BMK2 used in cathode-biased amp

[K O'Connor]

Hi Guys

Post-20 says how to size the bias range resistors based on doubling Rk as Champ81 did.

In general, the range resistor should be scaled with the bias pot value based on the increase of Rk that you incorporate. For example, if you increase Rk from 150R to 220R, this is an increase of 70R, which 0.47 of the stock Rk. We want the voltage across the bias pot to equal the portion of Rk added to limit the maximum current through the tube to the stock value. The range resistor then has the remaining voltage across it equal to stock Vk.

It is probably simpler to view the two parallel resistor pairs: the first is stock Rk at top with the added amount below, so 150R over 70R. Say you have 25k pots. The pot portion of the bias divider is the lower part and the range resistor is the upper part. We could simply divide 25k by 70 to get a multiplication factor to calculate R-range. 25k / 70 = 357. Now, 357 x 150 = 53k6.

If the bias pot was 1k, then the range resistor would be 2k14.

If we actually built the new Rk by stacking two resistors as post-6 suggests, where the upper portion is the stock Rk, then we can eliminate the need for this calculation and tie the bias pot across the added resistance to ground.

Since the example 2xEL-84 amp begins with Vk=15V, we can see that the 25k pot voltage divider dissipation is so low as to be unimportant even with 22V maximum across the new "220R" Rk. The divider total resistance is 78k. The 1k pot divider warrants investigation as its net resistance is 3k2, roughly, but total heat is 150mW, shared as about 100mW in the resistor and 50mW in the pot.

In Champ81's amp, Rk was doubled, so the total Vk could be 30V. You can see that the dissipations do not increase greatly compared to the example and 250mW resistors are fine even using 1k pots.

[Champ81]

Thank you for the explanation.

I am not sure what you mean by a stacked Rk.  Do you mean resistors in parallel?  Meaning the each bias pot "x" ties to each Rk directly and seperately?
Then each resistor would have it's own bypass capacitor as well?

For the range resistor I noticed in the BMK you provide very small metal film resistors. Like 1/8W. Is this in any way done to make them fuseable?

I wanted to summarize what I have gathered from all this in this layout.  I hope I got it correct.  The schematic is preferred but in case some need to visualize it for laymen like me.

[K O'Connor]

Hi Guys

"Stacked" Rk or resistors in general implies a series connection. 150R + 150R = 300R as you did in your amp.

Your drawing is correct.

Personally I only use radial-lead electrolytics as the welded lead to the can on an axial often breaks.

The small resistors in my kits are mostly 600mW 1% metal-film. I never use carbon. The tiniest resistors in the kits (usually 34k8, 125mW) are only used in places where there is not much voltage or heat, such as grid-stops and in switching circuits.

Larger resistors (1W+) are metal-oxide flame-proof or wire-wound flame-proof.

[Champ81]

In post #6 you also mentioned a variation circuit where a range resistor is saved. My understanding (could be wrong) is a second Rk is added with its own bypass cap and each pot "x" would tie to each Rk separately?

[K O'Connor]

Hi Guys

Read the post again now that know that stacked means series.

The increase to Rk is made by adding a resistor between stock-Rk and ground. The bias pots are wired in parallel to this resistor.

Whether Ck spans only the stock Rk or goes all the way to ground is up to you, but be careful of the stock-Ck voltage rating and life-use as subsequent posts describe.

In this stacked-Rk version, you still need the caps from the wiper of the bias pot to ground along with the safety resistors.

[Champ81]

So basically you mean in post #6 the bypass cap could span only one of the stacked cathode resistors. Otherwise the increase in Rk to span two stacked resistors might be higher and so you have to account for the voltage rating.

[Champ81]

Made a layout for the variation circuit.  The range resistors are removed and pot "x" is tied to the junction of R(add), R(Stock) and Ck.  

Do I have this right? The Ck could span both the resistors in series I think if rated at 450V?

[Champ81]

Initially I wanted to have an adjustable cathode bias to balance the hum. And that was achieved.  I am and was still new to this idea of biasing amplifiers but still learning. From what I gather now the adjustable cathode bias is varying potential difference between the grid and cathode. In a fixed bias amp its varying the potential difference between the grid and ground.  This was new to me.

Would it be worth it to convert this completely to a fixed bias amplifier. Tonally would this make any difference? Also because a fixed bias amp is referenced to ground you could also vary the current draw without the push pull configuration. Am I correct in saying this?

[K O'Connor]

Hi Guys

Champ81: Both of your drawings are correct.

Regarding bias: Tubes are CATHODE-CENTRIC so their reference is the cathode. Grid voltage has to be negative with respect to the cathode for there to be control over cathode current. Plate voltage has to be positive for there to be current through the tube. Screen voltage (when applicable) has to be positive with respect to the cathode for there to be current through the tube.

VOLTAGE IS RELATIVE. There is no "negative" or "positive" until you decide on a reference point, as explained in RSG Ready, Set, Go!

In both cathode-biasing and fixed-biasing, varying the voltage between the cathode and grid is what varies the tube current.

The cathode can be at what you decide is "ground", or any voltage above or below. For the tube the latter makes no difference.

[Champ81]

Thanks Kevin

First question:
Its the terminologies that I'm confused  with. Essentially both fixed and adjustable cathode bias are doing the same thing. Varying current between the cathode and grid. In both cases the grid is always negative with respect to cathode.  At first I was thinking the term fixed bias is when a current can be varied/adjusted via a potentiometer or trim pot.  And a cathode bias is a non adjustable bias circuit where the cathode resistor self biases the tubes.  With the circuit I currently have now would it be correct to say it is now a fixed bias? Or a variable cathode bias amp.

Second question:
I am also under the understanding that in a cathode bias the grid is negative with respect to the cathode via the resistor which makes the cathode more positive with respect to the grid with rhe grid being at zero. In a fixed bias a negative voltage is applied to the grid while the cathode is grounded. But I cannot see how you could apply a negative voltage to the grid.  For example with tubes out you can measure the voltage from the grid to the chassis and there is a negative voltage. Is there a negative voltage via the grid leaks?

Edit:
I have capacitors rated at 25V across the wiper and pot "o". The voltage measured across it is around 35V and sometimes goes up more depending on the bias setting. I ordered a couple more caps rated higher. I'm surprised that these caps didn't explode. I didn't measure them when playing and realized up until now.

[K O'Connor]

Hi Guys

1, VOLTAGE is varied between grid and cathode which in turn changes CURRENT from cathode to plate. It is how a tube behaves.

The circuit is variable cathode-bias.

2,Everyone confuses the BIAS METHOD with the BIAS CONDITION and also the fact there are two uses for the term FIXED-BIAS

FIXED BIAS strictly speaking means that the operating point for the output stage is solidly set or "fixed" and does not change during a signal cycle. This is achieved in two common ways.

FIXED BIAS is also a bias method using a negative voltage applied to the control grid with the cathode typically grounded. This DC voltage is applied to grid via a grid-leak resistor and the signl is applied to the grid via a capacitor, which blocks the signal source from the negative voltage. The signal adds and subtracts from the DC to allow the grid to have varying strength in the control of cathode current, which rises and falls with the signal.

CATHODE BIAS is a bias method that allows the tube to determine its own idle point (idle current) working against a cathode resistor and with the grid effectively grounded. Again, a grid-leak resistor between the grid and the bias reference (ground) allows addition of a signal to the grid.

In all cathode biased circuits, there is no current unless the tube is in-circuit. You cannot test the bias adjustment (if present) without the tubes as you absolutely must do with an amplifier using negative grid voltage control.

CATHODE BIAS should NEVER be called "class-A". Although the tube runs hot, as is typically the case for class-A, the tube can go out of class-A during the signal size, which it does in almost all cathode biased amps using a bypass cap across Rk.

The BIAS CONDITION indicates how much of the 360-degree signal cycle the tube contributes to.

[Champ81]

Thank you. I think I'm beginning to understand this a bit better now.

1) A cathode biased amp is where the grid (effectively is grounded) is negative relative to the cathode resistor which sets the idle current depending on the resistance of the cathode.

2) A fixed biased amp is where the grid is negative through an applied negative voltage while the cathode is grounded. So the grid is negative relative to the cathode through an applied voltage to the grid.  

Am I correct to deduce that in both cathode and fixed bias the grid is always negative relative to the cathode. Where in cathode bias the grid is passively negative relative to the cathode depending on the cathode resistance. While in fixed the grid is actively negative through the grid leak network.

So in both situations they are cathode centric. But in cathode bias the cathode resistance is "actively" setting the bias condition. Where in fixed its the grid "actively" setting the bias condition. But both relative to the cathode.

[Champ81]

I will going with the variation circuit where R(add) is in series with the stock 150R.

I'm testing out different R(add) for this circuit. The highest was 820R in series with the 150R. This produces a high DC voltage of 50V across the wiper and ground of the pot. I've tested different values and it seems that the higher R(add) increases the rate ratio turn on the pots. In post 6 you mentioned an Rk of 500R per tube. In this circuit would it just suffice to use a 100R R(add) in series with the 100R to make Rk 250R as a push pull. Since the pot max current is the same as stock I shouldn't matter how high the resistance?

Edit:
Please ignore the post above. The resistance value of r(add) sets the minimum current to the tubes via the pot. So a higher resistance allows you to bias it colder and increases the range.

These are the minimum values I measured with the variable eing R(add)
150R
Right side min 24.8mV
Left side min 23.4mC

510R
Right side min 13.1mV
Left side min 12.9V

820R
Right side min 9.4mV
Left side min 9.6mV

820R + 150R = 970R
Right side min 8.3mV
Left side min 8.7mV

820R + 300R = 1.12K
Right side 7.5mV
Left side 7.8mV

820R + 510R = 1.33K
Right side 6.4mV
Left side 7mV

The only other thing to consider I found were the capacitors on the wiper to ground of the pots. The voltage through goes up significantly to about 50plus volts so you would need to put in bigger caps. I think this would be a compromise between a larger range vs smaller sized caps on the pots.

[K O'Connor]

Hi Guys

It is probably best overall if Ck spans all of Rk, whether Rk is a single resistor or two in series. The alternative is to place a second cap across R-add when two resistors are used. This will keep the bias more stable with signal.

C-add should be a high value so that the effective Ck value is not reduced too far. The usual 22uF or so is pretty skimpy.

[Champ81]

I just updated Cathode bias variation (MV) on post #30.

[Champ81]

I've heard of people saying they use potentiometers rated at 2W or more for a bias circuit. I think that's way overkill. Or is it?

Let's say we have a 50K linear resistor and the voltage across the pot is 43V (in my case with a 620R Rk).

According to ohms law P=E^2/R
I get P=43X43/50k = 36.98mW

With a safety factor of 2 it would be 73.96mW.

So you can even have a pot rated for 100mW and it would be more than double the safety factor. Any reason why someone would go with a pot that high in rated power? The smaller pots are much smaller and inconspicuous. But more importantly even a 100mW pot is above the safety factor.

Of course it depends on the circuit and voltage variance between bias circuits. But I can't see anyone needing 2W potentiometers. Of course it doesn't hurt though if you have alot of space in the chassis. Where I think I would go higher too.

[Champ81]

I have so many questions coming up.  For a typical el84 power tube on the datasheet it states the typical characteristics have Ia=48mA.

I'm measuring 36mV=36mA from the probes with the pots turned to max so I should be able to have it up to 48mA = 48mV from the probes am I correct?  This would be max of course.  So maybe just under 48mA=48mV max. I believe the max of 36mV is from the 150R cathode resistor setting it to that stock max.  Could I lower that cathode resistor down in order to allow more current to go through for a higher max current but also limited to under but much closer to 48mV operating max of the tube?  Would this work?

[K O'Connor]

Hi Champ81

Tube data sheets should indicate the limiting values, but not every sheet does especially ones produced for tubes recently. Semiconductor data sheets ALWAYS include limiting values.

Tube data sheets generally show a few applications, which in themselves, do not tell you the limitations of the device. You have to remember that the "limit" is exactly that - the value that should not be exceeded. However, every value below that is okay and safe given that the device is kept within its dissipation (heat) limits.

Tube data sheets are also confusing inasmuch as the protocol used by the manufacturers changed in the mid-1950s BUT not all data was updated. Modern manufacturers still parrot the obsolete protocol, which makes their products look very poor indeed. The old protocol included tolerances for the manufacturer's own deviation of production PLUS anticipated sloppiness of equipment design where the tube would go.

Your amp was designed to provide a specific maximum audio output power. The PT and OT work together to make that power available. The amp designer decided to make the amp cathode-biased. You can change it to fixed-bias but you will not get more power than the PT-OT allow. Similarly, installing higher-power rated tubes will not get more power in this circuit.

TUTs explain how amps are designed. what mods can be made, what is safe to do with tubes. Other references, such as RDH Radiotron Designer's Handbook, follow more restrictive rules but also explains everything. The presentation may be beyond your skill at the moment, where TUTs make things a bit easier to understand and are on-point regarding MI application.

[Champ81]

I appreciate the information provided. 
In TUT 2 it calculates the current using 
Ik=Pa/(2XV+). 
Using the formula for example on another fixed bias build with two power tubes let's say with el34s.

Ik=25W/(2X442V)

Pa = power rating of the tube 

Which would give me 28.28mA

As I read in TUT 2 this is the reccomended value per tube which I usually bias at. It states biasing lower or higher is not recommended due to tube life and sound in factor.


In the datasheet it says for an el34 100mA. This is the maximum current? Although the datasheet might not be accurate for the actual tube on hand technically you should be able to set the current to 100mA per tube max?

---

[Champ81]

Hi Kevin

I just wanted to mention to you. I was on hoffman forum and posted your bmk schematic. On second i didnt think to ask you first if it was OK with you to post it there. I just wanted to let you know. Also if you would prefer I can delete it from the forum. Please let me know your thoughts. Thanks