also those HV zeners will increase in voltage 5-10% if you let them get spit boiling hot.

Letting the topic drift as I ramble on:   

A 6.2V zener rises about. 2.2 mv / deg C and a silicon diode goes down the same amount. If you series up a silicon diode (forward bias) and a 6.2v zener (reverse biased), then you have a poor mans temperature compensated reference so long as the two devices are glued together and fed with a constant current. The silicon diode can be the base emitter junction of a transistor. Here is a very stable +375v 100ma supply based upon a 6.2v zener diode:

http://pages.prodigy.net/jcandela/Sherwood_S8000/HV_REG/SS_HV_VOLTAGE_REGULATOR.jpg


Jim
WD5JKO

[Opcom]

regulators are always better but I still use heavy bleeds 10-20x grid current with good results. Wasteful I suppose.

[KM1H]

Ive been designing professionally since the 60's Jim; zeners were around then also and I used them as a reference for the SS regulated bias supply for a 2KW amp. Had to use tubes for the 10KW job tho

The TIP works fine within its limits as do IC's.

Use what you got or are familiar with or what the project specification calls for. No need to bring a gun to a pillow fight.

Carl
KM1H

[WD5JKO]

Ive been designing professionally since the 60's Jim

OK Carl, maybe you can share a little your vast experience and post something here to address Ron's original post. I tried to do that as did a few others. I am sure you can add something more constructive than say a 5W zener costs a quarter.

Jim
WD5JKO
   

[k4kyv]

There is no need for regulated bias for a class-C amplifier.  You just need enough protective bias to keep the plate current within a safe range in the absence of rf drive, either in CW mode or in case of excitation failure.  Broadcast transmitters often use overload relay protection only without any protective bias.

In the case of leenyars and class AB or B modulators, good bias regulation is of utmost importance to reduce distortion.  I prefer the heavy bleeder method to fixed regulated bias, since the latter causes the static plate current to vary widely with only slight variations in a.c. line voltages.  If an electronic voltage regulator is used, it should be designed to sample the +HV so that the bias increases or decreases at exactly the same percentage as the plate voltage.  This will further reduce distortion by several percentage points.

[KM1H]

OK Carl, maybe you can share a little your vast experience and post something here to address Ron's original post. I tried to do that as did a few others. I am sure you can add something more constructive than say a 5W zener costs a quarter.

Several already covered the original question to a sufficient degree. I see no reason to add to it just to get my post count up or engage in a childish game.

Carl
KM1H

[k4kyv]

If an electronic voltage regulator is used, it should be designed to sample the +HV so that the bias increases or decreases at exactly the same percentage as the plate voltage.  This will further reduce distortion by several percentage points.

And I might add, somewhat lessen the requirements for a well regulated HV plate supply, although it still should be a well regulated as practicable.

[Bill, KD0HG]

Here's a hollow state bias regulator that I wrote up in ER #10, February 1990. It's part of my 304-TL RF deck. If you want to do a stiff hollow-state bias supply, this one works great and there's none simpler.

Its output voltage is adjusted to right around cutoff for the final, then R7 supplies the remainder of the Class C operating bias of the final. When R7 is shorted out with S3, then the bias can be adjusted to run the final in linear mode (optional).

Just about any tube, triode or pentode, can be used for the regulator, the only important things are that its plate current rating needs to be more than the operating grid current of the RF final, and a low plate resistance is good.  I've had the same 6BQ5 in my RF deck since 1990, the grid current of the 304-TL is around 80 mils with full drive in Class C. The 6BQ5 sinks it easily. A 6V6 or 6AQ5 would probably work just fine as well. So would a 2A3.

You can use this same shunt regulator circuit with a push-pull Class B modulator.

Parts values I used for a bias adjustment range of -35 to -260 volts: (to be modified per application):

T1: 500 VCT & 6.3 volts 1 A.
R3: 7.5K 10 watts
R4: 100K, 5 watt
R5: 10K, 1/2 watt
R6: 22K, 1/2 watt
C17: 10 uf, 500V

[WD5JKO]

Bill,

   I like that shunt regulator approach to class B/C big triode output tube biasing. With a flick of a switch you go class B and provide a bias source with low source impedance for good regulation of the bias over an audio syllable being spoken. If the line voltage moves, then the bias tracks too, like what Don K4KYV was asking for.

   The shunt regulator approach might look like a complication to a simple need, but I see that the current requirement of the bias supply is much lower than the grid current of the output tube. If you just had a tapped bleeder resistor with the bleeder current 10-20X the grid current, then (20X .08A) * 260v = 416 watts!! That would be one heck of a bias supply. 

I like it!

Jim
WD5JKO

[Opcom]

Mine burns about 100W in bleed for the 4-1000's bias.